keptn Developer, 18 Feb 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keptn Community & Developer Meeting - February 18th, 2021

Description

Meeting notes: https://docs.google.com/document/d/1y7a6uaN8fwFJ7IRnvtxSfgz-OGFq6u7bKN6F7NDxKPg/edit

Learn more: https://keptn.sh
Get started with tutorials: https://tutorials.keptn.sh
Join us in Slack: https://slack.keptn.sh
Star us on Github: https://github.com/keptn/keptn
Follow us on Twitter: https://twitter.com/keptnProject

A

Hi everyone and welcome to this uh captain community and developer meeting it's february 18th. uh We have a couple of new and exciting news. As usual. Please just add your name to the meeting document, so we can keep track of uh attendance.

A

um There are yeah one one point also, um so we are following the cncf code of conduct. So uh please basically means uh please it's not be nice to each other in the captain community. um We don't have any discussions left from the last meeting, so we can already skip to the next point. um Right now we are in the process of reviewing applications in the lfx mentorship program. The application period closed last week and the reviewing period is until next week, so we received a couple of applications.

A

Thank you for that for both projects, actually so for the project for improving the for adding functionality to the captain, cli and the other project in uh improving um the connections between promethouse and captain and improving how how both our projects can work together. uh So thanks a lot and uh we will be reviewing these applications and then reach out directly to the uh mentees.

B

A

B

Know by the governance.

A

And the the portal of the linux foundation and this year, um one other part before I hand over thanks, also for all the feedback that we received on the website. We now moved the community page of captain to the captain website, so it's now on captain dot, slash community.

A

You can now find all the community content, um including a contributing guide. All of our meetings.

B

Some faqs uh all of our community rockstars, so we find everything.

A

On captain.sh, which is not really kind of a one-stop shop uh for uh everything around captain and uh what we also added is a couple of resources. So there are a lot of resources out there um on like. There are youtube videos. There are blogs. There are podcasts, there are public presentations and slide decks out there, so you will be finding a lot of those resources now directly in captain.sh.

A

It's a little bit structured into like more hands-on, focused uh webinars adoption highlights, like the user group of captain that we do every month.

A

So we have a lot of information on there, how you, how users and the captain adopters how they are using captain, what they are doing with captain so make sure uh to check this out. It's on captain.sh with all the new content. Now um so yeah. If you have more feedback.

B

Just let us know we are happy.

A

To to include this as well and just open the pr and we are happy to to review it and get in so with that, I think I'm already in the middle to the next part. um I assume this is uh johannes, that uh yeah, okay.

B

A

Yeah, I'm handing over to you, do you want to take over the screenshot? Yes,.

B

Please, okay, what I want to show you today is a little bit of of an outlook into the future, because yeah after the release is before the next release, and we already structured our thoughts about what we are going to build with captain 0.8.1.

B

You can already take a look into the milestone of captain 0.8.1, where you find already issues prepared for that and just want to give you a very high level overview of what will be addressed coming up the next month.

B

First openshift4.x enterprise readiness is on our agenda, meaning that we will run then docker containers as not root users and also want to provide um integration tests for openshift 4.x yeah. This should be this building block is, is maybe a handful of issues that need to be addressed and quite an interesting one.

B

Is that one where we want to drink our own champagne and where we want to get more familiar on delivering captain with captain itself, meaning that there should be one captain that deploys another captain into multiple environments on a daily basis, meaning that, with every commit that the community is pushing to the master branch?

B

A captain should be then updated in a dev stage, hardening stage and finally, then also be promoted to a production environment where we can or have um yeah stable captain up and running.

B

I provide here a very brief summary of this story. More details and also a breakdown of this epic into issues is ongoing. A couple of those issues our already created and a couple of mores will follow, but this is what uh the team uh will focus on um in the next weeks.

C

Is there a definition, or example of how it feels like end to end? So what does it feel like for the end user and what is the result on the target environment.

B

um Not yet I'll just provide here this epic for now, but I want to write it down into the user stories or into a user view yeah following or based on experience. We also make from testing order data zero.

C

uh Maybe a follow-up question: should a captain be able to upgrade itself based on this story.

B

No not itself no uh uh a captain should there will be one captain that takes over the control of deploying others, but this gentleman, who is kind of the master captain, should not update his self.

B

And also important, um based on this epic, we also want to learn more about captain itself, meaning um that we want to understand in more details how much cpu memory are we consuming? How many events we can store and also derive? Maybe some.

B

Improvements for for performance all right and when it comes to the ui, um there are two epics so to say uh for improving the visualization of the quality gates, uh because we have users that have quite of quite long sli names and with two long names. um It's almost or it's it's difficult to use the heat map or to the heat map rings and makes it difficult to use.

