keptn Community, 8 Oct 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keptn Community Meeting - October 8th, 2020

Description

Diary - what happened last week?
Integrations page in Bridge
GitHub master branch renamed to main causes problems with keptn create project
CommunityBridge - Mentoring Program from CNCF: selection has been made! Great to have you onboard the Keptn community.
Hacktoberfest - we are part of it as well as contributing.today
Keptn mailing list / google group: https://groups.google.com/forum/#!forum/keptn
Good first issues: https://goodfirstissue.dev/ or start here
Noteworthy Issues & PRs (Review)
For each PR, please tell us if you updated the docs, or in case it was not necessary please say so.
Florian
Dynatrace-service: Automatically sync services detected by Dynatrace https://github.com/keptn-contrib/dynatrace-service/issues/191
Bridge: Show true count of compared evaluations https://github.com/keptn/keptn/issues/2457
Allow to set any sli-provider with `keptn configure` https://github.com/keptn/keptn/issues/2483
Ermin
https://github.com/keptn/keptn/pull/2490 - Integrations page in Bridge
David Dyball:
Test Keptn on Kubernetes 1.19, Update travis-ci and cli for 1.19: https://github.com/keptn/keptn/pull/2414
Note: Microk8s 1.19 has a bug - https://github.com/keptn/keptn/pull/2448
Johannes:
Click on Keptn API - page not found: https://github.com/keptn/keptn/issues/2430
Improve documentation about Quality Gates https://github.com/keptn/keptn.github.io/pull/646
Christian Kreuzberger:
Tell people at beginning of meeting that GitHub renamed master to main, and this can cause troubles with our GitOps approach - https://github.com/keptn/keptn/issues/2487
Docs: Describe Helm Installation Values: https://github.com/keptn/keptn.github.io/pull/647
Docs: Remove distributor for 0.7.x docs: https://github.com/keptn/keptn.github.io/pull/657
Docs: Fix typo for assistant, rework delivery assistant page: https://github.com/keptn/keptn.github.io/pull/658
Ankit:
Improve CLI auth command by including endpoint in output
Roadmap → KEP / SPEC / Issues
KEP 06: The next generation of Shipyard: https://github.com/keptn/enhancement-proposals/pull/6
Work in progress:
https://github.com/keptn/keptn/projects/29
Release-checklist (Christian):
https://github.com/keptn/keptn/wiki/Release-Checklist
https://github.com/keptn/keptn/issues/2404
Backlog Refinement / Next Sprint:
https://github.com/issues?q=is%3Aopen+is%3Aissue+user%3Akeptn+user%3Akeptn-contrib+label%3A%22next-sprint%22
Planning Board: https://github.com/keptn/keptn/projects/29
Announcements:
Rob made an updated video for the Alexa skill: https://github.com/robertjahn/alexa-skill
Juergen and Rob are presenting Keptn tomorrow: https://devopsstage.com/

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

Right, if you want, I can take the first part.

B

A

Yeah hello, welcome to the captain community meeting um today. We don't have that many topics for you in general, but we have a lot of pull requests and I think we have a couple of items for work in progress. We will be discussing with you um first and foremost uh in case you have been well listening or watching or whatever reading the news you might have heard that github has renamed the master branch to main uh to be more inclusive. uh Unfortunately, this causes a couple of uh problems with our create project.

A

If you have get this in upstream, we are in the process of fixing this. I think there is a pull request that will be shown about this, or at least there is a pull request in our repository, and there is an issue in our issue tracker, um so stay tuned, and this we will address this in the upcoming release other than that uh yeah. I think community bridge regarding the mentoring program.

A

Emails have been sent out so feel free to start contributing if you're part of this program and if you're, not part of this program, there's also hacktoberfest, and you can contribute there by the contributing.today page, which is linked here, and we also have a lot of good first issues, which you can find where the good first issue, dev page or if you just filter for good first issue in our github issue, page other than that I'll. Kick it off with florian to show some pull requests.

C

All right, thank you. Let me just share my screen.

