keptn Community, 6 Apr 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keptn - Google Summer of Code office hours - April 06, 2022

Description

GSoC Q&A, Discussion of Keptn and GitHub integrations, discussion of the Jenkinsfile Executor Service project idea. Meeting notes: https://docs.google.com/document/d/1y7a6uaN8fwFJ7IRnvtxSfgz-OGFq6u7bKN6F7NDxKPg/edit

Keptn in GSoC: https://github.com/keptn/community/tree/main/mentorship/gsoc

A

So hi all uh today we have uh a special session to discuss google summer of code and answer whatever questions um and pretty much every any other topic in the community. If you wish. So we have several people on the call and if you want to ask any questions, just unmute yourself and we can start from there.

A

I'll share the link to immediate notes, just in case we need them.

B

Hey good morning, oleg uh brad, how are you guys? I'm very good? Thank you. How about you uh doing well doing well, uh yeah! I just had a a quick question about, um I guess, google summer of code- and I guess this is kind of for uh oleg, but also program as well. um I was wondering like um because I noticed, like oleg specifically you've, had like a lot of experience with uh google server code and like open source um as well.

B

I was just wondering like, what's sort of the um I guess like, what's sort of the optimal uh um kind of like the optimal result, I guess for like um participants, whether it be like for mentors or contributors or for the organization as well, for google center, specifically.

A

So you ask about expected results right.

B

A

So, first of all, google summer of code is a mentorship and outreach program. So the main objective for this program is learning and learning together, because google summer, of course, offers learning opportunities not only to mentees, but also to mentors just different areas. Different skill sets and working together having fun learning. Something is the main objective. uh Secondary objective is delivering something.

A

Google summer, of course, is really about code, and we expect mentees to participate in the community and work with the community to deliver some projects. And, of course, it's a really nice thing to do in many projects I see so there are two approaches, basically to gsoc in our organizations. The first approach, which I don't really like, is having a student working in a fully isolated environment to have a kind of second side project with an organization. Maybe some prototype is one of the approaches.

A

Another approach is to have uh mentee working on something uh within the main community roadmap um maybe still isolated to some extent, but that has high demand from users and high value to the community. So it's not like a sandbox project which just happens, and then it gets stable without any results.

A

I prefer to see different approach when students really engage in the community work on something valuable, whether it gets delivered or not, whether it's a prototype or ready to production at least being involved in some area. That is really important for the community and awesome. Personally, I think that all projects should be like that, because, again for community, it's a learning, experience it's coming into bonding and if it works, it's perfect. If you also ship something, it's also awesome yeah. I have seen many great projects done, for example in jenkins.

A

I can provide a few examples if you wish, but uh yeah even.

C

A

uh The final bits are not shipped, it's still a great uh for the community to have such projects.

B

Awesome yeah. That sounds great. um I guess just like a quick follow-up, then um I guess keeping that in mind like so then um sort of other than like technical skills um like solid technical skills. I guess like what are some things that you consider to be like. I guess like an important trait for uh being like being part of like an open source community um or I like being a picture leader.

D

C

D

Yeah um so yeah, like um the the great thing about the open source community, is that everyone's welcome, so that there's so many ways that you can contribute, whether it's non-technical contributions to technical contributions? um Yeah, it's really inviting, and especially the the captain, one as well um you'll find that after gsoc you'll find yourself hanging out there as well. So you can meet some real friends.

D

You know a great place to start is documentation a lot of folks start documentation and they sort of good places to understand what the tool actually does and just try using it yourself. And then you can see the areas of interest that you're into and then and then just all you need to do is like uh there's like a saying that they use is. The only way to get involved is to get involved.

D

So just jump in this like channel, say hello, say you know a little bit about yourself and then everyone will swarm and just just want to help you and it's you know. The power of the community is amazing, so yeah you can contribute in so many different ways and- and you don't have to be a rock star coder like that. That's what um we're here to help you with that as well.

B

That's awesome, that's awesome to hear thank you so.

D

Welcome it's going to be fun.

B

A

So yeah there are questions in the background from omar about uh how to get started, etc, so that we answered in the chat any other questions. You would like to bring up.