B

And this is one thing that we want to tackle um going forward and the second one is also about um quality gates improvements, because we want to give an user um the chance to get a project wide overview of how the services um how the quality of the services is, and I think that this would be nice to fit into the environment screen.

B

But I would also get uh your ideas and your thoughts and that yeah, as we maybe um want to discuss or want to discuss a mock-up in the next or once of the next meeting in here, to see also what you think.

B

But the goal is to give a captain user the chance to see at one uh in one screen how all the services are uh behaving or how the quality is of all of those services services all right, yep. Please take a look at this board uh and when you um have remarks or wanna add some thoughts, just ping me and I'm happy to discuss this with you.

B

All right, this was it from my side today, then I will hand over to the next one in the list. Who is there.

A

Any questions uh directly on this part: oh okay, um that johannes mentioned, as we have two new members to the community joining. I just allow everyone to unmute and talk. If you want to to ask some questions through johannes feel, free to do so.

A

Okay, not yet so uh I think we're just uh handing over now to jimmy all right.

D

Yeah, thank you very much. um I'm going to present the pull request for the bridge today. uh It's the service screen refinement um so yeah. The idea uh was, since we introduced the sequence screen null uh that the service screen can be reduced, and here in this issue I have an old.

D

Screenshot of the services screen, where you have the list of all tasks that happened in this root event or sequence and um yeah. Actually this was the same as we have now in the sequences screen, um but uh yeah. We wanted to have still the overview of the stages. The sequence went through and focus here on, the evaluation uh result and the heat map in the service screen.

D

So as a result- and I will do a live presentation as usually um in the result. Now, when you have yeah the service screen and you select uh sequence that was executed for a specific service, then you see actually some basic information like the artifact and the context. Here you see the stitches and when you click on one of the stages, you see the recent task and only important information like uh yeah.

D

If there is an approval open, then the approval trigger uh card or if there is an evaluation, and actually, I think in this project I don't have any evaluations, then you would see that.

D

Do you I'm sorry, do you hear me still? We can hear you um so I will just look for an example where I have uh evolution yeah here. You will see then the heat map as well, and we also provide the deep link here to see the um sequences or actually that sequence and this deep link is still not working because we don't have the links for the sequences implemented. Yet this will come uh yeah with next week, uh but um yeah.

D

We are working on that, but here you see like a overview of that artifact and the evolution in that stage.

D

Yeah, um are there any questions on that? One.

B

I mean for a user who was used to take a look at this screen.

C

B

I cannot hear you.

C

I think that's, I think it's in his.

B

For some reason, I can't hear.

D

C

I'm really sorry.

B

Okay, um it's just a note. It's not.

D

B

Not a question.

D

No, I can hear it.

B

Okay, um just for for a user who is used to this screen, this person does not lose any information because, with a click on the sequence screen, he gets what he was used to from the old version, but now it's reduced to mac. So to say, by just uh showing the the most important elements of a delivery and of a remediation um I mean what information are we showing when it comes to problem.

D

I can show the sequence screen, of course here we see the the whole uh sequence with all the tasks that have been executed, but uh yeah. Here we reduce it to the uh last task and the status so in case of a problem um you will see. Recent task is problem resolved and actually you should see here the message it's on the wrong place here. It says response time degradation.

D

So the reason why the um yeah remediation um or the problem, uh the remediation workflow was triggered yeah. um If so, you see that the status is the problem is resolved here and if you need more details, then of course you would jump to the sequence screen and have a look uh at the sequence itself, where you would see, then all the tasks that have been executed, like all the remediation pictures.

D

So if there are no more questions, um I'm handing over to christian. Thank you very much. Thank.

C

You give me a second.

C

Okay, all right now you can see my screen, um so I just want to show three pull requests. Actually, um the first one is not so much of a note, graphic contribution, rather a reminder for everyone, that's contributing and writing code to follow this so-called boy scout rule.

C

This hasn't been part of our contribution guide, so I've added it, uh but it should be kind of common sense or at least a good ground rule if you're contributing to open source or even if you're contributing to closed source, that you take a look at the code that you're about to change and to always leave the code better than you found it and I'm addressing this here, because I just want to also point out um some more things about this.

C

We all want to make the world better all the time, but it's just not feasible to do it always so, for instance, when you're fixing a bug, it's not your point or your part to refactor the code around the back. In fact, if you do it and we need to backboard the bug, we might get some merge conflicts and problems.

C

So if you fix a bug, just fix the bug and that's it and also, in addition, take a look at what you're trying to fix or what you're trying to do and make an educated guess on how much you should or could refactor. If you really want to refactor, for instance, if your change would require 10 lines of code, then don't refactor 200 lines of code.