C

Okay, so let's start with the first one, the first one considers the dynatrace service and let's just look at the story that was defined here. So basically, the requirement for this uh issue was that we wanted to make it easier for users that use dynatrace to integrate with captain more tightly and therefore one one potential feature we saw here was that we could use the dynatrace service to automatically detect desks that have been tagged according to a specific scheme within dynatrace and then import those services automatically into a dedicated captain project.

C

So you don't have to use the cli anymore in order to create a new service within that project and basically, let's just jump right into a demo that I've prepared. So we are here now in my diner trace, tenant and let's say I'm or actually, let's start at the bridge, to show the status we are starting with. So we have a dynatrace project here in captain with the stage quality gate.

C

So that's the prerequisite for this feature to work. So this project with that stage, has to exist, and now I want to import a status into that project and use it in my captain installation.

C

So here I have, for example, the service bridge, which is one of the cap and core services, and it already has the key value tag captain service with the value bridge attached to it. So uh this stack has been automatically created with uh automatic tagging rule, and now I want to enable this service to be used within captain for evaluating quality gates, for example.

C

So previously, if I wanted to do this, is I would have to use the cli and send a captain create service, minus minus project, dynatrace command, but now what I can do as well is, I can add, attack which is called captain managed, just have to write correctly kept managed. That looks better, add it and what the dynatrace service will do it's it is. It will check automatically given a specific interval. In my case, I've set it to 30 seconds.

C

It will check every 30 seconds if any services with those tags exist.

C

So, with the tag cap managed and with the captain service and then use that tags in order to import this service into captain and in the miracle of time, if we refresh the bridge yeah here, it is already so the dynatro services automatically created the recording uh service. uh You can ignore those uh three events here, because they uh are from a previous test run where I've tested the feature and then afterwards deleted the service in order to check.

C

If everything is working and what I can do as well is, I can immediately start an evaluation using the values provided by the dynatris metrics api, because the dynatoy service not only creates this new service, but it will also.

B

D

C

Sli config file for the dynatrace sli service that describes uh how a certain query or certain sli values can be retrieved.

C

So it will immediately have the capability to retrieve the response time, the error rate and the throughput, and now you don't see my console, but I just sent a new start evaluation event, and if I refresh the page, we see that event here and we see the evaluation already in that case.

C

In that case, no evaluation has been done because I forgot to deploy the sli provider.

C

But yeah, if it was up deployed correctly, it would have done the evaluation immediately. All right are there any questions about that feature.

C

All right, if not, I will continue with the next one. So my next issue that I would like to show you is the concerns the bridge.

C

So in a previous pro request, we added the capability within the bridge to show the previous evaluation done. Events that were used for the evaluation of the current evaluation of the current quality gate, and here we have the issue that the the amount of the number of previous evaluations wasn't uh shown correctly here, because it used the fixed value from the slo jammer.

C

So it used the field, compare with n results and it always showed those and, for example, here free re evaluation done events, but in case only for example, two past evaluations were available.

C

Then this information wouldn't have been correct anymore, so it will would have shown oh free always, and what I have done here is in my per request. Let me just show you here so basically uh before yeah, so in that case uh we had no previous evaluation done. Events and the the information displayed here was compared with the last three events and that's, of course, not correct.

C

So that was before the request and with this pull request with this fix, we now display that correctly. So now we can see the information here compared with last zero past evaluation, because that one here was in fact the first evaluation we did for this service and also if we only had one previous evaluation done event. We also showed it here all right. Any questions about that.

C

All right, if not, let's continue with another small pr, but I think uh rather helpful one. So the story here was that, in order to specify a certain sli provider for a project, it was allowed to do so using the captain configure monitoring command.

C

But we previously had a checking place that own the checked, the parameter, for example, this my sli provider and checked whether the provided value was actually supported providers, so either prometheus or dynatrace, because those are the two sli providers that were available at the time. And now the requirement came up to allow any type of sli provider to be passed on to the configure monitoring command.

C

And now we have removed this check. And now you can provide any value to that command and the lighthouse service, which also receives that command, will then update the correlating config map and then use that provided value for retrieving or for sending out the sli event.

A

So no more typos allowed when writing dynatrace or prometheus.

A

That's not a lot anymore.

A

Is there actually a way for us to um like? Is there a way for a user to see that nothing is happening? If you have a typo, let's say if you say prometheus with out the e in the middle so prometheus or something because what will happen if you have to typo, you won't get a reply from the service right.

