ROOM2

Speaker 1: Ah, tá. Uh uh uh uh uh uh. Yeah, I think it was late.

Speaker 1: Yes, no, I mean don't say when you fighting a chance to say something like this. Wait, can we do this?

Speaker 2: Iti

Speaker 1: masă.

Speaker 2: Core the door. Oh, okay. So don't modus them, the mode I show this.

Speaker 3: Take him once replace moss.

Speaker 2: Oh gosh, mute ,

Speaker 3: You could already also get open yes. Hi. Uh C C A Shinha demo, C C A, Jangriat Skrienka, can show Shakyimas, Pasto A PI demo, Tajimas. Nayonish Takata T you no took testok no tenano de CSA, the Zimbabwe Kito Shina, Tsio Y dokia, atanode, Konotema Shimasha. It to happy marino de

Speaker 3: simple. They say , Yumea, Krenai Novtanochkoki no de Khavisri Kototo, Kasseni Idjusri Kotok, Kanonište, Sumir Menzikyo, Scalto Sastai no to The senior type of the side of the Rio.

Speaker 3: Ai, Ijoi Kuruno Show Kai de Uh

Speaker 4: , Tomatsun Kazaki , do de Chotamaini, uh conference hap , shot kin on nayoto How will you cry to the indiscreto? They come to Kyoto Happy No Ma. Prompt on your command. Hi, so nakata today. Hi, on the same

Speaker 4: Are back. I know

Speaker 3: for the same shaggy glass. Hi, ish all got send the narim Like us. Hey, tada ano, t'skraste manzo gush e grasemanzo guster, to elavic to kto potaio de kina in de kaksimukta yet's kai byosan po teda mister nopp, you sage. De Oskio

Speaker 3: Tokao Jimun de Atemite Zmatraini. Shidiomu, CCI l'escurikaina, di noto, kyodo happywa jivun nachtulat, akina mistun nachtonya, o smilat di yunat. Kiko simple now. So since I don't know nay you not know that. So you nay you that tokenky datum, mite male to a design. Uy , no

Speaker 3: ah, it 's uh good taking a show It comes like AIDS you share. They to Rogun Shai to Tskaya Su Shikunain de Chokto Mendozainato Mutanske ,

Speaker 3: I'll copy Shef. Copy the The chat was created to hazard on that. Compose out

Speaker 3: she was dead.

Speaker 3: I'm good, sir. Okay, so getting out in the corag , moderat , kudot the AI Moderat, Kirodoga Eh , t'es en formotion. Decision agent

Speaker 3: , you can see that the case is a very good idea. This is valuable. Oh good you must it

Speaker 4: Should we have an application dissociate the kiton disked all?

Speaker 3: So

Speaker 4: we are an application. That's poor than the source network on the system. Oh, for the email. Goliath is so straight. Alright, but I'm putting put it

Speaker 4: there. This so they don't know that. Uh animal sent it. What

Speaker 3: are you not doing so they talk about the sky

Speaker 4: Tokyo more question choice.

Speaker 3: Oi , vai ganhar so dechango de costo. The kill it. Hi.

Speaker 4: Ah, por lo así , es que nunca Because you could just take the dog.

Speaker 3: Say there you sara yuri mental.

Speaker 4: Upon all

Speaker 3: this revike. Chuck to the Kaiba Surtout.

Speaker 4: Quite so. Nima and we are going to accept the car. Alright, mm-hmm. Nee, dan dicht hem , nee. Ooy.

Speaker 4: Second Well Python, yeah. Python components are to give me sure. Dynamite Ma fa io tampo, yo mi condiro no de to

Speaker 4: Dzentay oto stak kai shu o steg grek hazel It 's the melo. skijana in this law. You know, yeah, decorated must never. But oui

Speaker 4: Project on the store.

Speaker 3: So you could configure to kanisiru no.

Speaker 4: Taibu li china kanjini. So le pokonadi musei. Talk your through.

Speaker 3: Era

Speaker 4: adimasse. So this next temperature tingle naked They're custom filter on that shit. Except top.

Speaker 4: Imam Mike I accept the botan o'stji koshamaske do to set tenio tewa to zembel j dozi komodikiru. Macho tocco ainde, accepto y cuico, mina gara. Ya te más. Komm ik om temporal toch stick. Alright, gaan we zo misteren. Oh, they must have a Topa to

Speaker 4: prompt to ni no kideba. Tino ga. Hi.

Speaker 3: Ya, esto

Speaker 4: prácticas de ponemos Uh they file solo mamma

Speaker 3: Sai shop yarn.

Speaker 4: Because to practice you team to get to zero come out. Test on a snake desne. Testotto

Speaker 3: Chio Petano. Petno for the cara. Sai cara. Skudeba. Natattiin kansi.

Speaker 4: Toisin

Speaker 3: teketanga.

Speaker 4: Tois ne kun ajatullut jo. Näin mä sain.

Speaker 3: It's not for the kids. Configure the inokat for you don 't. Tada , free project , so no mama smitherity, to kologa, o inokanaat.

Speaker 4: Hierkimassa. Tässä vuorovälogy massa ne. Yeah, best of practice there.

Speaker 3: Aperty oh best of track. Kuisin

Speaker 4: yksiksi vuoksi? Mä kurkei, että saa itse kikkaa Delukana, Koleva, my projekto no. My projekt ,

Speaker 3: on

Speaker 4: atmosle kuleva. My project of settings are

Speaker 3: to the option again Ma. Ma to

Speaker 4: the corner not.

Speaker 3: Biggina. So

Speaker 4: you can motion a fast way, you must go. So look. Oi to

Speaker 3: item tutorial switch

Speaker 4: Ah, correba battery, must go night. You guys must get a mite car Imagine force. Scoot decorator Prompto E 'n Tony Kete.

Speaker 4: Et j'ai intelligent. Il va logique pour securer.

Speaker 3: Napasulaksha Sushinha Nustoa. Matkatahola Indian I can 't.

Speaker 3: So the

Speaker 4: package installed the stuff that I got. So that's next. Jungho was your kaitans.

Speaker 4: Application kiddos ten it 's got core de kid Should we not have a solace? Oh kidos dens assault Uh no main too Soy Al Bimbo, Bimbro not the Moy

Speaker 4: I 'm going to put the email to the upper note. Made Mako him with you know for a husband. Imagina toyunova. I quantum menu quad isn't it?

Speaker 4: Victor succession scheme.

Speaker 3: Oh, tame this. So you are niche to m is rakunarmi tight for the I

Speaker 4: 'm going to cast the AI

Speaker 3: ni.

Speaker 4: Astri ya i no

Speaker 3: lightem animaska tikata posiruto Kingdom detector. Don't want to detect C'était qu'il est allé te connaître.

Speaker 4: paissemassa il meille. Enkää kokonaassa ilmaa. Aita onkin kyllä näin. No

Speaker 3: cool seen to um tan va. So

Speaker 4: come off the skeleton.

Speaker 3: Maintenant height, you know. I saw the main cochet to announce

Speaker 4: it.

Speaker 3: Don 't argue.

Speaker 4: Zemir to Farga Kidai Kais

Speaker 3: Matino chatto deh, so no mean dot to pay some haira done and that de chatudeh. Kita Saba to paito kanhoga orkar seikamu desnetino. Tanjango projecting set the file no chotto nandakana

Speaker 4: Windows dokto kuno kakikataga areukana asumahana Tätä. Siete, että kaikki otte, joku tosminen tekee.

Speaker 3: Mm-jip to muna can this Ma eye that canate condemnus

Speaker 4: Okay, let's do it. So , meanwhile, let's get him a support. Do API doctor fight in special?

Speaker 3: So now No there is

Speaker 4: no and get the post of the master

Speaker 3: So the upper

Speaker 4: idol. Project okay on a token to an animals to this new

Speaker 3: avance Mana, some code minimum achievement inokana pick van display. Jiganga. Ez auto indiquet.

Speaker 3: I

Speaker 4: also work this night. Yeah, cool looks shut, yeah. We talk about nice. What are the jungle people that is going to mc?

Speaker 4: Thanks for screening as I told the net. Ah, what 's inputs? You'll see our quite nine.

Speaker 3: Ediamo. También se do. Well

Speaker 4: she okay , she was animation, eh?

Speaker 3: Happy Wood. Takai modelo na socksemikayna.

Speaker 4: It's gonna sort of sort of know it's been more Don't quite see it.

Speaker 3: Kissamos there. Hi. Hi. Hi. Yeah , we need it. They move forward spawning, fancy and this cup. So this name. I know you're not don 't mad

Speaker 3: and I'm not going to be able to do that.

Speaker 1: Kware ima jango tokam fast API in a kodo tetskut demasket tokano sono kodo tetsuono database tokasono sakino sousa tukama de modzembe. Dikirukan also someone would not

Speaker 4: command console the j codekiru kotovaz, zembekiru this or no database. Uh this tenure looks at the crin this guy.

Speaker 3: Test to flipping. Test to a fitting.

Speaker 4: Aye, kyuki

Speaker 3: o girito.

Speaker 4: Yes, that's good back. Ah, most core does it yet.

Speaker 3: Yeah. And katasya, it 's a bit pakaray to ask you. Ey, Hatchiroko. Sono

Speaker 4: charaj, sonuchi, sonno se ke ma wari mo , kayaku e gaioli na te, chanto, mukashi poi kanjio, yomikoma sereba.

Speaker 3: So you can take an idea. So

Speaker 4: we tried to test the background. Test the background, testman such a

Speaker 3: You know what I said to kill.

Speaker 4: Oh, and we use uncall wrong. Imagine James Karamuji quickly must be talking. It gas validation test on the parallel Dynamo chip to validation came up. Ne? Yeah, they kite creator on the snake.

Speaker 4: J soon in our cannon. Testo fasto dikto sakini testo kaitika gisona kimete kara Testokite, J Sony Narhazan o de Soachimpay this ne. Jitswa Saidai Haku Ichimade iketaket ikeru yoni natilikedo. So yosh to slimita l'amtisto kakiso

Speaker 4: danat di natuto o moj masa. Validation of this more. He

Speaker 3: wanted to take care of the key so you know what to solve it.

Speaker 4: So this

Speaker 3: name.

Speaker 4: Document to keep kakuna walakunya so paitayote.

Speaker 3: Oh.

Speaker 4: This

Speaker 3: took all a couple more She's like away. This took all the chinchists with the hotel.

Speaker 4: Storiko, story. I

Speaker 3: 'm the shide tede kak tenu. I don't know than such. So

Speaker 4: that's next. Which I core, what up this, ne? I'm going to show you a night's morning that time this one.

Speaker 3: Hey.

Speaker 4: Imanan Kayum Tokyo Dikitan. To hid ik inakunarz. Mukkaya tohidikina gunar haza. Wow. What is the updates next?

Speaker 3: The other way for it on your architecture So

Speaker 4: many are jango. I just want to It's wrong, AID. Jungle to screw it up.

Speaker 3: So API motivate. Scottin, I'll touch that Scott De Hog. Eatonash, I don't

Speaker 4: think I must. I didn't say that must. Yeah

Speaker 1: , who did what it's

Speaker 1: Imagine we call it softly more.

Speaker 5: Tokyo Kasana Haraghapios acetita kimas. Now he did to your account to make it hover to link to you at most forum and state that you say like this. But the has memorized the genre sang , the games of Tokyo Gasani ni suste jungle , Coting it

Speaker 5: ten years. It took us next day and to Iggy sno, krakenycono rosiz, to itukon is postary , stereo chitu. The OSS encrypted fields to a package of the video. We see subscriptive command form. Jazzband

Speaker 5: , Puriatskaritokon , to Macho to instancy. Tokyo Gasadisika. The ano ima sakin

Speaker 5: is a sibisawa eighthiga soyustakai mass. Cast to the asmatinum agasan disagre. The nise niche it is a keyhood cracking technologies to venga. Hi.

Speaker 5: So put a hazimate or the tanjo stuff. Tema who recame today. Ma varito system that the engineering incidents. Yeah, I think that's a good thing.

Speaker 5: Ya te sunde, onyak kizo u sabisha ataristi. Maste Ma data doon sikerit. Kraken to you know to kyogasano, I 've been a system to stay and it's

Speaker 5: not data Economic side of Stimme. Matokyo Gaseno system ariona soyo disté. So that you crack into your system, so you have a stick with application to the film. So you didn't know.

Speaker 5: I did. So they uh architecture. No, please, classic. Hi The ticks

Speaker 5: are taking a ticket. The kingitzka to must teponu maradekaponda to korola sono jango sostegazo. The Qita Kingwa interface in the moderate bubble to Matop interface on Dinoco Paranigsk , Infrastructure in Dorovia. Infrastructure commandment. Hi. What about it?

Speaker 5: JavaScript for JavaScript for the keyboard, masaki most It's the caranain de massive screen. Ma atua ma jango ma kist kedio on a frame a could anything. Motti must. Uh no , I'm not sure if you're not going to be able to do this.

Speaker 5: Scotty a green architecture, it's again silicon casual to the mass. But so this is you need to taste the cut.

Speaker 5: Hey, but to yeah, what if you're still there? I must a jungle can't even put easy you 're picking up. No use up, she's only o it tomorrow, mastering can link to the first time. Tanz

Speaker 5: Hi, the same one to pick up. Yeah, but I speed to the no bit in this game. My motto party jungle and what we 're going to do to must be a problem.

Speaker 5: Kengrai, jungwe can start, pake, arkutoga wakarkatumori mas. So the today is a moderate constant. Yeah, to changono databo database modelo to use it in its butt no business.