A

Then maybe uh it makes sense to take a look at the timeline and just so that everyone is on the same page. Oh sorry, so yeah just click update on the timeline. uh The applications for students. uh Sorry for contributors uh are now open and uh they will be open until april 19th. So this is the application deadline. So now you can go to the jsoc website and submit your application uh as a draft or as a final application, um but yeah. This is the format process.

A

Informal process followed in most of the recognizations that once you have a proposal draft, uh it's better to submit it in project charts like in captain slack, uh so that you can have a discussion happening there and you can have a few cycles of feedback.

A

uh The worst thing you can do at this stage is just submit your final proposal on april 19th without getting any feedback, because in this case you don't really use an opportunity to discuss your idea with mentors and without active discussion without community engagement. Usually these projects do not get on accepted. uh There is exceptional basis, uh so my recommendation that in the coming two weeks, uh use all the opportunities to discuss the projects.

A

If you have a project draft when submitted to the slack start the discussion in public, if you don't have a project uh proposal draft, yet maybe it's something to work on, even if it's not complete, even if it's not perfect, just trying to narrow down what you would be working on uh reviewing uh the project, deliverables, etc, it would be a good start, and if there is something unclear within the project idea, then again it's time to get. It clarified with potential mentors.

A

And yeah we have just two weeks left: it's not that much time. So please use this time and engage with the community. If you haven't done it yet and if you have already discussed it, we had a lot of discussions. You had some prototypes shown by potential contributors.

A

Please convert them to proposals so that we can start looking into that.

A

Okay, so this is what is going on now. Any questions comments.

E

Yeah, actually I talking about the time stamps. uh What's with their procedure like I just like to contribute in whole three months or is it like? You can take some leave in between the work in between the words and we can extend the workload, maybe five months? Is it like something? This makes sense.

A

Okay, uh so uh there is a strict timeline for google summer of code uh this year they made some adjustments, uh talk about it uh projects. So now you have an opportunity to have a short-term project which is 170 hours and full project which is 340 hours.

A

These projects happen within this time frame. So you can see that this year you can take extended timeline.

A

It's something you can discuss with your potential mentors and application, because it would require availability, not only of the student.

C

A

Also of mentors for these extended timelines, but uh it's possible these days will definitely based on your capacity. You need to decide what is the timeline for the project? uh What how much time you are ready to dedicate.

E

Yeah sounds good.

A

So it's actually the first year when uh extended timelines were introduced in google summer of code uh before that it was strict. So until the beginning of september, everyone was working in this time frame, but this year they introduced a few additional options just to accommodate with more flexibility.

A

Yes, it's up to you to choose what options you would like to. But if you really want to have extended period, then it's something you need to discuss with potential mentors because they need to also commit to longer time frame or we need to find some options.

A

Usually it's possible because actually summary is the most challenging time because of vacation periods, decay being off, etc, etc, but still it's better to discuss.

A

Does it answer your question.

E

Yes, yes, yes,.

E

But it all depends on the mentors actually how the commitment they have and uh how you are going to uh make your project uh in the next level.

A

Yeah, so this is something we will need to review so once the application is over, they will be almost one month for our mentors and our cartoons to form mentorship teams and to ensure that we can provide this mentorship capacity.

A

So there will be some time to close the gaps if they need it.

A

We usually tend to have three mentors at least per project, because it's normal that one or two mentors have to change their priorities, have to step down from gsoc life happens, and we want to have projects to be robust so that one mentor stepping down etc doesn't lead to the project being in limbo, and this uh one I think four can means to actually take care of.

E

Yeah thanks so much yeah exactly.

A

Well, it's still tricky every year there are surprises, sometimes uh well uh one thing. I saw that in 2020 for one project we had six mentors, but it was 2020 and five of these mentors had to step down over the first two months of the project. So this was quite a ride, but still the project was successful because we had a plan b. We had organ means engaging.

A

So sometimes it's not straightforward and be ready today, because it's community, so even if mentors do some kind of commitment, of course, some things may go wrong same for students. So sometimes, if you get ill, it also happens and the it's not like. You sign your gso contract with blood. uh So sometimes it's fine even to stop the project. Sometimes it's fine to make a break if needed, and we ask all teams to be flexible, but in your application, if you have an expectation about your commitment, it's really important to set it up front.