C

Yeah currently, honestly, you won't have any feedback other than checking the config map that has been updated by the lighthouse service if it was updated correctly, but yeah, maybe in in a future version of captain, we could uh think about moving the responsibility of handling this command, so currently it's being handled in the lighthouse service.

C

So maybe we can think about moving that, for example, to the api, which can give immediate feedback, because the api is called via http and currently the lighthouse service is covered from the caller, more or less because it receives the event for that inspiration.

A

Yeah, I think there is a follow-up issue, or at least there's a feature request by someone that asks if we can have like a command to to validate or to check if the monitoring provider is is properly configured. Maybe we can. I think we should add that for zero, eight or something like that.

C

Yeah absolutely.

C

Cool all right, if there are no further questions, I'm through with my prs and I went over to.

E

Yemen, thank you very much. um I'm going to present you uh one poor request from my side. It's a new page in the bridge, uh the integrations page and uh yeah. Let's see what the requirement for this was. Actually we wanted to uh show some uh information uh inside of the bridge, how to integrate yeah with the cli the api, or also how to use custom integrations and therefore I will show you the live demo. I hope it's deployed uh now when you click on a project.

E

You have a third tab here: uh the integration tab, and here you can see uh yeah how to integrate to the captain, cli or captain api and those are links to the docs uh to the cli and the api, and you also have some instructions in this case, for example the cli. You can download the captain cli, we have the instructions how to authenticate, and we have also one use case here- uh how to trigger a quality gate evaluation, for example. So you can use this command.

E

uh Yeah fill in your details like the project stage and service that you have and you're good to go. We are looking forward to provide more use cases here.

E

We also have the example with the api, where you can copy the api token here by clicking on this text.

E

We have also the link to the captain api, which will show you the swagger ui, and also a use case how to trigger the quality gate, evolution by uh yeah, calling the uh api link uh or actually sending a post request to this api. um Additionally, we have uh some more integrations, um which is uh so this. The next part is loaded dynamically here. So that's why we placed here a button.

E

We also show the information to the user, that we are loading this from an external source and because of that there will be also collected statistical data in the background. But when you click on this button, um yeah, we load an additional html page, which shows all the available integrations, which we have uh now supported. For example, we have here azure devops, captain integration, also gitlab integration, jenkins, jenkins integration and the links to the github repository and also, for example, if we have a tutorial for this use case, then also a link to the tutorial.

A

Yeah, the link for gathering data details can't be found. That is, is that to be changed or do we need to change it in our docs, because that is, I think, not working.

E

You're right, that's not yet deployed, since this is a yeah in development version. This needs to be documented here, of course, um yeah. That's uh still an open task here, all right.

B

uh I mean, I think, what there's also one remark that we should add uh from a security perspective, because when the feature of of showing the api token is disabled, then also copy api token to the clipboard does not work. Of course, that's already considered here.

E

So in case that the feature with the api token is disabled, you won't also say the input. So this this um yeah copy functionality here in the user menu and the same also applies here. uh You would see a link to the documentation how to retrieve the api token, instead of this copy api token to clipboard link.

E

So that's already considered here.

E

It will it would show a link to the documentation instead of this copy.

A

The clipboard function um we have a question from the community asking is: is this integrations page more like an online help page or what's the purpose of it?.

E

uh It's not a help page since uh yeah. You cannot post your questions here, but it's a page with um yeah a good to start point. So um yeah you have the captain cli here you can download it directly. You will also have some use cases here on the api token and how to access the api but yeah. The focus actually here is to show the integrations that we have and all other use cases are actually covered in our tutorials page on the captain homage.

B

I also think that we should put focus on the on the bottom part where the custom integrations are loaded and, as we have a growing ecosystem around captain more and more services, captain services will be available, hopefully in the future, and then this is also a place where we can promote external services. Someone can install right now, it's just a list of three items but going forward. Hopefully it will come up. um Yeah, bigger picture of integrations that are available.

A

So, who needs to be contacted if, if one wants the integration listed in there, or should that just be done by a pull request or what's the process for it.