Speaker 5: But then I need this get the more. My expression to kamot, I have a state, my case data. Attua, ma kriito a preditka, myskin to come disketum. A paruka kreito tukka, parukka preditka. A kidokide. Taking this to possibly no connection purum,

Speaker 5: to jungle no things call kaba for it , at the moment. Atła eto ma konatum date jango adomin to. Ma chango discovery. Adaming site to register to cluster the kid. At migration fire this migration fire to migrate to

Speaker 5: Migration firing is on ink on a Mato data migration by data based on the state. I just say for the migration file data to near no day and this video. Migrate to the data tone, migration higher.

Speaker 5: Maluka minna engineer ko una ziatayo, other steel, who is a steak. To zate commandoni zitinu. It 's a basic command of the case. And to Madoitsky Kataoselukati to Kayosin Tukini, which

Speaker 5: only been us. But a parchota toksina, yellka in this kid more.

Speaker 5: So no file yata or devi database. Sitten lukkest highlight. Ma who can crack in the attack , this is a savising API data together in this. Tonna monoga data who karate on show study. Anochos data is a good idea.

Speaker 5: Hi. This eh jungle adomin this guy. Whatever she engineer is kat tight. And data can still mock it. Eh matatonukas pastakatka. Matasanus pastaliu ananu katti uturumidio in state ismas If that's me, I'm saying the command of taita , atoming gaming commandos

Speaker 5: , state must be a good idea. Hi. They could see the guy deciding a commando comment, they know a actual commando comment. Adaming the commando zip code, screens there to go. Commando de pits kit parameter o fieldos And you look at the killing state.

Speaker 5: Say one, it 's it is The middle of mixing the grass of circuit. But I mean come to Kademu a Jiko. It 's a

Speaker 5: mi conday meetup in Alimas. Kotular in the skin. So they will meet the statements. I vento beats on eventually at the list to one second. Mayo, it caivasite. Ma non c'è mu, non kai un asiento great shuttle. Makura tabu and uh eventually to cade. Ukraine cut the canon say that is the body man in San Kaiuk

Speaker 5: didn't this give them a Masuta diagarano de Masurio Ano Yuski Panohoma De Nanka Muchatter Sirito Ano Suriba Mentuda Tercino de Surio Mituriade And this tea it is. My event ID got tara , whatever you know, database in the IDS event ID amatachita by anima accept it, total state I still

Speaker 5: Eventually , say you tamo tefinish the ideas. Hi, saiwani, matomidisni. Ah, talashi, ma jibund jibundia, you manu to jambua, several display , saikudanati if any more thing must. You have committed applications given you a machango skaputa, you yuke the arcota machina if you want to come like this.

Speaker 5: I

Speaker 1: got to say must stop. It to sustor.

Speaker 1: AWS no event. Hi. Eventually not to Kurote , so eventually , no I just did the buttons disconnect. I'm not sure if you can see it. Now come up data disconnect data to stay. Database.

Speaker 5: Hi, database is so discovery about.

Speaker 1: King Oh city star yoga. Hi , are you too shy Yes, he engineer was a command of the coastal SSH to Khan to Kite looked at command on the com

Speaker 5: So you got a kill and the other. Ano, kakanyo sti ano, matatin tatinaxima, masuchatimaga, stuttim feature sinal da sofu anyakticin taskar. Oh

Speaker 1: yeah, you got this. Jango Adaminjo, you saw no candy dosh certimos.

Speaker 5: Uh to ukta put skattimas. Hi, hi, hi.

Speaker 1: Hi. Hi. Oh, the guy. Well, you're probably sort of kumya sega jungle that keepanikum.

Speaker 1: Skowsi, Matra, Majikangarmoti, Totosmo, Sasete

Speaker 5: Italapokanato.

Speaker 1: Hi.

Speaker 5: Hi

Speaker 1: So this AWS take one taking you a shadow. Migration data oh, so no S3 took a karato tekitary to come S3 to my data. It was weak, to contact this. Kito su no votariano jungle storage is to ka odska. Mama

Speaker 5: touched pictures, sky cut that to comes, thirdness, my jungle scar. Makon migration kick , I don't say. Keep both because it's not the way of things.

Speaker 1: What is this name? Oh for moneticul.

Speaker 5: So there's a ki hong ima system to go None day what it do at the Rasmidi are in the hook over free. The

Speaker 1: sonar to the sonoji or motor system. Yes, why you guys are kaisha de jangulas karte.

Speaker 5: Maskamu um migration stemu Khanalic Ikiganalakanisona.

Speaker 1: And

Speaker 5: you know, project in that.

Speaker 1: Migration te kiko quai hat Surtoquite Saichokrajujanai disco

Speaker 5: This

Speaker 1: soon the coachikan kakatari sortum, this kiddle more token in kaas you could study to go.

Speaker 5: So the same month studies kicking with the state. Uh can sit of it, massive machine water paging list, but you're open up.

Speaker 1: So this is your day. Hi.

Speaker 5: So this is your

Speaker 1: day. Oh, so no. The topic is supposed to go in on this.

Speaker 5: So listen , use us to put you to sense some back in the She needs to cut it, mother annual, mother, yakutukha, jakman chuttuka, masu tatukomanot, mother, tai jugunan designer.

Speaker 1: Ai, ai tuu saimas. Tuuonassa su potiskaani. Ai, ai. So this nature one third one or two must not

Speaker 5: be a

Speaker 1: Hi, I think I must start. They were quite a studio to the testmaster.

Speaker 5: Hi, I'd be mustard.

Speaker 1: Hi

Speaker 6: Hi.

Speaker 1: Hi. It's giga e to kolore karatsgiga e delukes under the kilojango testov. Nisen Nijiko, toyunu, Hajimetai, to moi masu Roshku Nega ishmas ai. I thought you must Hajj mitte good as I.

Speaker 6: Hi, Hajimitic Titan my mouse. Your skunagash mass.

Speaker 1: Hi, Hajj Dai Jobis.

Speaker 6: Hi. To the Kiru Jango de Test to Ay As we take it to my mask what they're saying, what topic on the mute must have finished in it to walk you or not this Aye. Although my daddy ought to pro that we probably should go we probably should

Speaker 6: program to consider it all in Basil. Kore kar? Ah, hi. Eto. There you go.

Speaker 6: They could come in a sang compass they are taddy, Pi Q theatry, adopt translative system castles of the cars or sustained mass. Hi. Timokteki Disney. The Tamanin is the skiddle. And the Kuno Shiru have to go to Dive Mo Ziru to Hot De City. Uf kui me te apete nato mutte, keo hapsaste, it's

Speaker 6: a takimasta. Atosin e-cogentate, testokite, mitanekazi, marinagestrachtu, mitanako tu e stodimas Kion mokyoto still this near hold on the contain dika. It's not helpful not away you can automok the massive On the order in this program. So yokbua, test to Kakasiman

Speaker 6: stadium or de Sunkumite et to kill you sus data diet. Kukana teorimas hai. This to Kudul has your Solomon washima saying. What do you say? How do the party in Iberride the Christmas in a jungle project once the Hanasotomotemas So at particularly together, to jung on a bajum of kusu e kin testostery to ka. So it hanashim high ticket to toks on Hanash within a giganakata is in the skill moi, Molikunda in the end of the Sunhewtai show guide to sustain the side.

Speaker 6: Hi, it is on the show this year you need to test on Holocaust, you need to test the train of coconut cross test, to get so you are good. But to go test the kids, you go test on Hanasimasimus and Sikino test on Hasimukimasimas. The unit testing and just so why to share kakuny putuana could be half planning or hazimar kuttop. Test to code jitaign, Aquarius , my content you do, Chanta Kita or in Utiku, Kaku in the Kirk, Okinaw Ktiana to Mottimas.

Speaker 6: Then Mokojes. Hi. The PyTest no harasses. Nisan Jimajan and Twoj mustiness. The PyTest PyTest to the Niki library skill. Now you library the attack isn't right. You know, can't you desk? This guy 's in it, that's no kayisham of meteor, that's in a sterile jungle project of mete. Pi just to scatter project in a hobo naika you can 't see that thing with The Python the test to coctaminos, the fact that the factor standard and python not there.

Speaker 6: Jungwan Hyojin no test , kai kuika natoj tikanyo. Not the body to go nicky no got to gas on how to cash coi just run out this yeah. The show sign RMS is the Asato the Mat Python Kekokutsu Kyrukuton. But the hopeful of it could chuck in with the key to the side. Yeah, to such what is it?

Speaker 6: Test of views and test of the test of views by the country, Kite Kotogama, can't let us talk about it. Can you do an production coordination directory in your adaptive test directory that you sort of select the product more Hi. Oh, my data given out of my mass. Hi. Conga con Todeba console testoste, the Baikonakonsa Gardus Must. Dihuda the Uromodate or two date on it They took date object.

Speaker 6: Yes. It important you need to test. You need to test my cut to Hanshia sustainable more. Test to different test to case okay shouldn't misotone So the actual designata is actually sky mask design.

Speaker 6: You need to test a check with the kidum, could the check goes. Cross the case. The Python is the fixed chart. Can 't decorate this name. Target to target to the katasisne.

Speaker 6: So the question they could target diffuse the console mama in your demos. Hi. They saw the sato y code. The case you're says , I'm gonna test the case that they couldn't just say multiple jungle by the motif. Today's giving it.

Speaker 6: The pyrons are not test to guide this. Pyramid to create dantai training. It was the ten tanti test to guide the current skull pies on the test of colour to kinese into the test on it to kinashkur, also smaller can 't be a very strong. And we could test the title module scope of the import s we don't. Modular no scope the importance.

Speaker 6: Class got to buy any. Hi. Parents test guide on the core who you did function under the testometer. So it's a bit new students. A pie testing fixed judge. Take it staggered bar, target to our keyways. Keywordish kiss. We get total timing with the cono import of in portal as it was a very important thing to do. Button case of mirror that the Mozir not to the small import different import the colour cutting import.

Speaker 6: No, ga era narkaz. Hi. Test the case, I shouldn't do the test or to do this. The temple test of passion does it test the wash in your pick more gang of caritziguka. Yonna testokies to mean jamma shoot. You do this. Hi.

Speaker 6: Hi. It code taitukka. Kuramofts unjuni tajun tekna case su gotovo docibunkastai. Kyokoansiken, testukara sakai metuner joking, which antop testocus committee agreed with you, could I not take us? Test of the education test of the kill cai took one security.

Speaker 6: It to noon to Ichiga Kairo kyokai noah, Ichinichibun na to actually kyokai o test of the kid. And on a kairodin, has metin no kairivna, let's go, chanto, yoga, test of case nigga. T must. Hi.

Speaker 6: Asatu got you thought the boy. I saw to go shish, but the boy equal as I mustarimus. None day. Iran in a takaratu, the cocoa naush de motatsing, asatu de Iran notte, motosyo naushte, mechanicele naranai, kanusem ardu. You they kill do then was hakimash. Parameterize test. If Xumon you look put on the duiton test of case or j cross to Python parameterized to you. Cancer, but they couldn't go at Sakurai.

Speaker 6: The Python mark parameterized input test tier. They couldn't era not to parameterized testwah hidrokike waiva terijangula terekashu sanda teritimu, testua ziko sariru naru de sakita, to should she pashish mahu to gana kudariba. When I have this T10 with a sky could again, you need to test to colour.

Speaker 6: Test to you is a king of money about the kissing this. Burugitous de kozin the Burugday who caused the creator bong, the hymn medical tasai. Hi. Hi. Django models DB needs on your test list. I test the mark jungle DB to mark audio she must.

Speaker 6: Maka decorator state sky mass. I testo Maku Jango Dj. You can't cross any test to skate. DV should kikash to give must. Test to reduk to tampus. That tiny eco, eco shit, data height in it to go, or so sabetical.

Speaker 6: On a jango DV no markups got the database colour test to data with a test to case or j go through to you not any mass. Pi test to mark the answer. So this is a mark. Kurason nominitol medokta baywa, jung would decoritaan, mesotonobaya, mesotoniskewa, mesotonita ketekiosarilto.

Speaker 6: You can't use the They could have submitted a Mozurini Kansas. Data is a very good idea. Hi, the test to go through this. PyTest or Jango the PyTest on a PyTest jungle py test library. Jungle setting modules near PyTest.

Speaker 6: And any options they know set test , sorry guys. Python why you don't know Python 50 mm. Test to colour has the model python while Yeah, why test is goes in choice? How many how much to buy the choice to eat and described more default to the test

Speaker 6: to case in test to Prefix what's can I to test a case? XX testometer and I would sketish. The kids couldn't test the conjugation about this. No more review wages matters.

Speaker 6: Jungle South Party and Debra and Test to the T. South party and test system. The test of the show is a component or component. You think that the surrounding component to test to take away to test on the mock take this. Can't this images.

Speaker 6: By test of chart and more fixed chart. No, my unit test. And my short auto show you know who k mo is set up a mess of the teletyada or mesot. Target to age on this kid more, the buttoketo.

Speaker 6: It to test the data. It also matters, it's a good ide No, the guide around Pyrons are Safnosukede, Helipomesu. Hi, you don't think it is to case on the fix chao is this for a son of stereotypti nagarega? Hi. Good to case this.

Speaker 6: Et to test to auto genius, it possum de destroyed. Oni wat tiktokiniwa. No day, colour test the case and the boy, good do data and a pickish channel code of data shakes on the stain there, sake takes. Hi.