A

So, for example, if you want to have a full-time uh project, so say that you are going to work full-time 340 hours and then it gets discovered that you also signed the internship contract. For the same time, then you aid a hero who can work 80 hours per week and also study in peril.

A

But hint 99 of us are not such kinds of heroes and uh please, if you have such significant commitments, uh make them explicit upfront so that mentors can take it into consideration uh and, for example, for students in india, it's normal to have placement exams and the university in parallel with gssoc for some months. So if you have already have expectation of these periods of unavailability, it's better to include them into your proposal so that mentorship teams can plan for that. None of that is a show stopper, especially for uh short projects, so 170 hours.

A

But it's better to know about that. So all sites informed.

A

Okay, any questions or if you want to discuss specifics of particular projects, we can dive in.

F

uh Actually, I do have a question, so I.

C

F

Have actually any prior, I do not have actually any prior experience with open source, so would that be any? Would let cause any hindrance to my proposal.

A

So for proposal uh there is a gsoc student guide which explains what needs to be in your proposal, and this guide is quite explicit. Let me show it to you.

A

So this is my screen.

A

So how to find it uh um yeah, maybe it's something that needs to be linked more explicitly because they worked this side. Now it's uh hard to find everything.

F

Actually, I do know where to find it so.

A

I'll just take a shortcut from my other organization.

A

I believe it's here, yeah.

A

So this is uh the guide which is somehow linked from the pages, and my take away is to actually have it linked from our landing page, but it provides a lot of information about what needs to be in the proposal and there is a section called elements of a quality proposal.

A

So here you can find what needs to be there, and there are two important sections so, first of all, deliverables so be exclusive about what you are going to work on and what you it's going to be delivered as a result of your proposal.

A

So it can be new code, new plugins, new integration services for captain new documentation whatever. But it has to be explicit because this something that converts text an idea to a measurable criteria which you will need another important thing to always keep in mind- is benefits to the community.

A

So when you have a project so for some project ideas you have listed, uh there is explicit benefit to the community, but still always think that if I make this proposal how it helps captain end users how to help captain contributor community, especially if you want to come up with original project idea, which is totally welcome.

A

But you also need to keep it in mind. So these are probably two main sections and the rest just what is here in caption, we don't have specific additional guidance for this year. We don't have specific templates, because this guide is quite self-sufficient, in my opinion, but yeah. What I mentioned before that, if you have some details about your availability, etc, it will be recommended to put it in the project proposal.

F

Does it answer your question uh yeah? It actually does answer my question. Okay, uh I.

E

F

I also have one more question, and that was that: can you actually elaborate on the idea that improve github integration in captain.

A

um Okay, uh so maybe we discuss it a bit later, so I will answer one question in the chat and then we can uh for focus on this idea. So there was a question from nishant how many applicants is captain taking this year?

A

uh I don't have answer to that right now, uh because how it happens. Once we have uh project proposals submitted, then they will be reviewed by the organization.

A

So we, together with mentors, we will review all the proposals and basically decide how many projects we would like to have and how many projects we can support by mentorship teams etc. The enemies have been two numbers to google. One number is a minimum number of projects we would like to get another.

A

One is maximum number of projects we would like to get, and then google processes these requests from all the organizations and comes back with a number of projects we get, and then we do the location so right now we don't have exact number for how many applicants we are taking.

A

It's not up to us to decide to some extent, but good news is that, depending on the number of quality proposals, etc, we can justify high number when asking for slots from google.

A

At the same time, it's important to know that the captain participates in gsoc first year and it means that it's unlikely that there will be many project slots allocated so usually for first year organizations they allocate two or three slots, maybe sometimes more. If we can prove that we have experienced a mentorship team, etc, but usually the number is relatively low for the first year.

A

uh So I wouldn't expect that we will have let's say 20 students working on captain this year. Not at all, we don't have so many mentors. We don't have the capacity for that. So I would say that the expectation, the maximum will be maybe five, and if we don't get enough slots, we from google. We also have an opportunity to run community bridge projects.

A

So, in this case, you shouldn't be too concerned about the number of slots we are getting, because if you have a good proposal, if you have potential mentorship team and if the only the reason is not having enough slots from google, it's the easiest possible situation because we can run our lfx mentorship program. uh Getting budget for them is relatively easy. We can shake some freeze and get money for that and we can run a program so right now.