B

The idea is that we all that we list those integrations that we also have on the captain.sh page here and in this part, meaning that when someone has an integration, it should go first into the official documentation and then we also consider it as moving it into this part. Because then, all captain users will get kind of notified that there is new integration out there.

A

E

Cool thank you for the questions and uh that's all from my side. I'm handing over to david.

A

Yeah, so I think david you're on the call, do you want to quickly talk about your pull request.

F

Yeah, hey everyone, uh it's good to good to be here and hey everybody, um so I I basically just did a quick test of um the latest kubernetes version, 119 using uh kind, kubernetes and docker. um I know that it was tested with micro gates and uh mini cube and all of that, but um the easiest tool at hand was was kind.

F

So I figured I'd jump through the hoops and get that working um hit a couple of issues, but I think it was mainly because I built from master and therefore when it, when I did captain install it installed all of the latest tags for the docker images instead of actually tagged versions. um But once I applied the helm template with the latest version, it all worked fine, so yeah. It was a good first issue.

A

Well, thank you very much, uh I think so far it has been working. I think we only hit like one issue with, I think, was micro, cuban editors or something, um but that is a documented, well-known issue for micro q and it is 1.19, so we just disabled it for now, and everything else, I think, seems fine. If I remember correctly.

F

A

F

um It turned out to be a breeze to get it all set up and that's actually the first captain install I've done. I've been lurking for quite some time, so it was a good first issue for me as well.

A

All right, yeah cool yeah thanks and if you.

F

A

To keep contributing, obviously um yeah feel free.

A

All right, uh I think, next up, johannes.

B

And it's me all right, yeah. I have two pull requests to show you the first one is a quick bug fix um in the in the bridge.

B

I touched angular at least for a little bit, and it was a really really a minor, because the captain api link was broken. There was a slash at the end and this slash is not supposed to be there and I just removed it, and then I flipped the links to have it aligned with the buttons and then the labels or the the fields are down there to have to keep an api on top. Then the cli and then also the captain api token and the cap now's command yeah.

B

I think I don't want to talk more about that, one as it's really just a quick fix.

B

But I just want to show you more about the improvements in regard to the documentation of the quality gates uh there. We added more content also to help users to to make their way through the quality gates, and when you now go to the documentations page here to the quality gate section, then you get now how to get started with quality gates.

B

Where you first get a quick introduction. What is actually a quality gate, then here an example of how a quality gate looks like at the end, also with an explanation of the different elements that are used in this evaluation tile to explain you the outcome of a quality gate and yeah. I referenced this in the screenshot, and here you find, then the explanation or the question uh the the item is entering.

B

Then it's about defining the project stage and service for your quality gate. Here it's explained based on an example, then you actually do that yeah with captain and when you're done with creating the project, then there is a four step guide or how to where you finally then trigger your quality gate so that you can really ex experience. What it, what it means- and this is just a way to get started and for more details in working with quality gates and also in integrating those into your existing pipeline.

B

You will then find a section below- and here I have it explained, based on the viewpoint or from the perspective of the cli and also from the api, and there are always the two actions of triggering a quality gate and fetching their evaluation result.

A

Maybe some feedback on that quickly. I jumped across this today, so I I saw that and I was like okay, I see step one to step four and then I don't see any commands and I'm not. Maybe it's just me, but if I see step one to step four I was expecting commands. So maybe the steps I don't know, maybe it's wrong wording in terms of steps down there or.

F

A

It needs to be moved to the other. To the second part, I don't know.

B

A

B

A

What I mean step sounds like something I can actually actually do and not just read.

A

But I understand the separation of explaining what a quality get is versus executing a quality gate that makes sense.

B

Yeah, the second part is really focused on hands-on on working with the quality gate, and here it's more like getting an idea of what I have to do to get to the point where I can then trigger the execution, but uh I totally understand also your feedback, but I don't have any idea for getting a better word right here.

B

A

Let's, let's take it offline, maybe let's take it offline.

G

A

Has an idea and can maybe tell us why it's like.

G

Maybe it could be prerequisite, so a pre-requisite is an slo, an sli and also can project. So maybe we should formulate it this way.

B

Edits to prereq is a project service. Then the slis must be available, the slo must be configured and that that's not a pre-work, that's really, then the execution or the kickoff and where you triggered, and they qualificate good idea.