Speaker 6: Library. Jang Winganim is a skill coming at the engine. Just my gender taking a model that I'm Mary Roska night. Mammy to Ammaroshi Rosknai to you. Factory boy modes. Factory boy. So no , PySamanga Faktori. Oh , this.

Speaker 6: That you have attached to buy that Dukonitzkukata Kimotinin is getting more. Test tool directory in Shani Factory is the crosskup import of factory TV minor joy to come to the sub factory team was there, account test factory, use a factory methist , to no factory no Usa data, Itajota idea, test

Speaker 6: to the kibos. Create about the king of books to say. Octave how many points. To us But you fragile test. Tata

Speaker 6: test to display, get to display price, middle cancel. Target and still go to more. Hi A price.

Speaker 6: Kiba Doshmount. You can't. Then you have data out scroll to kinyi e do you data nanukat. Oh hi. Hi. They took me factory point factory one how many point to you.

Speaker 6: Skruta manywa who ga factory goshio, fugamoder , fugammodel is the moderativa test to go to kite imasta. Oktori wonava, subfacteria, she can so catsiv surputes, kiss the model take the kirony step to koyku to

Speaker 6: And the fact that we don't die in the sky got to watch. Hi. Tikini mokudes. But test of my test to the unit testing cousin, she eyed the magic kosher. So many design API and it's ridiculous.

Speaker 6: The testosterone , it's a very good idea. The dummy object , object to this. The cancer test to tighten, testosterone, testostery wait, cancer.

Speaker 6: If most you are in this card, so you have to figure out the same thing. Mukato yuna magnificken wata saraju de bendi destru. The karakutax incredible price can do where I don't go to the get mokongae.

Speaker 6: Target to cancel your genius. Expect it to the kitchen iniquotes are test to go in a mass. And to concept the strip, Kiroka Kenshokano is there. As the code with item. But you can 't case.

Speaker 6: You call it this. Important spam they chocolate spam module says spam cara import safe spam import safe spam does Idiot Atara night, you can't do nothing.

Speaker 6: No day, you want to get testosterone. Okay, cohabare the paths got the goods got a kill to skit last time. But you can must pamut the parin spammuzy rutti, no Yeah, spam was you not going to say but spam with it. Mä oon tekin nukaan jo ikanalta. Kosin mukkunatka jatkaa, että

Speaker 6: toisen kun arnut on kijelmitata ikanalta. But the food is done to type on the sorry test to test the cotocodic, so But the HTTP request on the bio responsive driver yes or no, the question response is taken to driver. But our data is in the bio hate could read this.

Speaker 6: Tada, canonukagi can't get this most pirate. Fikusha Mukino Akuma the Damina , Kanuna Khadi can't kiss the ways. Mukutskai Sikh the Gavin Nanutish and way, you go a carnival to go to UK's garden of there, Mukutskai Skinwalking Sky Sikkim in O shut in a chance of the The colour cobblestung, and the code cabalists, testosterone, to the crypto still got case of colour cabalitz, could you test the cry technological?

Speaker 6: Me on a kudoga, naikatuka. You know , idiot tamino, she would have de mass Tamini Kabaret, can't the Nizetani Totat from Nakodom is gotash day Nightarios Mazga, Kodu Kabaret, Midnish, you could have hacked in the kid. Case of Kusurkato Yunuga, Kukala Kugu the Kabaritchini Narinas.

Speaker 6: Motum to coverage library garden discussions. The jungle discovering you are migrations to things to come. Okay, covered the case of questions. Hi. They cabal to nanny medium got with the passing being the same. When we cover the body , optional task data more content by the mid pioneer.

Speaker 6: Punkimura Madikesoktakareba Kabeburanchi. Taishwashwutte testoste ma kabajas uhat shamatrikiamo su test. Digipast. Unajoni coloji, zinka shpahtuko , divig and the day, says only shpa

Speaker 6: statues, the coast. They test the case of some but the test automatically, motherfucker , option. The coverage of document to cut you on a discussion. Coca-Canai to use the can't see this. The code system and the system discovered to meet, the canoe de factoring nothing else is good out.

Speaker 6: Hi. Kabalachikach on under testoga who shouldn't code miscarriage. Joo Kutomaga, otaisin kulun jo hiviit kanaton kanssa. Kentupeikka maatsin faudraa sottotteessa. Hai! Ai. Kukma de Matomedis.

Speaker 6: Fast You can do my testosteristicano , Arigata Barka. You catch this. Hi

Speaker 1: Totikanga mussun arino esti geoto jutru juttu asti tätä kirappaatuivas. Kiko

Speaker 6: Naka, test to case you a zimbu testuk, what's the metana make outsinu day?

Speaker 1: Mutuali What does she have to look in the shiran show? Oh

Speaker 6: , that's my test today. Mita Koto and the point is that they killed Janait kanat de futomo.

Speaker 1: Yeah, it's all this thing.

Speaker 6: Hi, Python test by test random show cai shooting must shoot the bug in co bug with chatter in the skill more. The test test to kess doshiga is going at Tarituka. No de testosure sus state and bar, Iranio

Speaker 6: Narana has not says, has me to kiss , Yes.

Speaker 1: Hi, I got to go through my

Speaker 6: tools I must have.

Speaker 7: Oi. In so no security approach to them, they want to take a document. Ya te must.

Speaker 7: A kumakai sit menus radio taki upass keepish must the shit wakesh they must not kill the command, and we couldn't desk mari for no sokanatom, they must for your show could say that the money is a very good idea. A hundred Hana's coton and the ski domain at the website, Kanoyona, Yo Ken Garabai no, a jangodeno, you za ninjono se care pro to ni tic de Hanaosimas.

Speaker 7: Stop stop The Korean Sabisino to Keto Ste and the Usar Shibitzaga , that we zoo and by doctor, Hambay Shakti and Mazi de Shohino Torukoka, the economy. The koyo yokengade to kino jungle de no use anisho no se key approaching to take your anas type to mimust. The target of the same django, you can use a mic kanatu mute, you must cathete.

Speaker 7: The hung up yon mukuti to stewama jango, then of course to use a shiveth no n show not to nicimata de so ka de reputant to kayak. Matamate girl, yes, not the motivation. But the Hana sanites, the Yazhivatina, Microsoft Hana Shiva Kiwa, Shimasen, Monorisiga, then I'm not sure. The Madiva Matante was in Chenga Nikara.

Speaker 7: Majiga kidoni kidoni no kidman of sang or in there, mafe tekini, monolisos and take for you, service king to the other. I to ushinaga kinan deskido. Yompatans, te sai, then matom it. The mother, my webcast, panhanasto, defective, can write a potential

Speaker 7: thing, you three can write it or more. Kusseyuza sveta nincsono, fikeno so danga muchumaru naatti kotoninanga, kizuki masta. There's a start. Saavissotte, kuna jukotskalked , amateagit kan laittaa naina, mutta

Speaker 7: sydavett era koturaat , maakete. Jos te saavissa pokahuita, te jo se koija koodimasta. The cono matekikara vacuukotto sta visiting. The kore ika to ipotina ma sabisno te kieosade tahottina savisno ketete ika rate kanato umai masu. Nande savisutina kotovano tegikaste keosato ketega ma iranjana i katte umoimasta.

Speaker 7: And they a ma because the user shivetun is no word button. If you can see the other thing, you can see the other thing. The Django team is a very important thing to do. Shingo Yozano Ninja Kantani tis so the kiru in the skeddo, Yakunjana Sorega Cho. Kukaizas tokatikrto, se jak kuni munaru tikotoga.

Speaker 7: Are the mass nande? The cocoa chat jangodono se keep patan no e shokai ni haite iki mas. She must a type of feed of path, n show moder kill your path, ni show moder, kirica patan, the my patin thing. The mawatasino, masankomangyong keke na skedo, manito yon no se ked skuraleta humbanka do ster systemka, ali

Speaker 7: masu. Naka samper kodwede naka di signor service. Ma minimum no is site to come by shot.

Speaker 7: The mother comegaito type feed of the pattern and the skeleton. A type of feed is a pattern, magaio, tosteva, ma user model customize the type of feed or take. The type of feed on it, use the SMAF. The cocoa user should not feel the more sweet te user modell y e te shimas. It is modell no the sole e itemitomeske do ma user modello customize the cutoff to use or ma cash to usergi.

Speaker 7: Pas ansarete, itai cast te hostein te cao, and it's a good canato , oymass. The kukarva production development, this is a patent stage , the mass , model you patanticara suka is mass. Eto Kueno Gaionandeske do Eto O C moderni show model or Statisho. Use a moderator that you use.

Speaker 7: It authentication to your table or guy key, so you set the cocoa and canin show up. They must think I think you can't talk to you much to call the source the muscle there, Madukasimitore, Majesty source the hogar, you know, this de faconic source the country.

Speaker 7: The modern story to modern authentication, and the so basically, you can see that the other one is a very important thing.

Speaker 7: That there is an authentication to this car. The soit taba ni naka ko shopoyuza no in a meme shelto naka. Matt is to maintain the mass. Of the structures that use that to a cash of country and winning the musto jungle water takes

Speaker 7: a mistoga, not the massive There's not too authentication. Modeling edit pasado , I've seen that there are stuff to name the correct. Mother margin to model to custom moderos in Peronye or Tegi

Speaker 7: Simus. The Konotuki Niao Chanto Kono Juan Bango to Koriskedo. Io ci vediamo in tocco in un filo, non moderne, tecke nuga, type fildo, poi tanto questo pochino meritoni na rimasto. The auto ma authentication gani ko user shivet so motasito kto bendi, desculato that's quite must be a Hi, it's a kokumadedec to moderiotegi standards gay sinki kain to

Speaker 7: kino this so simas The Hambai Shagato Kukas rather than use an orbit no URL. Yeah, authentication successful authentication water machanto success. The head to amigini kittens in a photo post the user authentication. Humbai Shag the Sodas in a Uzama, Pony Sarizakatoro could give him the rock star commodity to you, put together a bar and a perfect ste Ai, peraí, você vai

Speaker 7: moverse. Hi , it's so this. Jungwonkoski document on ka logoin distance katamitan authenticato te cancerto, logoin te cancer, mason tori ni just so that you can use the same thing. Titsgini Ano Psion Nandeskin Majatawa to the set, the Kozimi Dorwano this to Kinoka Adi Master Kunande and so it's away in the skedoma.

Speaker 7: Marchant by the Marchant, the Marchanton business logic to category to cast, or you're going to follow my cycle to another in this. Dio in Watarmani, Chota-Sonjšte, ittosio yoku Diktarajnatinva a tequizimidorweldi You carawa, or the question to customer or to sanction the gades, kid, and the other thing.

Speaker 7: Ai, te con successive. The corona crash view the cancer view is the correct customer. Yeah, sucky dokimit away, just so driver, more chantal customer

Speaker 7: needs to make the something checked. Aguerto А ну, другуинсти, амбайчашка, а кто-то на йони таки массаж.

Speaker 7: The authentication model is a katahono yoda si vetuno, ninsuhskila, kai taiva and kumarunza kamita na gimonga de kamo si demasi.

Speaker 7: Ton kick out of Kastama Baidatara. Use the ashibitan katahoga , kanu ni adi massive.

Speaker 7: The patan ma medito de medito , kanla ski do ma medito bastila, jango no, niso kino, mafunitskai masa, the yuno. Magij tekinifkus shalletun you though, you gangster like the naivete kinwish totunno, and so you have the cara, jangono, later not perino de maanshin, kanato umoymimas. Mato customized this mixing on the bounds. The medicine of carnacet on the kats.

Speaker 7: Presere the set of username to passado nanka. Say a kuzeo, nakaye authentication tever, choy us the sickinazar or name at the program ,

Speaker 7: Ai, tem some modelo kyo y patanamato steva, mago ni up tete, ansi de, son na chimite kina ketinwa na nine de, magitio tiki de, you should not. Dovadisho casi a proxy categories, the proxy modello. Ajattelista, sopimata, että ruotsin onde, kun oletamo miten, huono happiu, too, kyllä omoi mass. Ai!

Speaker 7: Mit skedó ninja moder, kirica i patanti nutrini shukai smath The Kuna Patanoke can go nine there, or to POC, mahana s nine death. Is it to modern you at the modernization?

Speaker 7: Now, the good is the user take must.

Speaker 7: If you have a lot of the other things, you can see that the other one is a good idea.

Speaker 7: Kan'kiota customer machanto or se usa if we must be a The Kuraga, Naka Hombank, the Kado, so it took a do you imagine your hands, should even Sabatuka container Yoishte, Modomain, Never de Vakita Dukka, Fruki Sabha Denaka, based passed. And authentication to machantoni lokale nib. Nin show model kyoyo patan yelmo a simple niggist to the kitani.

Speaker 7: Machant to skut till the get the kanji. The center to ste ma ifanteh to jungle naperiva, naktio, they will get the hoga, it

Speaker 7: 's a matu ali masu, to coreano. Kun no marchanton no view data, it's just the product view to you, and so hint to marchantoska and access the king of view

Speaker 7: the key is there. Yeah, customer and custom access to the kinetic deck or standing other and be got Don't know you that's it , no cocoa, docarium.

Speaker 7: Massayak, so score the colour tight should not use the shivitz of car in this. Janwa de Hammetikirutin Giooni Natte. Tigini, to testuno hana si o si mass. But I'm not sure if you can see that the other one is a very good idea.

Speaker 7: Python as a py test of the filoto and machante pip the co sterecara or says the modeli and machanto height Because you know more majority, then most of the model no, we can't use the same thing. Mazvetina testoni over the settings of schedule and the sketch.