A

I would rather focus on just preparing good proposal uh working with the mentorship teams getting feedback, and if there is joint interest, I will see how we can make this project happen.

A

So, yes, all said that one gso contributor participants per project- this is a general rule, but there are some exceptions.

A

So in theory, if you have, let's say two students, two contributors who are interested in a single project- let's say we have this case already in a documentation engine.

A

It might be theoretically possible to have two students working on this project, so that many times I still keep thank students for gsoc. I apologize for that, but to do that, these projects have to be independent these projects, even if they pursue the same idea, they should be able to, firstly be successful separately.

A

uh They should be able uh we should be able to evaluate them separately and they shouldn't block each other, because every project may not happen may not be successful due to any reason.

A

But if you see that there are multiple students sort of contributors interested in the project idea, uh it's open source. This is why we recommend uh public discussion etc. uh If you see that there is a kind of competition, you have two options: firstly compete.

A

Secondly, you can collaborate work with each other find potential solution, and this is also how open source community works. So if you come up with potential plan, how two projects could happen independently, how you could have separate deliverables, etc? Then there might be two projects for one project idea: it's rare, but it happens.

A

Okay, does it make sense.

G

It it does make sense. Thank you, yeah.

A

Okay, so if you're in this situation better to start early, because it's not always trivial to come up with projects that would be independent enough, potential mentors can help with some guidelines with some guidance, but generally it's up to potential contributors to work with each other and just showing showing ability to work with each other and finding a solution is already a quite important step for having this proposal justified.

A

Okay, so what else we have in the chat today? uh So, let's ask for github integrations then uh jenkinsfeld runner, project idea and.

A

Let's just dive a bit in the project specific details.

G

um Excuse me, can I I have another question. Can I answer.

A

G

um If I do want to review my proposal or how is there a certain procedure that I would need to follow,.

G

If I want to review my proposal, can I just contact any of the mentors and ask them for the review or.

A

My recommendation would be to put your proposal in the reviewable format in the chat associated with the project idea, because it would be the best way to facilitate more feedback. uh What is the reviewable format, for example, my recommendation would be to use google doc where you allow public comments so that people can share feedback, etc. Writing in the proposal.

G

And this will be on the channel itself right of whichever idea. Thank you.

A

So it's the best way as we discussed before. If there is specific interest, we can organize a meeting to discuss uh project idea to discuss some proposals, but all of that happens on the case-by-case basis. So if there is enough interest, we can set up something, but we need to see that there is such an interest.

G

G

A

So this is my screen.

A

Conveniently enough uh energy soccer points to this page, okay, so here what we have, um we have a number of project ids. So again, uh these are projects.

B

A little bigger.

A

B

F

Is it better now.

D

A

Okay, uh so here, for example, we have uh the first question when we got about github integration in captain. So this is a project idea and, let's be honest, it's a white project idea because it has a lot of topics and it's widely open-ended.

A

So the program array is that many projects use github and github is used for multiple purposes. For example, uh there are some apis which you can leverage to provide to improve by the developer experience. So one of examples is in deployment environments. So, for example, one captain does deployments etc. You can propagate this information to deployment statuses so that you can see what's hap going on uh on the github site.

A

Another example is github checks, so, for example, here you can. You have a number of checks, including some checks like netlify deploys the preview environment and it fires you here so that you can easily evaluate this pull request.

A

So all of that could be also added to captain so that you can invoke this notifications statuses as a part of sequences, or maybe you could propagate all sequence stages as checks. It depends on how you wish, and it could be a good improvement for those who do development and who rely on captain, for example, to manage staging environments or to do quality gates as a part of your continuous integration flow, or to do continuous delivery and deployment after the pull requests are merged.

A

In any case, you can get benefit from these integrations just by providing some data etc.

A

So this is one of items that could be done. Other items for integrations is, of course, actions because a github actions is basically a pipeline engine and you can invoke captain quality keys from github actions, for example, for your ci build.

A

So one of examples here would be captain core, because captain core itself has a continuous integration implemented as github actions. So you can see a bunch of various checks and pipelines here, and some of pipelines could also invoke, involve working with captain, uh invoking quality gates and retrieving results, so such integrations could be useful.