B

Yep all right, thanks for the feedback yep, that was it from my end, then I hand over to christianity.

A

Thanks, so uh I have a couple of doc issues that I want to present today, um wrong button.

A

There we go um so my first thing I already said at the beginning of the meeting like the github master, main stuff, um that's being dealt with right now, then the second thing or the next three things actually are dock issues, I'll just go over and quickly.

A

The first one is describing the captain install options that we have with our helm chart.

A

So, basically, if you go to is already deployed, so if you go to install cli and captain and you go through all those steps, then you see after choosing the ingress or load balancer or port forward. You kind of scroll down- and you see this part with advanced installation option for the helm chart and if you click on that there we go, you get a table that has most of them.

A

I'd say listed, might not be all of them, but these are the ones that make the most sense and we can easily extend it if we have to. um For instance, the cli download link, which is a configuration option for the bridge and bridge, is obviously part of the control. Plane can be any string and the default value is null and basically yeah. If you set this value to something else, then this is going to be displayed in the in the user drop down menu on the top right.

A

um Also, for instance, let's say every control plane service um has the tag called or the config option image dot repository.

A

So if you ever want to deploy a different version of let's say bridge in this case, you can set control plane.bridge.image.repository to, for instance, docker dot, io, slash whatever, and by default it's docker dot, io, slash, captain and the service.

A

This is, for instance, something that's used and also described below. If you want to execute helm, upgrade without internet connectivity, then, for instance, you set control, plane, dot, bridge dot, image repository to like this, so we have some examples for it below, and it's described above what it is any questions about this.

A

Okay, the second one I was working on is a prerequisite for the current release. It's basically a cleanup task, um so I went through the docs and I saw the term distributor is mentioned a couple of times: um that's not bad or something, but we don't use the distributor as a dedicated deployment anymore in kubernetes, but we use it as a sidecar to our micro service that we deploy.

A

We had a couple of instances where we referred to the distributor for checking if everything was deployed properly for troubleshooting the installation, so I just cleaned that up and it's gone and the last one is still work in progress. But I want to show you a quick preview so um similar to what johannes did for improving the documentation for quality gates.

A

I started to actually kind of rework the documentation for our delivery, assistant or manual approval and there's a pull request that will need some review and some comments, um but essentially I added a delivery assistant page in the continuous delivery section, um because this is right now uh the only place where you can use the delivery assistant. As far as I can know. As far as I know and yeah, so I've copied the part with bridge and extended it with some more information.

A

A screenshot like this is an approval, triggered event in bridge and how to now send an approval finished event via the cli via bridge that was already existent or even wired, api um yeah. So for the cli I added the command and some of the output that's being generated same for. If you want to send the approval finished um directly via the cli, you get like the option how to how to use it or which one you want, and whether you want to approve and decline it and for bridge.

A

As I said, this was already there. um I just copied it over, and this one is new for the api. Obviously, if you want an integration without the cli without bridge, you can just use the um what is it called v1 event endpoint of our api, and you send this cloud event directly to the api, and I added this here. I think it's not yeah. I think it's correct now.

A

um You need the sr captain context and you need the id of the ex of the previous approval triggered event and the essa captain context goes again into the sr captain context and the id goes into the triggered id. So if you want to build an integration with that, I think that's a valuable information and I hope it's correct.

A

Okay, any questions or comments.

B

About this, there is no question on that issue, but rob was asking um if there are, uh when you do a helm, install and you change the setting.

B

um Does this work and I mean: uh are there any already? Are there any um gotchas or tricks tips and tricks to update a helm, chart.

A

So what should work? At least that's, I think, would be documented. You can always reinstall. Captain is maybe the wrong word, but you can re-upgrade, I think the same helm chart and just change the the values as far as I'm aware, but shouldn't be well. There's no problems known to us so far. So if the installation worked the first time, it should work the second time, but maybe andreas has some more input on this.

G

Right, so you always have to be careful um if there are any database- migrations, of course, which we do not have for 0.7 to 0.71, um then I have an upgrade always works basically, and you can even reuse the values which you have set by 10 upgrade minus minus 3 use.