Speaker 7: The coconut canon, the testosicovite command of the seminar schedule, but the merit of us, my guide the melee the correct of the Malachi Kun on customata, custama IPP ,

Speaker 7: kun tässä kun kastamaan haikata kenysioku. Testova ano. Directory , or the kineticino, jungle, or kernel you have to use the same thing.

Speaker 7: But if a decorator goes to hova, and tie hand none of that, and you know why so now to you know what you can just The it 's a migration , customer IPP, Python manager migrate ,

Speaker 7: The Honto Kostakata has composed, Adam MPP, the Python managed by migrate to stay to the modern, or series of migration kicking. The Nanderson, Ico minus customization. Adimas, the Kunaion each to Kuno eto, Nakini Kilika Patanato, eto, donokankyo demain jiko siny metabaiga adimassi. Hi , this is the cutomita.

Speaker 7: They came up meditating. Ninja model customized resource video. The table is custom can be doctorities as a machine.

Speaker 7: He also just sold a relative to the kiss animate. No, Django Gaga so test it or no kanato mimasta. They also use a model, I the customize the use of the model, you show module, to stano ni sortaminomono the source code, they can't kill you, use the okay carriage automatically, and use case so they document ,

Speaker 7: Django doesn't have knowledge to use as the Namade, means a modern kid cypatano, Mamma Tometostava. I can mean some mother kill you patano in keto The jango gaga so tested or seized the modern sky cotton hunched a techukto naka coaching taking a hakogimina patan and the other.

Speaker 7: Это джимаджи сабатан, гайондеске до джаго контроль воссон со коду санкуниште, не шокину, джимаджи со. Yeah, and the most kind of jungle means you would sky must.

Speaker 7: Jungle got take your still in shokino cancer, but they must decide not the so you need to more than the other. The Hompatan is so. Jung will control that moderate so it's not so called them with even the imaginate. The ma medi to connect one in so much. Ma jango in so kino, security

Speaker 7: yokim, ma tayo de kirutoka, marukama sele naikido, you mata korota swa si voto de soyokini, de atta kutoka ali mafem. The majority of the cote and cousin, taking I can touch the raw mouth , came of my time, I cannot cause an jungle. In show the cousin is so shining you thin or scrapbooks in kitchen for e canato. The matter should kick a feds and the sketch, the jungle controls, some conservative tricked once colour, you'll meet

Speaker 7: the master. The son of the jungle controls the security task. The unfair to control security particular kinetic or the kid is you with the kinaiboyga or the mass. The motor projects doesn't content on the skill.

Speaker 7: So can we show this register that could be signal to your head when you can up demo? Django sikko nämä sumatia mari, jotka mitään uskotte masta.

Speaker 7: Mustang when you took to Mazda, Ninja Mother Course, the Ocono. The Ninja Mother Kilika Patan Watchto, Hakka Green Patan Cosin Takina Kansitanka, me at an discourse or the quite dinato. The U to Ama so kommode is call cause motivation. The sign when you know this , I shouldn't wait to understand, good is not case to

Speaker 7: use the ancient girl on the tokenic , got to my talk or sino. Matokyomode for Nopatanskat Taiotino, Sure the Kantamerito de Melitoka, Matov Kanyiko Tamar Sotoka, soy Noka at the Razi Nanka, feedback to feedback , and the

Speaker 1: Tomaza jutano komentę z ust mi temi mas. And

Speaker 7: Doga no patan de show questions of car with potion and there.

Speaker 1: Hi.

Speaker 7: Hi.

Speaker 1: If you can see the same thing, you can use the same thing. Hi, I'm not going to be able to do that.

Speaker 1: Hi, but you can this note.

Speaker 1: I mean that hey to Sri Gaussian Holy Nature. Okay. Restart your presentation. Your scrolling ice mass Thank you.

Speaker 8: So Kwani Chivan Namaste, hello. I am Apur. I am from India. I will be presenting a talk on Django Congress Japan 2025. So the topic that I have chosen is getting knowledge from the Django heads using Grafana and Prometheus. So as a developer, we all have built so many large applications, but the real work comes in when we have to maintain it, we have to monitor it. and how to manage those things. So that's where Grafana and Prometheus tools like that come in, where we are keeping the monitoring tools and we are also observing the apps and we are trying to figure out what went wrong in some problems. in Samba. So moving on to the next slide, we will be talking about the introduction, Prometheus, Grafana, and how to integrate that in Django.

Speaker 8: So before dive in, I would like to tell you about myself. I'm currently a software developer at Pells Fargo and I'm also a technical steering committee member at JD Robot. It's an open source organization where we develop robotics development using Django React ROS and we have I have been participating in Google Summer of Code as a mentor for JD Robot for 22 uh for for the year 23 and 24 and I pass participated for the first time in year 22, which has been my tremendous experiences there. I have learned a lot and that's how I got into the open source. And I have been learn I have learned machine learning from Amazon experts using the program that they launched called Amazon Machine Learning Summer School And there are several open source programs in India. One of them is Code for Government Tech, where we are building large-scale application that would be used for

Speaker 8: government government of India that I participated as a participant in 2023 and I have learned my development experience from startups and companies like Deloitte, Resilientia and Top Dog. So that's all about my software development and uh software development and career. But beyond that, when I'm not coding, I'm a very big fan of animes So my talk, which is on Prometheus and Grifana, is uh out of uh one of my favorite character, Itachy Ujiha, who emphasized on looking ahead and always be prepared. So let's get into Django and how we can foresight and prepare for our Django application using Prometheus and Grafana. Thank you.

Speaker 8: So I've came up with an analogy to you know walk through this whole talk. So it would be like getting to uh getting to know these tools like Prometheus and Grafana because they are big names but uh when we try to uh when we try to think if think of them with some analogy, it helps us understanding what's the basic crux of it. So I like baking. So I have been into cakes and pastries. And once like, you know, there is an anxiety comes up whenever we are trying to cook a cake, bake a cake, sorry, when we are trying to bake a cake We have this constant anxiety like what if something then goes wrong? Is everything running smoothly? What if the over overheat? So how do you do that? You will keep observing, you will keep monitoring your bakery, and that's how you will find the results.

Speaker 8: Okay, okay, everything is going fine or if something is going wrong. So same way we have to do it for our Django application. So let's have a look at one of the tools which is Prometheus. So here it's an overview of Prometheus and it's an overview of Prometheus where we can see at its core we have At its core we have retrieval system, time series database and HTTP server. So uh time series uh these three are the main tools that Prometheus is using to monitor the applications. So we can see that it's using pull metric system mechanism to take all the metrics and it scrape it from the target endpoint to get the knowledge of what's happening in the system.

Speaker 8: So if there's an application running, it would be putting all the data somewhere at the endpoint, uh, which is generally the default one is slash metrics, and Prometheus server would be using its retrieval method to Creep that data and store it in the time series database in some loc in some disk storage and the HTTP server uses PromQL to represent those time series databases and analyze it in the manner of graph. So here we can see uh the pull mechanism but for the uh let's suppose we are scrapping the system at an interval of 15 seconds But we have a short-lived job which is running just for the five seconds and it will then exit. So how will you scrape such data?

Speaker 8: So there's a uh component that they have created which is called push gateway. The short-lived jobs would be pushing the metrics at exit. And and we will be scraping the data from these push gateways. And there can be certain scenarios, like if you want to monitor the Linux system where you are running your code, your application code, and you want to uh monitor your Linux system then how will you do that for that also they have exporters uh like node exporter is there which will be providing you the pull metrics according to what Prometheus scrape. So the formatting would be help would be done with the help of such Prometheus exporters. And we have service discovery. This can be easily integrated with Kubernetes AWS and other services with the help of service discovery component.

Speaker 8: Now okay, you have monitor and you have figured out what went wrong with your application and you need to alert your uh team like something has gone wrong. So how will you do that? With the help of alert manager we can send an alert using email, maybe Slack, and notify the uh Notify the team where 's the fix that is required and with the help of data visualization tools like Grafana, Prometheus Web UI, and also it can be integrated uh a in customizable dashboards using the API clients. So if I talk about the traditional monitoring systems, there is generally the push mechanism. Like you will go, uh the system will push the data and there can be chances that the data got missed.

Speaker 8: And you won't get the latest updates on the system. But Prometheus relies on the pull mechanism. So it will be scraping all the data, hence it's more reliable. So like uh like if I talk about Prometheus in a straightforward manner, it's like uh It's like a data automated automated automated data collector which would be collecting data from the particular targets and it would be in a it would be It would be putting it in, you know, it would be uh using the vital science of the application and telling us what what is wrong. And with this we can identify some uh issues before it becomes into a bigger problem So in Django, what the things that we would like to track?

Speaker 8: It may be the server response time, number of requests, and how much memory you are using. And how do you want to identify some SQL queries also if it is running for a long time? So, as discussed in the overview of Prometheus, there are the these are the key features of Prometheus, a multidimensional data model, which is a time series database, and it can be uh And the dimensions can be increased into multiple dimensions with the help of metric name and key value pair. So the labels are very important to distinguish the distinguish the metrics and it would definitely help in identify bugs easily. Then we have this PromQL which is a query language

Speaker 8: to to leverage the data that we have scraped from the target endpoints. And then we have time series collection that is happening over pull model. And for short-lived jobs, we have intermediary gateways and target discoveries are there for Kubernetes, AWS, and several other services. Now, I would like to go to some prompt uh promptual examples. Like if you want to identify the number of HTTP request errors which with the status quo of 500, you can simply write some and in the bracket HTTP request total and status. equals to 500. So this will give all the all the number of HTTP errors that have happened in our system.

Speaker 8: But if what if you want just want it in the past 30 minutes So you can add uh square parenthesis, square bracket sorry, and put 30 minutes and colon. Then it will tell you total number of http requests like this. You can add 30 minutes And now if you close it, now you can get all the past 30 minutes HTTP errors. So with the help of PromQL you can easily visualize things in better manner Then if you want to have some rate in the five-minute rate of the HTTP request. So that is also there with the help of a rate keyword. Now, I would like to ask, like, what is the benefit of having this total number of HTTP

Speaker 8: 500 error We can have alerts which can tell us the sudden increase in server failures. And for this particular query, what would be the use case if we have a normal traffic which is coming around 10 requests per second and suddenly suddenly someone tries to DDoS attack or maybe the bot attack. This number would go up and it would be like 500 requests per second. And we can have alert setups already there which would help us identify that yes there is some attack or it can also be a sign that something went trending on your application

Speaker 8: Like uh maybe on some application there's some video that for trending and you can also know that with the help of this But what if this rate doesn't really stop to drop to zero? Then you can actually know that there is some server failure. You are not getting any request, so it might be a possible server crash. So Prometheus keeps an eye on all the activities and tells what's going on every few seconds So we have all the data, but what if the data you are seeing it's not good until and unless you try to draw inference Like we were seeing the examples of how the data was set. Like if there's 10 requests per second, we are saying it's normal traffic, but if it

Speaker 8: suddenly spiked to 500 requests per second, we are afraid it might be some DDoT DDoS attack or bot attack or something just went uh viral on our we on our platform so this data If we are visualizing it, we will be draw more influence uh inferences from it and we can have the data uh data in more uh like In more specific manner you can track a memory leakage or or server failures. Such kind of things would be there if you visualize your data So I have I've added one slide. This is uh this is upon the memory leakage, like how to identify memory leakage. if happening in your system. So the sudden spike in memory

Speaker 8: uh here the sudden spike in memory is there and there's the green Sorry and there is this green uh which is telling the trend which is normally there and if you see the sudden spike might might say yes there's a memory leak And if even if you see the corresponding CP usage, that is going high. So that could be also telling us about some of the um Some of the high system usage, or maybe there's that's the time when usually when people are using your app. So that that kind of information can be drawn from uh visualization easily So with this uh we are able to uh keep our bakery thriving.

Speaker 8: We have all the things that we know and we can also fix things before our customer even notice. So the main part comes in when we have to integrate Prometheus and Grafana in our Django application. So It's the architecture that I have taken where we are seeing our EPI server is running. Our EPI server is running and Prometheus would be using a pull metrics and it would be pulling all the uh data and it would be generally from a metric endpoint oh so sorry Metric endpoint, it would be trying to scrape that data and it would be putting it in the storage, which is a time series database.

Speaker 8: And Grafana is not directly connected with any of our API servers or client server. It would be you integrated with the Prometheus where it would be using the PromQL to create the beautiful dashboard. For the example, what I will be doing is I will be running the API server and client server on my local machine, but the Prometheus and Grafana would be running inside Docker. So we can see how you can uh so this this can be uh simply seen for different OS also So the real question comes in how to integrate Prometheus in Django. So I will be using Docker. So first you have to create a Prometheus varmil file which would be telling what are the target points, what is the scrape interval, and

Speaker 8: and uh what how many jobs we will be uh basically monitoring with the help of Prometheus Then we will be creating a Docker compose file which would be running our two applications. One is Prometheus and the other another one would be uh another one would be Grafana. We are not running uh we are not running our client We are not running a Django server inside the Docker. We will be running it outside. Outside the Docker. We can run it inside, but for this example, we have taken it outside. And there is a Prometheus client that has been developed by Prometheus. It's an open source tool where you can actually go and look what they have developed for the Python version of Prometheus and how to integrate it. And that's all you have to done to integrate Prometheus in Django

Speaker 8: via Docker. Now if I talk about how to integrate Grafana You will just go to Grafana Dalgoth and add a new data source. But the real question comes in, what would be the Prometheus server URL? Because we are running it inside Docker. So what will come here? We'll answer it in the next slide. So this is the Prometheus YML file. This is the step one. This is step one. Where we can see we have global variable which is scrape interval. So it will be telling us How many uh at what intervals we will be scraping our Django application? So currently if you focus at the image here, we can see at the timestamp 95824, it uh it took the first uh it took It scraped it for the first time, the metrics endpoint.