A

Last but not least, there is one important thing about all these apis, so checks api deployment. Apis I've shown that to work properly, they need authorization, sorry identification with github application. So, instead of having users and token.

C

A

Can register a github application, let's say captain that actually uses special apis and gets access to this apis on github, so that you will get more benefits to show you some examples of real how it was done so jenkins.

A

Two years ago, one year ago, we had a project for hub checks, so basically it's a project that again integrated jenkins with all these github checks apis, and there was a lot of uh changes done there and it was a gsoft project by karjic young and this project has been really successful, uh widely accepted by users, and it basically goes through various checks, api and authentication support that actually makes it happen.

A

So one of interest, if you're interested more in some success stories, how it could be done, you can just go to jenkins and there you can find this project in archives, and so it's I believe it was 2021 uh checks, checks, checks, checks.

A

Now it was 220 10 place uh yeah, it has checks api. So it's just one of examples. So here you can basically go through stories. What was implemented, how it was implemented again, it's just an example of success story. uh It doesn't necessarily map to captain directly because jenkins and captain are slightly different, but in principle you can explore these areas.

A

Also, there are some other opportunities you could consider. So, for example, as a part of cap 60, we recently delivered support for open id, so maybe rating some proof of concepts demonstrating how captain can integrate with github open id and maybe later, this github security model for authorization once cap 60 is delivered, could be also an opportunity.

A

So there are multiple integrations and if you're a github user, probably you can come up with even more ideas, and this is basically what is this project idea about? It doesn't set particular deliverable. It lists many options that could be done and, if you're interested in this area, maybe what you could do is explore. Development flow with github with github actions, see where you could see improvements for your developer experience with captain and building your project proposal based on that, so it can be quite narrow idea compared to what I described.

A

uh So if you want to deliver anything, it's probably a task for many senior engineers for several months, but taking one particular area that could be interesting, for you is something you could start from.

F

Yeah thanks for this detailed answer and it actually answers all of my queries. So thank you.

C

I had a question: um where does the various parts of github integration uh sit in the captain? Workflow? Oh yeah? This is a good question.

A

Because uh it depends uh so, let's take a look, what we have uh so, uh if you talk about, let's say: github actions, etc. Maybe it's something that could be done as integration service or as documentation for web hooks, and if you want to integrate otherwise to have a github action for captain as polarized. We already have some, but it would be nice to productize them.

A

Then it will be just a separate component that integrates uh github action for captain, probably by integrating captain cli- or maybe you talking to captain through rest apis, but it would be a separate repository.

A

It helped out opportunification service most likely it would be integration as an integration service with captain so basically having a service that drops, requests, etc. Talks to captain core. Maybe it will require some changes in captain core to make it straightforward. But in this case it would be rather integration, service, reporting, application, statuses to deployment api, etc.

A

Again, it would be integration service or maybe some documentation for webhooks. But to make this possible, you will need to have github up authentication.

A

So this is something that will need to do at least some kind of integration to make it possible.

C

A

So again, it's something for you to explore and the good first step for this would be to just try out captain and github and see what you would like to change. Also there are some user requests so, for example, provide integration with github actions. It was created by jurgen a while ago, but I think that it's still relevant.

A

So if you see items like that, just comment and discuss in the captain integrations channel and we can see what exactly could be done. There.

C

Thank you for answering my question.

A

A

Any questions on github integrations or should we move on.

A

Yeah there is comment from suraj about that. It would be easy to understand the integration ideas if you study how captain uh works under the hood so kept as event driven orchestration engine. So if you study captain cloud events, how this thing integrate with each other, then it could be a good addition and architecturally otherwise yeah.

A

Hopefully we can decouple everything into separate services through integration. Apis. Let's see.

A

Okay, there was another question about uh executing pipeline with captain execution plane using changky's file runner. So I can briefly describe this idea. I guess I'm one of the potential mentors.

A

So it's this idea jenkins, pipeline executive service, so how it works. You already saw a job executor service right.

A

A

uh So, basically, it allows to execute various kinds of tasks uh using uh job executor, uh so how it works. Captain I itself has a control plane that does all kinds of orchestration, but if needed, there is another execution plane which allows executing various heavier workloads.