A

Yeah, I think, uh rob also made a point. I just read the comment here. We should maybe add last updated or something or date added to here and maybe at in the end this this might not be a list that is really specific, but this might be a list that's outside of the release, and we just add like uh when this value was available or not available.

A

So, for instance, this one has been available since zero, seven zero, uh but the bridge version check enabled, I think, has just been introduced with the release that we're currently working on.

A

If I'm correct, true yeah, that's true um ankit is asking: where can I find the helm chart templates um so the help chart itself? We publish it to storage.googleapis.com captain installer, but if you want to have the source files for the template, they are in our repository in the installer sub so directory.

A

This is the installer sub directory and there is manifests, slash captain, and this essentially is the whole hem chart with um the sub charts like there's, continuous delivery and control plane and then control plane again has a chart for mongodb. um So I hope this is yeah, where you should find the the charts.

A

Maybe also quite important, we do publish the index yaml on google storage with the helm chart the latest or the master branch release is basically always in sync, and the release branches itself should be in sync. So for zero, seven, zero point one and the upcoming point two release: there is a separate template or a separate release. I think on the on the storage of google cloud, but that's the setting you can control with. Where is it there.

A

I think correct me: if I'm wrong, if you say version equals master, you get the current release from master.

A

Okay, that's it from my side.

B

Next, one is actually um is ankit anki. Do you want to present your pull request and your the issue that you were working on this week.

D

Hi hi yeah, I'm like haven't I'm working currently on one of the issues that.

D

D

Currently, working on this issue to add the post installation steps for the service if the service is not port or door balancing service, so I think testing is already reviewed and left some comments. I'll try to close this up today, other than that is one more issue that I was that is actually merged. That surprise. Supra was working on the issue to uh provide a captain config a file uh environment variable so that we can pass api, endpoint, uh api, endpoint and tokens directly from there. Let me show you so.

D

D

Yeah this one so right now this pr is already merged, but it the pr doesn't support the mac. Doesn't support for mac and windows, so surprise editor code for the uh integration with linux. Only so I have extended this uh feature in for mac and windows as well, mostly I'm. I was working on these two pr and that's almost close. We will close the pr as well thanks thanks a lot.

D

One more thing, while installing captain cli, so I was facing some issues. Let me show you so this is one service or api service. uh It contains distributor which has some topics for net cycling, so that service that container, particularly that one of the containers on getting up because it doesn't contains topics, so it was uh going back in cash back loop, so I have removed that particular container as of now and that fixed my issue. But I would like to like uh to raise the pr or the particular issue related to that.

D

If, uh if, if that's a problem with the api service,.

A

Yeah, uh actually, that's exactly the problem that david was uh running into. If I remember correctly, I think it was the api service running into a crash back loop.

D

Yeah yeah so that doesn't contain uh topics that is being required for it. So as of now, I have removed the container itself, so it fixed my issue, mostly. I was working on these things.

A

Yeah, if you can make a pull request for that, that would be nice thanks.

A

D

A

Don't need to create an issue uh like really just make the pr, I think, was it like one line of change or 10 lines of change. Sure no problem, thanks.

A

Okay, um is anyone else here that has something to present that we didn't mention yet.

B

In the attendance list, I can't see any member um who was working on a pr this week right.

E

A

Yeah, let's jump to the working items, do you want to present johannes yep.

B

I can share my screen now all right yeah, I moved already all those that have been presented to the to the done column. Let's just quickly go over those that are in ready for review or finished by someone else um prepare release, 0.72.

A

Yeah, so I I think I've completed the bullet points from the really uh from the issue. um As it was a meta issue, there was no real prs, except for the one where I removed the distributors.

A

In my opinion, we can close it.

B

Okay, yeah, then I'm going to close it and, as you mentioned it in the demo, we can also move it over to the done section.

B

Then here is one issue that is that florian is working on it's all about the the renaming of github, already changed master to main.

C

The status here is I've already implemented a fix for the current release branch for zero. Seven two, that's already been merged, so that's not blocking us anymore in the current release. But the reason I have not shown it yet in the community meeting is that I have implemented a different approach for the current master branch that is not merged yet, and I would like to discuss both approaches next week.