Speaker 8: And then at 9. 58. 39, that's the in difference of 15 seconds later, it scraped again and got the data from from our uh from our application. There is this. uh this particular sorry this particular you uh target docker. for. mac dot localhost It's basically localhost 8000. But since I'm running my application outside the Docker, I have to put the target point endpoint like this. But if my application was running inside the Docker, I would have simply used localhost 8000. And here we can see remote write. We are uh there's an there's an authentication set up on the Grafana dashboard.

Speaker 8: So it would be running in in localhost and for that I have just put the basic auth as admin and admin. Now, this is the main crux of app services that we want to run, which is Docker Compose. And here we can see we have created two services. First service is Prometheus, and another service is Grafana. So coming to this, we are using the Prometheus image, which is available on the Docker Hub, and we'd be using the latest version of it. We would be naming our container as Prometheus. And then we would be running and mapping it to the port 9090. This is the default port that is generally used by Prometheus. So have kept it like this We will be mounting uh the volume to keep the data persistent. Even if we restart the server, we'll have the data running.

Speaker 8: We'll have the We'll have the data and we don't lose any data. So there is this network that we have created which is called monitoring. And we'll come to the use case of this And then we have this restart which is in the Docker Compose unless stopped. So unless this ensures that Prometheus restart automatically, unless it is manually stopped or explicitly stopped. Now coming to Grafana, all the things would remain the same. There is this environment setup. As I told you, Grafana has uh authentication, so I'm just trying to bypass it with passing the password as admin, and this depends on. And if you look at this depend zone, it tells that Prometheus server read

Speaker 8: the Prometheus server, the container needs to run first before running the Grafana server, as it depends upon the Prometheus server And again we have defined the net monitoring network. So what happens is uh in the clo in the Docker network we will be running Uh the container of Prometheus inside the same network. If I would have not defined network, then what would have happened is Prometheus would be running in some Docker network And Grafana would be running in some Docker network and they would not be able to communicate with each other. And since we wanted to put what connection URL would come here, it won't identify. Even if you put localhost 9090, it won't recognize it Because it is not running inside that particular network.

Speaker 8: So now we are using this bridge which connects uh our networks between Prometheus and Grafana. But the question still remains, what will come here? It would would would it be local 9090? No, it won't be that. The reason would be because We are using network in in the Docker and we have to identify with the uh ser with a service name. So it would be Prometheus. uh colon 9090 because when you are when we are divide uh creating a docker network we use the container uh we use the service name prometheus and grifana to connect with the to connect with the ports that are running inside that particular container. So

Speaker 8: let's look at some changes that we have to do with our doc uh with our Django application So far, we didn't modify any of our application code. We have done Prometheus , we have created one file, Prometheus YML file, where we have set the scrape interval and what jobs we will be scraping. And then we have created a Docker compose file. It's not necessary just to create a Docker Compose file, but it will make things easy for us. Because all the things are at one place. You can def uh definitely launch Prometheus at uh one uh uh separately and then Grafana separately and then find a way to connect them in the same network. That is also uh that can also be done but docker compose file is like the docker compose method is the most

Speaker 8: uh most uh like uh easy to use and definitely it's able we are able to understand it so it would be better to use this method Now coming to the what modifications are required in the application code. So this is where we have to put The Prometheus client and we will be creating two metrics. One is counter. So if I talk about the counter metric, so sorry. If I talk about the countermetric, it is generally used when we want to know how many times X thing happened. So if I'm saying HTTP request total, I want to know the number of times HTTP request got uh got uh

Speaker 8: like happened and if I talk about histogram I would uh we generally use histogram metric when we found we want to find how how long things are running for, how big the thing uh that we have run for. So So just if you want to identify how long this particular SQL query is running, you won't be using a counter, you will be using a histogram metric to identify how uh for how long this particular process or how long this particular SQL query ran for. So for that we are using histogram. So we can see as latency we want to know how the question is how long for latency. So we are using histogram. When we want to know the count, we are using counter. So we have created a we have created a decorator

Speaker 8: which is monitor request and it would be used to count uh it would be used To count the request and latency, request count and latency. We can see we have a start time, and then here we would be passing the request and we will be capturing the response and then we will be saying what what was the duration for it with the help of time dot time minus start time. So it would be telling how much time it would take for that response request to complete the process and how in how much time we got the response in And then we will be using these label. These are very necessary to identify quickly what was what's the what's this particular metric is about. So this would definitely help us debugging easier when we are going for

Speaker 8: if we identify any problems And there's the endpoint request. path. So we will be segregating them with the request. path and what is the status quote driven. So when I was showing you some prompt quill you we we saw that we have http request total status 500 like 5x6 So this status code would be passed in the request count dot label. And then we are completing this decorator. Now I have created three endpoints just to uh simulate what we have done. So we have this welcome to Prometheus monitoring and I have put a delay to show the latency and an error request is also there which simulates error and All of them we have put the decorators and we will be seeing what uh what would be happening in the next slide.

Speaker 8: But you can see we have created one more endpoint which is called metrics Whatever metrics we would be generating with the help of Prometheus client, we would be putting it in some endpoint like metric. So for this We have created a HTTP response which generate relating. This is the function present inside the Prometheus client and we would be using it to put all the metrics inside one place And that is all. You just need to identify the put the endpoint in the URL patterns. Seeing metrics, would we be using our metrics function that we created? from the Prometheus client and that's all. And by making a few changes in your application code, you are able to monitor your application with Prometheus and Grafana

Speaker 8: So this was the particular endpoint that I have been talking about in the whole series. So we can see HTTP request body total bytes bucket This all this data that we are seeing currently is the way actually Prometheus scraped from it. So the Prometheus server understands this format. If you would have provided some different format, it won't be able to understand the whole. HTTP thing. So that's why Prometheus have this client's SDK, so which we can easily use and export the data into the required format. Same way for Linux, they have node exporters For AWS also they have uh exporters which will help identify and put the format in the correct manner. And this is one of the query

Speaker 8: HTTP requests total that I am seeing. I simulated it. I have this linear trend. I created a Script to uh show some data how it is going. So it is sending request every 15 uh seconds. Uh I send one request and we can see it here. It also tells us about the target health And this is the graph. So if we look at this Prometheus web UI, we can simply say, like we cannot conclude much of the details from here. But if we look at the data. The same data can be visualized in two different manner and we can and draw inferences differently. So number of requests that were coming is what 0. 133 requests per second. So such way we can use uh tools like this and identify DB query errors, uh

Speaker 8: what is the response time and what is there any uh particular place where we can see identify memory leaks or CPU if high? where that would mean either our particular application is being used by the user at most of the time. So such inferences would be drawn. And uh definitely uh like as I told like uh Prometheus highly depends upon the pull metrics. So it's been Also advised to limit the uses of push gateways. It's always better to rely on the pull metrics But whenever there's a short lived job, there's only one thing that we can do, then we have to use a Then we have to use

Speaker 8: our push gateways and alert manager and then other components would be there. Yeah. That that completes my presentation. I am open for questions.

Speaker 1: Okay, thanks. I'm seeing it in on YouTube comment. Why are you coming in tomorrow? Okay, I have

Speaker 1: some questions. I think um using prometels and grafana um is needed a lot of computer resource. So I uh my question is how can I start observability on production uh easily?

Speaker 8: Oh so uh like uh that uh if you want to like do that, we have uh Like uh we can also monitor Prometheus as well as with uh like Prometheus allow it to monitor itself. So if you are Running it the production side, I would always advise to not run it on the same uh same place. That's why I try to move this particular piece from uh outside of the From our main application code. So even if something on if we are running on the same server and our server goes down, then even our Prometheus might go down. But if we are running it on separate uh server We might uh even if the something because generally it's used for distributed system

Speaker 8: and a distributed system one one uh fails There can be a pro uh the cascading effect like all the other services starts to fail. So there might be a chance if you want to save that. I'd I would advise to put it in some different server and it is tracking all the the things. And there's also one thing Uh when we are doctorizing applications, we really need to be very careful what we are monitoring for. If we put the wrong target URL, we wanted to uh monitor our application like if I would have put localhost 8000 instead of Docker. 4 it would be start uh scraping the data it uh maybe some something there for which is running inside the Docker So such things would be also be uh needed to take care of when we are going for production.

Speaker 8: Like uh we are putting the correct endpoints and to make it easy I'll I'll go with it's better to go with the Docker app. site and uh keep it on the sep separate instance.

Speaker 1: Yeah, yeah that's a question. When I uh when I use the G Unicorn with multiple processes and threads. Is it possible to collect metrics on the power process or pass rate basis? For

Speaker 8: multi-process, right?

Speaker 1: Um process or power to uh metric connect to metric I

Speaker 8: yes. The answer is yes for this. Yes. We have in the settings. py of Django, we can put uh Prometheus Prometheus client have this multiprocess uh feature available. So if we look at the Prometheus client code we can easily uh put in the settings. py and as like multi-process true prometheus slash multi-process equals to true and it will start and we have to identify one uh put one directory also so it divides and segregate the code according to the processes and it will work around that and it can easily monitor the multiple process and thread there.

Speaker 1: And

Speaker 8: there is one more thing I would like to add here. Generally, Prometheus client is uh used uh when you are if you want to uh top use Prometheus according to your needs But there is one more package called Django Prometheus, which is built on the top of Prometheus client, which can uh like what I have done in the code like this process uh here This could also be avoided with the help of that. It is a library which uh which you will use as middleways and it will uh give the metrics Whatever metrics that is defined inside that package would be delivered. So that is also there. Like you don't need to make any application code changes, but I wanted to show the flexibility. So that's why I use the open source which has been officially declared by Prometheus but that is also there and it would be more helpful if someone doesn't want to jump and make a lot of application code changes

Speaker 8: And that would be setting some, it would require only changes in the settings. py and URLs. py. This this would be still required because that we have to expose from our side.

Speaker 1: Can I set up the uh Prometheus and Grohana on Kubernetes?

Speaker 7: Sorry, oh Can

Speaker 1: I set up Prometheus and Cura for my environment on uh Kubernetes?

Speaker 8: Yes. Yes. They have uh service direct discovery uh uh like if you go to Prometheus official website, they have this uh discover target uh where they they have made it accessible across different services and one of them is Kubernetes they have which they have mentioned especially in their diagram also so it is definitely extensible there also And many other services. There's a whole list of it and they keep on adding new things there. So it's it's definitely extensible

Speaker 1: Yeah. Okay, thanks. Hmm, it can go around katani hongodemoda i job this Can we embed embed graph final dashboard into other applications?

Speaker 8: Uh can we embed Grafana into as an application?

Speaker 1: Yeah, yeah. Into other application, Grafana dashboards. Can we embed Grafana dashboards into other applications? Uh

Speaker 8: like uh I haven't done it myself. But I think this would be something uh like good if we are able to like this dashboard this particular dashboard like I have done done it myself but I think Grafana provides such Uh flexibility. I remember seeing it somewhere like it uh it is there, but I have not done it myself But these are very uh like this particular dashboard is also like uh uh some open source uh de developers have created their own dashboard and if someone doesn't want to do that hard work of writing all the prompt QLs It's like customizable and it's already available like like if I created one dashboard and I put it there, people can use that. uh

Speaker 8: ID and simply use the same dashboard also. So I think the extensibility is there, but uh I I don't know how how I have how to do that. But if someone wants to use Prometheus and they have their own API clients, so they can write their own wrappers, which will definitely be a bigger work But uh that is also there. And uh I I also think might l like a lot of people like must be thinking of logs also. Where are the logs? And I there is one more tool which is called Loki that goes very handy with Grafana. So if we go with uh Loki also, then logs would be in picture there.

Speaker 8: That is also one more tool I would say better to use and if I talk about some of the best practices while working with Prometheus and Grafana that would be like In any logs also, in any monitoring, we should also know what to log. Sometimes we have this habit of logging everything and that becomes a big issue later going on forward. So it's always a good thing to log the necessary things And in Prometheus, as I told you, the target endpoints, they're very necessary. The scrape interval is also something that we need to figure out what's the best scrape interval for us. There might be a situation like 15 seconds might be too much. I I need to put it like maybe five to ten seconds. That is also something. We'll need

Speaker 8: really need to focus on what would be that and there are authentications also blocking like if I'm uh using some uh authenticated application I really need to bypass those. So Prometheus also provide in the client side also they have the authentication feature. So we can authenticate with the help like server to server authentication is also there.

Speaker 1: Okay, thanks.

Speaker 8: Like I was thinking like when uh when I'm exposing an endpoint which is showing all my computer's health like on metrics like this. uh I'm showing there but if it's not protected uh might be some some other person like if he tries to scrape this data it would be a big uh it would be a big loss for my application. They are needing all my data. So they that's why we have server-to-server authentication enabled also. That is also yeah

Speaker 1: Okay. So finish this session. Thank you.

Speaker 8: Thank you

Speaker 1: Yeah, those I must have. See you next week.

Speaker 1: They're streaming to YouTube Okay , please start to your presentation.

Speaker 7: Hi, hello everyone.