A

So, for example, if you use geometer integration service to run load tests, this execution plane can be used on another example: job executor service, which is basically an opportunity to run random tasks in random utility containers that can integrate various execution steps, so some examples of job executor service being used. So basically it's low level extension point, but if you go to captain integrations, uh for example like here, there is a recent integration. I added, uh for example, powershell.

A

I believe adam gardner submitted it and basically what it does. It allows to do. Powershell execution as a part of captain sequence. Basically, here you can see what happens so it's a task that takes a powershell container and then just runs powershell with particular options. So here it prints hello world, but instead of that it can be quite heavy. For example, running ms build or running a build, make file, etc.

A

So it can be a heavy workload and it executes scripts at the same time executing scripts, it's fun, but at some point you might want to execute something more serious and for that, for example, you might already have a genius job. So, for example, jenkins pipeline is well jenkins, is an automation, server heavily used in ci cd and there might be pipeline jobs, for example, for building your projects for doing static analysis, etc.

A

So I'm just opening jenkins pipeline and just looking, for example like that, so here's an example of jenkins pipeline the way on agent we execute build using, make then run tests, run deployment, publish the reports. So this is basically a pipeline.

A

It's similar to the captain sequence, but instead of being declarative instead of being orchestration focused, it actually does a lot of execution steps and a lot of reporting uh it can integrate with capital as well, by reporting, quality, gates, etc. But this separate story right now. uh There are ways to actually trigger jenkins execution from captain uh just by using rest api by using web hooks, but it uses external jenkins instance, which runs which consumes resources, but there is another project in jenkins: called the jenks file runner.

A

Which basically allows to do it's a portable engine that can execute jenkins pipelines anywhere, basically as cli or so whenever you have support for docker or other container runtime or whether you have java, you can execute jenkins pipelines in cli, so he's just give happening executing pipeline and basically the idea for this project is that you can take this engine and they use a job executor service, so basically put this engine right inside job executor service, so that you execute jenkins pipelines uh inside the captain execution plane uh as a single shot by jenkins pipeline execution.

A

So this idea is easy, and actually you can do it by a few tweaks here, which is something I suggest to try. But then there is, there are infinite opportunities for improving developer experience. So, for example, uh streaming some results back streaming uh reports returning unit test reports or static analysis data as quality gits as sli providers back to kaplan, et cetera, et cetera.

A

So for this project uh you could start from creating a simple poc, which is probably a few hours of work, uh gluing things together and then you can think about developer experience and how you could improve the things by providing better integration, layer.

A

Is anyone still following.

A

A

Yeah, so basically jenkins today is just an example of uh let's say, common line executor so jenkins file. Runner is quite established as a tool right now. Obviously we can talk about various things: how to pre-configure the images etc. We can talk about that if you're interested, but uh the most straightforward example would be to just integrate the bits with each other and see how it flies, and then you can iterate on that and maybe create something more exciting from user experience. Standpoint.

A

Okay questions comments.

A

Looks like none so any other projects or questions we forgot to discuss today.

A

Okay, then again, we have only two weeks left if needed, we can organize more meetings. So tomorrow we have captain developer meeting where we can spend some time discussing jsoc as well. Next week we will have community meetings, uh but do not wait until these meetings. If you have any questions ideas, you would like to discuss just drop them in the chat and it's the best possible way.

A

So um bright mac would you like to add anything.

D

No just um thank you. Everyone for coming, it was really good to everyone was asking some really good questions and we're really looking forward to to spending time with you and remember what I said. The only way to the only way to get involved is to get involved so jump on slack, say, hello and, and just get talking men's like.

C

Plus one yeah- and I just wanted to do that on the related note, any question you asked here or on slack or anywhere you can search it publicly. It's a country contribution, recapping, loving time same questions later on asking very few questions.

A

A

Okay, then thanks everyone, so I will publish the recording soon and I just realized that I still owe you recording from the previous week, because I wanted to copy edit it a bit and then forgot to click the publish button.

A

So I will get the recording from the previous week published so.

A

So yeah thanks everyone and see you in the chat.

E

Thank you. Everyone thank.

C

E