C

It's it's just not balanced, so make sure your prs are balanced and the last thing I want to point out if you file a pr, always ask yourself what changes within this pr. Would you expect if you were the reviewer? So if the task is, please fix the typo, then the reviewer would expect a pull request where the typo is fixed and not where I don't know, 20 other things have been changed.

C

Does anyone have something to add on this or to comment on this.

B

Okay, I think it's.

D

B

Good reminder to to rethink or to think about the such guidelines and best practices when it comes to contributing absolutely and.

C

It's it's true for every open source project out there, not just captain so yeah, um all right next thing. I did some changes with the help of florian and I think, bent to the onboard service workflow.

C

um So it's two prs because one of them is for the cli and one of them is for helm service, and I think this one is the cli yeah.

C

So I actually want to dig a little bit into the code because I think that's that's a good thing to know, um but let me actually take a look at the issue first. um So the idea was this is the wrong issue?

C

Okay, I'll just talk about it. The idea was to get rid of the captain on board service command. We found out that it's not that simple to just get rid of it.

C

The idea was to replace it by create service and add resource, but ad resource had some issues with helm, charts that were supplied in folders, so it was in my pr. I did address this in the cli, but I wasn't able to fully do what was uh in the initial issue definition.

C

Instead, what I did is, I have moved some code for onboard service into the cli, so it has been moved away from helm service and it has been moved into the cli. uh That's essentially part of this, and so how would you do that?

C

Because that's the one point I want to show you would initialize a so-called resource handler, a so-called stages handler a chart storer, you iterate over all the stages or actually you get all the stages and you iterate over all the stages and you store your helm chart in all the stages and I'm showing it is because there is many use cases where it might not be a help chart.

C

But it might be a different file that you want to automatically that you want to store in your captain configuration repository and essentially that's that's the code for it. So resource handler stages handler chart store or if you have something else in a chart, then you don't need a chart store or you can work with the resource handler, get the stages iterate over the stages and just add it to your config repository.

C

This is all implemented in captain utils or in the captain, go utils by the way.

C

Any questions on that or comments.

C

Okay, I'll go over to the next one: um that's the helm, service specific! No, it's not sorry! ah I know I see this is the part. I've updated helm service and the cli at once.

C

So here I've changed the cli and, I think, down below yeah. I deleted the onboard handler in helm service and we have removed some cloud events or some events that we listen for in the deploy service yaml and in the helmservicemain.go.

C

So the second part is actually rather boring. It's it's really just deleting things and then the other issue, the other pr. Here we go, we deprecated the onboard service command, or at least that was our intention. We added a warning now when you use deprecate when you use onboard service, that it will be changed in the future and we added a minus minus all stages to captain ad resource.

C

I think the code itself is not too interesting. um I've provided an example here, so with this, this actually goes into the direction that johannes was mentioning of deploying captain with captain.

C

So if you wanted to deploy captain with captain, you would have to add the captain helm chart and you would have to add it into all stages right now, so this would be the command to do it. So you do a captain add resource to the project captain to a service. I've called the service captain control plane and I will add it to all stages, and this would also do the same thing that I've shown you before it would iterate over all stages and call the resource handler to add it.

C

About deprecating the commands, I think we didn't deprecate it here, because it's just not deprecated right now, but we added some output yeah.

C

Okay, any questions on that or comments.

C

Okay, then, I'm done and I'll hand over to the next person. If anyone has anything else to present or show.

A

uh I think am I no I'm not muted. uh I think uh we are good here. uh Let me just open up the agenda one last time, uh looking at the agenda um thanks christian thanks airman for for sharing this, and also thanks uh to be honest for sharing uh more on the roadmap.

A

We have folks joining from the community. If you have any questions, um please please ask them or if you maybe also uh want to quickly introduce yourself that would be cool, so so the captain community gets to know you. um I have allowed everyone to unmute themselves and.

B

A

Us here otherwise, I just want to let you know that we have the captain slack. uh We have all the information, also in this meeting document that has been shared uh in the in this meeting, uh so we do have uh the site channel and uh which is basically our main way for communication um during the week uh between our weekly captain development meetings.

A

So if there are no more questions and no more um announcements, I think we can close it a little bit earlier than usual um thanks. So much um see you all uh next time and I just want to share- maybe a link to the latest captain.

B

User group, because.

A

This was a really great addition, where a captain user, where they were showcasing, how they are leveraging captain for scaling the performance tester. This is just one thing. I will share real quick here. I know the chat will be gone once we close this meeting.

B

Or you just add it to the agenda, then it will stay. Oh yeah, absolutely.

A

I would like it to the agenda as well.

A

Okay, so then I will edit to the agenda, so you will have to link also for later use. um Then thanks. Everyone see you all next time, bye.

C

C