Speaker 3: Uh I am Priya. So I'm a software development engineer at Beyond IRR. I'm a session organizer at ShangoNod Space and a co-chair at Fundraising Working Group, Chango Software. foundation. You'll also find me wearing red hoodie as a GitHub campus expert or a GitHub octurn. So yes thank you for having me here today

Speaker 7: to have The privilege to speak here at Shango Congress, Japan. Thank you to all the organizers and the audience for joining in. So, before starting with the topic culture eats strategy for breakfast, why psychological safety matters in open source? I would want you to take a moment. Don't worry, you don't have to answer it in the chat boxes. I just want you to think about these questions out loud. Have you ever held back a question because you were afraid it might make you look inexperienced? Or have you ever witnessed a microaggression or unconscious bias?

Speaker 7: Let's say someone treated unfairly in a tech space. Maybe a small comment that made someone feel uncomfortable even if it wasn't intended that way. Or have you seen someone leave a community? Despite being talented. These moments, they happen to all of us. And an open source where conversations are often online, written in public, and visible to everyone, it can feel even harder to speak up. That's why the success of open source projects doesn't solely rest on technical brilliance. it's deeply rooted in our culture. So that would be

Speaker 7: the broader aspect of the agenda for this talk. We'd first cover how to understand why culture is crucial in open source, psychological safety, what it is and why it matters, we'll cover up next. Then we'll dive deep. into the practical ways to build a psychological

Speaker 3: safe space in a community. How

Speaker 7: Django space embody these principles, then we also Get over the reinforcing psychological safety through feedback loops, and ultimately we'll talk about the role of Django working groups in onboarding new blood as well as maintaining old guards. in the community. So uh what does it take or what does it actually mean when we say culture eat strategy for breakfast? I mean you all must have whatever lunch, dinner by now, but yeah, by the phrase, what is culture eat strategy for breakfast? Imagine this. You're leading an open source project with detailed detailed roadmap deadlines and goals

Speaker 7: all right in front of you but if the culture isn't safe If the contributors don't feel heard, if there's fear of criticism, will the roadmap succeed? I suspect strategy is the plan, but culture, culture is what determines if that plan actually works or not.

Speaker 8: This is why companies and open source communities that ignore

Speaker 7: this very important aspect often struggle. People disengage, innovation slows down, and projects lose momentum by time and time. But if we get the culture Right, everything else just falls into place. Now the question is how to cultivate such a thriving culture I'd take up the very beautiful metaphor that is close to my heart when I was reading and going through the Django documentation. Let's refer it back to it. Django's own documentation uses a beautiful, beautiful metaphor. It compares open source to a garden. Yes, a track as the community tended garden.

Speaker 7: And as in any garden, sometimes there are weeds

Speaker 8: to be pulled, and sometimes there are flowers and vegetables that need picking.

Speaker 2: Even in the most

Speaker 1: garden, there are still snails and insects. In a community garden there are also people who with the best of intentions fertilize the weeds and poison the roses sometimes. Sorry, uh you changed uh slide or page of slide. Job of the community as a whole To self uh again, I guess uh there's some problem with the presentation. Let me reshare

Speaker 1: So yes. Okay.

Speaker 8: So I'm really sorry for the technical problem. But yes, as referring with the garden metaphor of the Shango documentation. As in any garden, sometimes, you know, there are weeds as I was mentioning by week can check it out on the slides as well. As in any garden, sometimes there are weeds to be pulled and sometimes there are flowers and vegetables that need

Speaker 7: picking. Even in the most beautiful garden, there are still snails and insects on the go. In a community garden, there are also helpful people who with the best of intentions fertilize the weeds and you know poison the roses. It's the job of the community as a whole

Speaker 8: to self-manage, keep the problems to a minimum

Speaker 3: and educate those community into the community

Speaker 7: so that they can become valuable contributing members. I absolutely love this metaphor. Let me just See, yeah. So the here is there on this on your screen of how this example that I was talking about It reminds us that software is never finished. It's always growing. And just like in any garden, what we choose to nurture matters. Now if we talk about any gardener will tell you that it's not easy, just uh Give me a moment.

Speaker 7: I hope my screen is right there

Speaker 8: Yes, I hope you're seeing the unseen challenges just right now. I'm seeing unseen challenges of uh handling this screen sharing. Uh now any gardener will tell you that it's not easy Easy. If no one pulls the weeds, they take over. If snail starts eating the plants, progress slows down.

Speaker 1: And if the soil isn't rich, plants struggle to grow. Yes, green is okay. Could you re-share your screen? Uh I hope it's good now. Exceeds top A No

Speaker 1: cosloidom.

Speaker 2: Oh. Share it now. Are you able to put?

Speaker 1: Ah, okay. I can see yes, right There's still some oh no, okay. Um if is it is it fine? Is it or Can you please give me a heads up if you're looking at the right screen?

Speaker 1: Yeah, I can see maybe good. Psychological safety is the key to uh sleeping garden.

Speaker 3: Uh is it like streaming right? I'm in doubt that it's it's streaming right or not. Okay Uh I am very sorry for the technical mishap right here. But if we continue with it with what's the what that's why the psychological safety isn't just a nice to have it's a necessity in the community. When we talk about psychological safety, what do we actually mean? That at its core, it's the belief that you can ask questions, share ideas, and admit mistakes. Without fear or embarrassment or retaliation, the sense of safety and willingness to speak up is not just an

Speaker 3: individual trait even though it's sometimes you do feel the and everything and experience at the individual level it's an emergent property of the group

Speaker 7: It's what contributors say, I don't know how this works, or can someone help me with it? Or sometimes it's like I think there's a better way to do it without hesitation. An open source where collaboration happens across time zones and cultures, psychological safety is the foundation. And if people feel safe, they contribute more, take creative risks, and engage meaningfully. But if they don't, silence takes over And when science replaces contribution, projects stagnate. So how do we build a culture of psychological safety in open source This is a group level phenomena.

Speaker 7: It shapes the learning behavior of the group and in turn affects The team performance. Yet taking risks and vulnerability are at odds with what social psychologists say: impression management. And especially in the tech

Speaker 8: space we all are in, there is mostly that instinct that my force PR should create an everlasting impression. It has to be be perfect or what impression would be there if I failed to deep up properly in a code pairing session or what would folks think of me if my PR gets closed, you know, without being merged, etc. etc. If one of the work by Profession Brown dare to leave, she shares, we all use this impression armor to protect ourselves. But what armor is heavy and prevents us from growing, so having the interdependence in the community is important to cultivate a safe spot space and break that barrier. There are three key principles.

Speaker 8: Encouraging curiosity, constructive Feedback and leading by example. When it comes to encouraging creativity, it is that every great open source contributor was once a beginner. But often newcomers hesitate to ask questions because they fear looking inexperienced. We need to actively create an environment where asking questions is not accepted but you know celebrated so a simple shift like saying there's a great question instead of You know, you don't know how to do this or you should know how to do this can make all the difference really. When it comes to constructive feedback over criticism, what we actually

Speaker 8: mean

Speaker 7: is that feedback is essential for growth but the way we deliver it matters. A comment like this is wrong can someone can sometimes shut someone down While I see what you were going for, here's another approach that might help can keep the conversation growing.

Speaker 8: And what I personally love is the leading by example principle Culture isn't dictated, it's demonstrated. If experienced

Speaker 7: contributors openly admit when they don't know something, if maintainers set the tone right for respectful discussions, and if community leaders leaders model empathy, others will follow simply.

Speaker 8: At this point you might be wondering, does it really matter? Does it really create a difference? The answer is a resounding yes.

Speaker 7: Communities that prioritize psychological safety retain contribution. longer, attract diverse talent and create a welcoming atmosphere. But it is high time that we understand Psychological safety isn't just a switch you flip or a policy you enforce.

Speaker 8: It's a result of a community that actively chooses to support, actively chooses To include and actively chooses to uplift each other every single day.

Speaker 7: And in the world of Django, I've seen this work in action, most notably through the Django Not space. If you're wondering what is Janganath Space, let me brief you about it. Janganath Space is an eight-week

Speaker 8: group mentoring program.

Speaker 7: Whether you're just starting out with your journey in Django or looking to deepen your knowledge within the framework or related third-party packages, this program empowers the contributors with an accessible, inclusive, and structured path. Being a Janganaut myself in session one and now having had the opportunity to be an organizer for session three and ongoing session four, I can say this by heart

Speaker 8: that beyond just onboarding contributors, Changarat Spay is also a living example of how psychological safety fuels long-term sustainability in a community and potentially create leadership roles in the future. So if I talk about the program goals of the Changanod space, for those of you who might not be aware of it, it is about increasing sustainable and regular contributions. and community involvement with Django, provide an accessible, actively inclusive space for the development of Django Narts, improve sustainability of Django's development through empowering

Speaker 7: others to progress into Django's leadership. roles. The structure of the program is uh straightforward like the Chang'anauts are the mentees who are passionate contributors of the Chango B

Speaker 8: ecosystem dedicated to improving the framework through code, documentation, or the community involvement. The navigators are the folks who are there to help you find the right tickets discussed the technical details of contributing to Django or sharing helpful tips and tricks, they offer the support Django Not needs to make meaningful contributions and grow faster in the Danger. Changle community. The captains are the community managers steering the ship forward, working closely with navigators to ensure everything else runs smoothly. So in this structure, we support our contributors both technically as well as personally Now the point is how do we achieve our goals within the Changanath

Speaker 8: space? It's through accountability, connection, sponsorship and inclusivity. being our four integral themes which we aim to incorporate while wearing a little space suit. So the first and foremost, our community thrives on accountability And we maintain it through weekly catch-ups which encourage Shanghanauts to share their progress, celebrate it, learn to

Speaker 7: overcome blockers while staying motivated. uh you know

Speaker 8: create a space where we hold ourselves accountable by supporting each other's growth. The next step is the connection by recognizing that behind every line of code is a real person learning, evolving or striving.

Speaker 7: This understanding brings a sense of humanity to our work, creating a community built on support and empathy.

Speaker 8: And for the sponsorship, unlike corporate settings where we are talking about the financial incentivizations, in open source communities, things rely on the goodwill, passion, and expertise. of contributors scattered across the globe. So through sponsorship we shine light on the accomplishments of Django Nots, connecting them with experienced professionals and leaders in Django or Python. We organize talks, encourage folks to get involved with Django newsletters, grants, blog posts, or simply encourage them with a little nudge that I think there's that sparking you to go for that particular conference or let's hop in and do some co-working session together.

Speaker 7: Another thing that we hold very dear is inclusivity. There's an old saying, make the space and they will come. But we believe it takes more than just creating the space. It's about extending an invitation, but we believe it is

Speaker 8: this saying that goes by make the space and invite them in. Active inclusion is far, far more powerful than passive inclusion.

Speaker 7: When we reach out directly with opportunities That aligns with the skills of the Janganots, the motivation to contribute follows naturally. So if this is something that you're excited about, keep an eye out on the socials of the Janganath space Uh so

Speaker 8: what are the key takeaways of of why I was referring to all this? It's about The takeaways that extend far beyond the program itself, that extends within the Django community as a whole Each line of code contributed, each discussion on the forum initiated, each ticket on the track explored, or each talk, podcast, blog shared becomes a Django North. And

Speaker 7: like we are here today

Speaker 8: talking about Django and being involved with the community, these knots aren't just lines of code, but they are threads of connection binding us together as a community. and then contributors to ensure the framework sustainability. Here no one is expected to you know have answers to all the questions in the whole universe. In fact Mentors themselves learn alongside mentees. This desta stigmatizes not knowing something, normalizing, admitting gaps in knowledge and encouraging

Speaker 7: a culture where asking questions is seen as a strength and not a weakness. If maintainers openly say I haven't figured this part out yet. Let's experiment it together. It rewires the culture.

Speaker 8: Creating Slack threads or GitHub discussions titled Things We Are Still Figuring Out.

Speaker 7: encourages a collective learning process. One of the reasons why people contribute is the hesitates to contribute is the feeling that their work is too small to matter. But

Speaker 8: as said in Janganath space, every contribution Even a single type of fix gets recognized. And this principle applies everywhere. The Jumbo newsletter is an amazing example of treating contribution

Speaker 7: Like a collaborative performance, the lead roles and the supporting cast everyone matters. Shoutouts for documentation fixes, post-time PRs, and even thoughtful issue discussions reinforce that no effort. is wasted at all. Now

Speaker 8: that we have talked about the importance of psychological safety and practical examples of it, let's discuss how we reinforce Beyond just having the good intentions, one of the simplest, yes, most powerful tool we can use is structured feedback. The way we give and receive Feedback determines whether contributors in an open source environment feel motivated or discouraged.

Speaker 7: This is where the sandwich feedback algorithm comes into picture. It's a simple yet effective way to ensure that feedback is constructive and encouraging. You start with something positive, acknowledge that worked well, then provide the areas for improvement. With clear, actionable guidance.

Speaker 8: And finally, end with another positive reinforcement or encouragement. Feedback done right

Speaker 7: isn't a one-time thing. It's a loop process. A strong feedback loop ensures that contributors always know where they stand. And beyond feedback

Speaker 8: A thriving community needs clear expectations and shared accountability. And this feedback is reflected on every PR review, on every PR comment that we we make an open source but if we food pairing sessions that are held in in within the community are outside Now, another aspect of creating a psychologically safe space starts with us, the people who make up the community. It isn't dictated from the top down. It's built through shared responsibilities and shared beliefs. Every company, every community Has its own unique culture, has its own unique rhythm and way of getting things done, of course.

Speaker 8: But across all high-performing cultures, certain truths remain constant. Which is that there might be a balance between autonomy for individuals and teams and alignment with the broader organization. This is where the Django working groups come into the picture. The Jango Software Foundation, that is

Speaker 7: DSF, delegates sovereign powers to working groups which can then act on behalf of the DSF. without both votes or the approvals on certain specific items for which those groups have been spun up So uh some of the working groups, if you haven't heard, might include the social media groups who take care of the Django's official profiles and

Speaker 8: amplify important announcements on the community

Speaker 7: Then there's the fundraising working group focusing on the corporate and major donations to keep Django financially sustainable. Then there's the Code of Conduct working group who handles reports of violations and ensures that Django remains. a welcoming space and these are just some of the working groups that I've mentioned which are active and currently budding. If you're interested in helping out that That's great.

Speaker 8: Each working group has a charter available on Django's GitHub where you'll find details on how to join. Just head over to DSF Working Groups

Speaker 7: repository and Check them out. And if you have an idea of a new working group, even better. The process for proposing one is outlined there too. Now why I am highlighting this right now, because you know people

Speaker 8: like me, let's say, who have been just started out with their journey in Chang. sometimes feel when I joined it initially might feel that okay this is something really really far fetched is it even for me so that's why my friend I'm here to tell you that The working groups are really, really welcoming if you have any amazing crazy ideas of how the Django Software Foundation should Run, propose that. Go ahead and take the initiative because there's no denying that open source is the technology of the people, for the people and by the people. If we rely on Django to develop our web applications, then Django relies on us too, the community of people who ensures its brilliance endures

Speaker 8: and maintainability remains. And so, you know, it's at the end of the day, Django thrives when people step up, take initiative, and help shape its future. But a strong community doesn't just happen, happen. It's built through intentional efforts. A lot of what goes into creating a psychologically safe environment. environment or good management practices, things like establishing clear norms and expectations so there's is a sense of predictability.

Speaker 7: This is where the code of conduct plays a crucial. role. We have seen that code of conduct mostly at every community out there, at every conference out there, at every open source project out there. But it is not just for the written piece of paper. It is the responsibility of every individual who is using that project, who is involved. in that project

Speaker 8: to respect it and abide by it. It's not, you know, just a set of rules. It's a shared agreement on how we treat one another It provides a safety net ensuring that when conflicts arise, we have a framework to handle them with fairness and respect. Because at its core, Django or open source normally isn't just about code. It's about people. And when people feel safe, they create, innovate, and build something far greater than just software. So I leave you with this. The next time you contribute to open source, whether it's Django or any other project, ask yourself.

Speaker 8: Am I making this space safer for someone else? Am I lifting others up the way I was or wished I was lifted up when I started my journey in this program or in this project? Because

Speaker 7: the real measure of a thriving open source community isn't just the quality of its code base, it's the quality of the culture we cultivate together. And always remember, culture doesn't just cultivate on its own on a fine sunny day. It's the choice that the community

Speaker 8: makes every day. It's the choice that we as the responsible people in the community make every day. And when we get it right, we don't just build better softwares, we build a better, more inclusive future. So thank you very much. That was it for the day. If you have any questions, then most welcome.

Speaker 1: Yeah, thank you. The introduction slide was probably skipped. Can you show it? It's oh maybe f first. A D of

Speaker 1: two or three Since I know there were some issue with the

Speaker 8: presentation, I'm happy to drop the link of the presentation in the comment. Comments below

Speaker 2: if folks want to put it back in

Speaker 1: Thanks There's some questions. Japanese people are often not fluenty

Speaker 1: at English language skills and we were uh worried about the language. But uh when joining a community outside of Japan. Yeah. Do you do you have any advice on how to join a community without worry? See, yes, language barrier is sometimes the initial one of

Speaker 2: I would again say is one of the initial psychological barriers that we all have in our minds. mind that having not having a certain good language, let's say if the community is running in English, not having a good English might harm you or might uh Pull you back from joining a community but honestly

Speaker 7: that does not happen if you choose the right community and the right guidance It's always the diverse and inclusive workspace that has been built across. the community. So don't feel that you have a certain language barrier which might you know harm you from joining the community. The translations are present everywhere and especially when we talk about open source, most things are virtual. things are you know happen publicly so the translators come handy and the folks are really really helpful so the language barriers won't be a problem So don't you know have that fear in you that the language would stop you from it.

Speaker 8: Go ahead, take the pride in your culture and cultivate it and make others learn too. So it's great

Speaker 7: The next question we have, do you have any advice on how to join a community without worry? Yes, uh it's always about the first initial step you take. Always remember that you know time is a non-renewable resource. If you're giving your free time volunteering for a community, you have to choose it right you have to choose it what aligns with your passion and goals so it's always about taking the forced step it's always about taking uh the rest of the group together of who you join. So think

Speaker 8: go ahead and without worrying about what would go next down five years in the line, just go ahead if you want

Speaker 2: to join a community, be the change that we want to make Then we have another question from Sarah. What is the one thing you feared before joining and participating in the program that you no longer feared? So uh I would tell my personal thing, okay, what happened was when I started my journey in open source Back in the sophomore

Speaker 7: year when I was in my college, it was mostly like I had to submit a comment on a public uh, you know, GitHub repository. I was so afraid that what would the other person think because there was some installation issues in my uh in when I was running that code based on my system So there was one of the friends who mentioned that hey, nobody's going to come and you know scold you of why you put that comment on GitHub. Put it nicely. And let's see what happens. And the next day I really got that comment, uh, reply back. And so joining the program When I had to join the Jangat Space

Speaker 8: program, I was so confident in putting out my uh desire or passion or what I wanted out of the program straight there and also going ahead

Speaker 7: The same with when it came to the fundraising working rope of how publicly the PR or the system goes. So that learning in public and not fearing of the repercussions, what it might might have later. So that's one thing that I really really learned during the program and I no longer fear. So

Speaker 2: we have another question that when I'm really tired, I sometimes feel it's a hassle to answer questions about my open source product

Speaker 7: Should I try to answer them anyway? See, uh again the concept of psychological safety and from the maintainer POV From the maintainer point of view, there are the cases of burnt out. There are the cases when you are you have filled up your plate so much that you're like Should I answer this particular question or should I even get involved with it? Why you have no back in the line you enjoy it of course? So it's always fine to take a break and Take time out for yourself

Speaker 8: and thereafter whenever you're refreshed, whenever you're back, reply them. It's not like you're not, you know, taking a

Speaker 2: stop altogether, you're halting that. Just taking the right break at the right time.

Speaker 1: I hope that answers your question. Okay, thank you. My question. Thank you. In Japan, we welcome bigness in communities on this code. So it is interesting for me.

Speaker 1: Yeah bueno

Speaker 1: Okay, just finish this session. Thank you. I'll just stop the stream. I am really sorry for the presentations. I don't know how it made up a big mess. Oh okay. Yeah don't worry. Okay.

Speaker 1: Okay. Yes, I can see your uh presentation. Okay. And you see my presentation, right?

Speaker 7: Okay.

Speaker 1: Right. Okay. Can I start, right? Yes, yes, yes.

Speaker 7: Okay, hello everybody I'm uh really uh glad to be on this conference. Uh please once more uh time feedback if you can hear me and if you can see me and if you can see the uh presentation. Please feedback. Can somebody give me a feedback if you can see me and uh hear me? Hello Probably in chat or uh some there. Okay. I don't have any feedback. I hope it works. Today I want to tell you about

Speaker 7: technology which I uh uh name uh micro jungle. And uh this is uh not something new, this is not invention, this is only the pattern how you can organize your work with Django to create, I think, microservice And uh before I start, I want to say a special thing to my family for their supper I want to say thank you for my wife, uh Elena. I want to say thank you to my children, Mark and Maya, and I want to say thank you for my animals, Marcel and Kisa. And also I want to say special thanks to my children, Mai and Mark, because they created all illustration for this presentation.

Speaker 7: Thank you, uh thanks uh Mark and Maya. And right now I want to tell you who I am. My name is Maxim Danilev. I am more than 27 years in commercial development And uh more than ten years I worked with Python and Jenga and also I have um I worked with uh reactive uh JavaScript frameworks on the front end And

Speaker 8: since 2026, I worked like a mentor in engineering.

Speaker 7: And the

Speaker 8: last three years I am mentor for free for young developers And you also can book uh session with me if you find uh my person in internet and uh if you find uh if uh you can simply send Send a message for me. Link to all slides I provide here, and also you can get link on the last slide. There I put

Speaker 7: all examples from this presentation.

Speaker 8: This example works, you can take it and use it. And also you can see I am a professional alpine skier uh skier, and I like uh ski. I like mountain ski. Okay, uh the preamble is finished and uh right now we start to speak about microgen. Microjango. Which idea is behind these words? Micro Django, this is a description of Django project Which contain only one single PI file.

Speaker 8: But if we speak about these projects, these projects have full Django functionality This project can be run alone or can be run it in monoliths And this uh testable, migratable, all uh or or you can install some additional um Batteries uh for for Jungle projects. It's all only one single TI file. And it's interesting uh because this file is pretty small Let's see in detail. What is in this MindPI

Speaker 8: single file? This is the full text from Pyle And if we concentrate on the uh on the file, we have three blows at first. We have import Uh there I import important uh parts uh from Django and on the line number two I import ISDI uh handler which run for us our project. On the second block, we have the endpoint which I've writed in I think more. This endpoint stands for us small Jason uh hello world.

Speaker 8: And important part also we have uh we define URL rotor uh

Speaker 7: for these endpoints

Speaker 8: it all

Speaker 7: and um

Speaker 8: how to run it I use uh

Speaker 7: Ubicorn to run uh this um project Uh in this case

Speaker 8: uh it's um

Speaker 7: required

Speaker 8: to set up uh and in environment uh jungle settings model uh variable and uh after that I run the command uh I run UVCorn which run for us IDI handler and you can see I run it with four workers Uh uh at first I uh set up uh the uh s uh uh the variable with settings uh and uh and the second I run my application with four workers. And you can see the results from terminal I put on the screen.

Speaker 8: And It means this endpoint works asynchronously. And you can see how it works, how it works in browser. I simply write uh the uh path to my endpoint and it works and also uh I um also I uh uh show how it works uh with um uh okay I say about uh it later. Simply you can see this endpoint works All code examples you can find in the repository, and you can try

Speaker 8: right now if you already catch the link on the start of the presentation. Right now I show you how how it seems and um microjungo is only technology which helped me to create uh the um I think I think uh endpoints. But we have on the current mark uh development marks a different competitor who works the in the same mode like jungle. If I start to compare, uh mm on the left side on this slide you can see example from uh n written on jungle, on the right side you can see example

Speaker 8: written on FacTapi. And you can see we have the same quantity of y to get the same result. But MicroJango, if I run it in uh Docker container, my uh container contains around fifty megabytes And fast API container uh only for um for this endpoint takes uh uh uh in fifty percent more Uh uh this is only if I speak about who. But what about performance? Because uh the performance is important Right now I provide for you

Speaker 8: two quick words to my previous talks or live coding session On the left side uh you can find a link to my talk Django FTL Faster Than Live Live. And in this talk I describe How to create Django which works faster than for example PyDantic or Fast API or any other Python uh Python web framework. In reality I cannot say faster, but um comparable fast, uh I'm sure it can be comparable fast And on the right side, you have a link to live coding

Speaker 8: session on XMPython concerns. And in this live coding session I uh made the direct comparing uh the performance of Django, uh microjango project and performance of of Fast API project. Uh in I think um This I think uh paradigm. And it depends on settings, you can find it and you can see it's uh comparable History about this technology it's important to understand. This technology appears Uh in two thousand fourteen, Julia

Speaker 8: Ellman and Mark Lavin uh created the book Light Stay Jungle and on the first Part of this book they present Jungle project in one file. It means this technology is already uh mm uh uh in life uh usable is already ten years. But mm If you can see about micro jungle, people start to speak a little bit later, and I tell you why. Uh this is an amazing book, I like it. Uh you have there example of uh how you work with Django. mm uh with uh web socket and so on, with uh

Speaker 8: uh with salary. It's perfect book. And um a little bit later I meet this technology in uh uh Europe um the Armin Wolf Wolf and Flo Air Motor uh from Isengard Ate um presented for me a small uh MVP project and they also used simplifying architecture of Django project. It was around five PI files and not more A little bit later I miss this technology in Python Web Framework Benchmark, which uh created by Material

Speaker 8: Klenov and idea from these benchmarks they run uh he ran Django project in one file And in additional they have uh he has uh uh staking spI in separate file. And the next explosion in about uh talks about this technology was uh war uh um happens on the Janga conference in United States in 2019. Carlton Gibson, Will Vincent, Peter Baumgartner, they all discussed it all and presented in repository Django MicroFramework. And

Speaker 8: only what I see what changes they simply improve ideas of life they life they gender. It's interesting because they simply repeated it But uh they used the new name. And in two thousand twenty-three Paula Melchiora uh this is the uh Django Fellow right now. Uh Paolo presented again on DjangoCon United States 2023 the same project with name Micro

Speaker 7: Japan. You can find it in uh in

Speaker 8: on GitHub, but uh the

Speaker 7: important steps which he

Speaker 8: did, Paula They presented I think like the jungle project and offer also to use new name, Micro Jungle. And this is uh it was the important step, uh for example for me, because after that I start to use this technology And also later I found I find many different uh repositories there they use Uh this report uh this uh project used the same paradigm, light-based jungle paradigm, and it was in uh Europe uh teeny jungle it was a nano

Speaker 8: jungle and I I'm sure you uh have seen different other ideas how we can create Django uh file the Django project in one file. probably with additional libraries or not. But I speak about raw jungle projects. I don't install anything else Okay, micro jungle. How it works in real-life application or why we can start to use it And there we can use this example on the right side , this example from RealProje.

Speaker 8: In my experience I work in big uh projects, it was a messy monolith uh with around half thousand enfoils But to run this monolith was uh not uh uh um not uh this monolith was not always stable and uh special endpoints we want to extract. And these endpoints uh with my uh help was transformed in micro-jungar endpoints and they was run it separately from Monolith outroof These endpoints still

Speaker 8: uh present in monorepo It means I can develop micro-jungle endpoints in monoliths and I can run it separately or I can run it in monoliths paradigm. This is important and that's why my Janga can be interesting for Jenga project because I don't I'm I'm not refactor anything, but I can run single entry points uh for web project uh standalone and uh uh and I can improve the stability of this entry point And exactly this uh

Speaker 8: this example uh I or we use in uh course As claress -based project, there we want to separate or segregate tweet comments and write comments. Uh we have one entry point for uh based on micro junk which only write to database. They work Stand alone, don't matter what happens with project , it simply writes data to database It works, it works fast, it works in I think more, it's amazing. And after that uh the big monoliths uh get the data

Speaker 8: uh compute uh the data and uh send results uh back to the customer. Also As I already say, I can use microjungo in modular monolith projects. It means I want uh still to work in uh Monolith, but I want to run my um my entry point uh separately. And also there I can use micro-junga technology, I can use it in async-oriented jungle projects. Uh what I mean if I want if I want uh uh simply to switch uh to switch my uh project in a sync

Speaker 8: uh mm style I can step by step um migrate from uh sync entry point to a sync entry point

Speaker 7: Uh probably you can find uh many different other many different other

Speaker 8: uh uh ideas, but this uh these ideas uh uh which I have Uh I already use m uh in these ideas I already use micro jump. Or uh m please one second.

Speaker 2: Okay. I hope I can continue. Uh and if somebody writing chat for me, uh it's all okay, um it uh would be mm mm very very good because I don't have any uh um I don't have any feedback for me Okay, and uh what I uh what we can uh uh

Speaker 8: how we can pick about full functionality in uh uh our project To add full functionality from Jungle, I mean the possibility to run some management commons, we should add only three lines in our core in uh in our micro jungle probe and this at first. This dih we should add this behind Uh because uh because the standards uh management omens work in sync paradigm, not in a sync paradigm and the second uh we should add the possibility to run uh to to run

Speaker 8: our um our file uh l uh in python manage pi uh style. For example, if I

Speaker 7: uh uh uh if I name this uh file mm minus pi

Speaker 8: mm I already

Speaker 7: can can

Speaker 8: run these uh these uh mm done the project and I already can perform some some management comments like uh uh create super user or migrate or mm collect static and so on. But in real uh projects If all uh commands already perform, I don't need this line in um uh micro jungle file because uh this uh I need only this file for uh to to work with entry points more and we don't use any any

Speaker 8: manage PI comments uh there. Okay, uh this is usable example. This uh model data serialization. In sync mode, you can see I uh uh it's written uh it's uh uh declared uh pl generic class rate view which render to respond uh serialization of my object and in this case this object is user. And also on the line 16 I this declare the URL to this uh this view. You can see

Speaker 8: uh I use to serialize data, uh I use standard Django serializing. I don't need any other libraries to serialize my uh object in JSON And also you can see I use the special object from uh uh Django. This object is simple lazy object In reality, in Python, we should have uh only lazy objects, but not too much developers work with that. And um what is this simple lazy object? Uh in general this is an important uh element because

Speaker 8: uh We declare some uh or we want to use some models or some objects which not existed in uh in project right now, on the start of project.

Speaker 7: And this uh simply the object raptor help us to delay

Speaker 8: uh

Speaker 7: to de delay um declaration of objects. In in this case I create the revelation to objects uh before this object appears in my project

Speaker 8: This is important and this paradigm you should understand uh if you work in micro-genre, I think, paradigm

Speaker 2: with your project. I hope you understand what I mean Uh but I don't have any any feedback from chat. Okay. The next example which I show Uh this is this is a sync example and here y you can see the uh this is the same uh the same entry point I seralize uh light uh

Speaker 8: my user objects uh in JSON. Uh but uh here I use Here I I try to use a simple possibility from Jenga and uh the problem about it we can uh discuss a little bit later. And um It works. And one small trick which I want to show you, this is the line number six. The line number six is a decorator to uh to organize the um uh to organize the URL

Speaker 8: fund for this view and it's written in decorator mode. Exactly how you work with Fast

Speaker 7: API you can also decorate your views with uh Ural buttons. I don't like this style with Lambda and with mobile But it works. If you want you can use it, but if you uh worked with uh Django

Speaker 8: In normal modes uh

Speaker 7: usually we use simply URL uh we only describe URL patterns at all some there in uh uh in file and uh below this view. View should be um uh should be created before Okay, but we can use it. And

Speaker 8: also this example, this is how I write a sync test for my approach. The test I put uh put in other folder. Uh we don't need it on production, but if you want to test it, it is possible and I think testing is uh also possible in Django and that's why that's why I I uh I provide the example here. One second. Uh one second and uh it's also work, you can use it. And some work

Speaker 8: is about second. Settings PI. It's interesting because I don't have second spI in my microjungo project. Because uh usually we have test settings PI, we have broad settings PI, uh we have the environment. This is a system type And after that I publish uh I publish my new micro jungle project uh in Docker file and simply I provide broad settings for this uh uh jungle micro jungle project Uh and that's why uh past settings or quote settings I don't have in mm in in project. It's uh

Speaker 8: placed uh somewhere uh somewhere on production server mm and so on. And also for testing I have the uh uh test PI uh And uh it means on the um test stage we need also test PI file uh but on production I don't I don't have uh on uh any other files only mine pi which I run with Ubicorn in Docker And I provide some uh guidelines how I run it on the uh Windows and you can check how you can work it with uh, for example, with Linux.

Speaker 8: This is an example of how I run single service. MicroJung , I have the entry point in the folder in application second step. and I run uh it with test seconds or with second and also how I can run it in modular monolith I have the um application first service, uh micro junk application, I have the application second service, and I have also the uh general uh mind PI this fold uh this uh general mind pi goes through f uh through f whole folders collects every mine pi and add it in the router and after that it will be run

Speaker 8: it And every time if I go back, every time a router calls microjung entry points from first service or second service accordingly to the request. All example you can find in uh repository uh this uh it works Okay. Use case and Docker. Uh uh the Docker file I also provide in uh mm in the repository, and you can see I don't have any requirement PI. It's interesting because um we use uh we install uh the important library

Speaker 8: directly in Docker file. And in this case we don't need any state in PI or setup PI file. Uh it works like a charm. I like it. You can um uh also to two or we can if we have time we can run it uh we can run one server uh

Speaker 7: service which uh I publish uh public But in my experience, right now it's around uh four hundred twelve entry points works with microjungo paradigm in my project. Okay, Micro Django. Oh

Speaker 8: too much words about it, but problem. There is the problem why Micro Django not used too much The first obstacle, lack of documentation.

Speaker 7: You don't have any information uh about uh we have around ten percent

Speaker 8: of uh the uh uh possibilities are documented about Django. And too much in far uh too much possibilities are not documented. The second problem lies in hair because I have here Many, many, many times the opinion what jungle is tall, jungle is slow, jungle is big, etc. Outroof, jungle is small! And uh it it was uh it was amazing uh why I why so happened. Uh the oh sorry. Uh sorry I

Speaker 8: Uh uh sorry I pressed the wrong button. Okay. The mm uh Next problem is height around other frameworks. For example, I want to start to use uh uh

Speaker 7: to to create the

Speaker 8: um the entry point, I think entry point and if I write async entry point party you can find the million articles and million me million opinions which frameworks should use and you cannot find Janga in this list This is a problem because Django is comparable with other frameworks.

Speaker 7: And also obstacle which you cannot avoid right now This

Speaker 8: is the big problem. Django is not ready to work. Full in a sync paradigm. For example, if your model has related fields to other models like a forage and key or many-to-many

Speaker 7: fields

Speaker 8: In this case, uh you your project cannot work in um I think paradigm.

Speaker 7: You get always uh

Speaker 8: always uh this uh

Speaker 7: warning like here.

Speaker 8: And in this case to work with many mm binded mm binded uh uh

Speaker 7: uh uh m

Speaker 8: objects and models

Speaker 7: in this case Every time it should be uh it should be asked the first uh first object, after that it should be requested uh uh asked the next uh binded object And probably if you start to transform it in a sync mode, you can see how the performance or complexity increase uh grows in your project. Okay, and the last obstacle still exists, and last obstacle it's probably the biggest problem for a person who wants to create micro jungle async

Speaker 7: project. Okay

Speaker 8: and the next problem uh which uh I also want uh want to describe The I like to work in generic class-based view uh style, but they are completely not a thing. Every uh every um Methods in generic class-based view which you want to use in I think mode should be overridden.

Speaker 7: And this is a problem. For example, you can see the standard getObject method in detail view uh this getObject method should be overridden completely because these methods are not uh created to work in a sync mode This is a problem also. Okay, summary about my tool and summary about microjungo uh paradigm to create a sync mmm uh

Speaker 8: entry points or async microservices. If this

Speaker 7: pattern is ready to work, it's already exists ten

Speaker 8: years The second , why we can use it or w why why I speak about it?

Speaker 7: Because It

Speaker 8: allows me to create endpoints without switching technology, without switch to light

Speaker 7: uh

Speaker 8: lightest

Speaker 7: framework for example or uh to some uh to create some endpoints on the goal uh

Speaker 8: for example

Speaker 7: and

Speaker 8: If I uh serialize

Speaker 7: uh only flat objects without any relations In Django, in last version, Django 5. 1, it works. I can create I think endpoints and it will works like a charm. And also it comparable with other competitors. But problem is not all goals in my project can be realized in a sync paradigm. And the also the problem? This all partally not documented. For example, how I test my async views You cannot find too much documentation about it. Or

Speaker 7: there is the problem or how I should redefine the Descriptors for related objects to work with uh

Speaker 8: in a sync mode, it's completely not documented. And it's hard to change it.

Speaker 7: But

Speaker 8: I see how the community works with uh Django paradigm and uh

Speaker 7: I'm happy to have in

Speaker 8: next version of Django uh

Speaker 7: I think cache

Speaker 8: decorator because the cache uh problem

Speaker 7: with caching is already solved.

Speaker 8: I can create uh um I can up uh cache to my async views. It's amazing and uh more entry points I can create with MicroJunger right now with a new version of Django. And this is all what I want to tell you today about micro jungle, about microservices and make microservices not var Thank you for your attention.

Speaker 7: You can ask me anything and link to slides a code you can find on this slide. This is all. Thank you

Speaker 1: Okay. Thank you.

Speaker 7: I hope uh I pr I hope it was clear what I want to tell about Janga.

Speaker 1: That's interesting. Uh I have some question. Uh

Speaker 8: please.

Speaker 1: How do I get started with microjungo? Uh is there uh is there a temperate I uh uh is there a temperate I can use?

Speaker 8: At first you can download my

Speaker 1: repository and you can take the mine PI. This is completely fully created

Speaker 7: template. At second, you can visit the microjungo repository from Paolo Melchior, because there it was the first example, and of course, how you can start Simply you can read the book Lightway

Speaker 8: Jungle, because in this book this technology was presented and completely described

Speaker 1: Yeah, is there the uh documentation about these patterns?

Speaker 8: Uh only I I know only one book about it, Light Faced Jungle. I tried to show you I tried to show you this uh uh

Speaker 7: this uh book uh sorry how it works. Yeah

Speaker 8: Uh simply like

Speaker 7: V jungle

Speaker 8: and You can you can you can buy this

Speaker 1: book

Speaker 8: this book is already ten years

Speaker 7: uh old.

Speaker 8: It was created in two thousand uh

Speaker 7: uh two thousand fourteen Yeah, but uh uh it uh this technology is still not famous. Don't worry. Welcome. Welcome.

Speaker 1: Nani Kahukanishumu Aruba Areba Hu comentoto demo Kaiti tada Karepa to you must ni mm demo the YouTube.

Speaker 7: If uh you don't have uh question right now you can ask me always later And in this case, um

Speaker 8: I should say goodbye, right?

Speaker 1: Hey , uh so this night so gonna Ah, wie heißt der Comments?

Speaker 8: Okay , I put in chat um the link to uh live day jungle and I put in chat also

Speaker 1: There is a link link

Speaker 8: link to the to the mm I think what's the better?

Speaker 1: Are there any services released on the site using this microjungo? You are using uh microsion of pattern in production, right?

Speaker 7: Uh I use uh I use uh micro jungle in um in uh In production?

Speaker 1: Yeah.

Speaker 2: We have around uh four hundred uh twenty or four hundred fifty In uh entry point

Speaker 1: In the point. And so we can

Speaker 7: check it. Yep, four hundred

Speaker 1: Those four hundred entry points are wait a second, I should deploy um With the Unicor or uh Ubic.

Speaker 7: Um uh wait a second about Uicorn or um uh

Speaker 8: please ask

Speaker 7: more time I use Ubicor. Okay. Um View Oh, there's my repository. This this repository

Speaker 7: Can you achieve it? Or it's public, not public.

Speaker 1: Go ahead.

Speaker 7: Uh this is the private, okay. This is the public right now right now. This is the public right now If we finish, uh I want to say thank you for conference and see you, right?

Speaker 1: Yes.

Speaker 7: Thank you very much.

Speaker 1: Okay, thank you very much. Thank you. Thank you. All right. Bye. Bye.