B

All right, perfect, then amen that one uh that is related to the integrations page. We leave it open since there is still the documentation missing right.

E

uh Yeah, um I will integrate the documentation page and uh yeah, let's uh close it after that is finished, perfect.

B

Then uh there are three actually three pr's from arno cool yeah. Let's just open doesn't.

A

I can talk about them if you want, um so if you can open the pr, I think it's easier.

A

Yeah anyway, this one is also good, as uh sercho has created. This feature request for us and cool has finished it thanks santa cool. Essentially, uh we want to present the captain server version um in the captain version output of the cli and yeah. I think he's shown this to us last week in a short demo, but he's made some some changes uh since then, but uh it has been merged. It works fine. It's now basically querying the captain api and using the swagger yaml, which contains the version info, and this is the primary output.

A

I think it's working nicely. I've tried it out so nice feature great job and.

B

Yeah thanks a lot.

A

Then the next one I see this is merged already, as well. um Yaml input support for your eyes, maybe florian! You can help me out here a little bit I'll do the intro. You can do the conclusion.

A

So, uh as you can see, this is a feature request by adam and he was asking whether we can supply the shipyard via a url instead of having it on the local, hard disk and anukul also implemented this nicely. As far as I can tell, and maybe florian you can add some details. If I miss something out.

C

uh Yeah you pretty much described it correctly, so anukul made the request where he used a library- I don't know the name yet, but basically this library he used, provides a function to determine whether a certain string is a url or not. Based on that the result of this check, he either continues on to use the provided input as a local file path, as it's being done now or if it is a url, and he sends a http, get request to that.

C

Provided url retrieves the content of the file and then uses that to to create the project.

A

Nice should we change our docs johannes. I mean like for creating projects and and stuff, because this is a nice way to not having to check out the examples repository, for instance,.

B

Yeah, that's true. uh I I mean I expect that it's documented right in the code, at least to explain it. Yeah at the command.

D

B

Should also think about adding it to to an example.

A

I mean I mean essentially this allows us if we have like three or four shipyard files lying around somewhere to just reference them via http.

B

Yeah, that's true yeah. I think we should. We should sell this feature to users and how can we sell it by documenting it and we should definitely think or we should provide documentation here yep. Let me take a note.

A

Yeah thanks for that feature as well, and the last one is ci include pluto to automatically check for deprecated, kubernetes versions, um I'll just quickly talk about it. We don't need to look at it. Essentially kubernetes deprecates certain api versions from time to time.

A

That's probably a good thing, um considering how many api versions they have that contain the word alpha or beta, so pluto helps us in detecting those deprecated cumulative api versions, upfront and prints as a warning or essentially, the ci task fails if we check in some code with a deprecated api version and in the future, this also helps us. If we say we need to support a new uh cuban leaders version, then we can just use the newest pluto version and configure the kubernetes version in there and then yeah.

A

It should check for deprecations um so yeah. It fails for us in ci before it fails for the end user when they try it out. So that's a good thing.

B

Yeah there was now actually a concern coming in from david, as we talked about the feature of of pulling the shipyard from an external url. uh Do we support https and basic authentication.

B

To authenticate against github.

A

We have not tried this out in this case, um but I think so far this feature is only available for the shipyard file, so uh I I don't see a huge demand for having basic aud in there, but obviously, if we extend this feature for more than just shipyard files, let's say slos slis, then yeah. We should probably redefine to also include authentication, headers and stuff.

D

A

For now, I'm pretty sure this basic art might work, but I'm pretty sure it's not defined to work so, um but if you want, you can raise an issue on github with that and we'll have a feature request for that. Obviously, and we can tag it as a good first issue as well.

F

Yeah, I think the main use case is just onboarding projects um having having sort of ci itself to onboard projects or automation to onboard new projects. The first thing that jumped to mind when I saw that feature.

A

Yeah, but it could also be for for demos or for workshops if you just want to quickly create a project onboard the project whatever um without having to check out the git repository might also be nice if it just works out of the box, but yeah there might be some pitfalls with basic authentication.

A

Next thing that comes to my mind would be proxies if.

F

I'm not mistaken, I think github are actually deprecating the basic goals anyway. So all right.

B

Okay, then, let's go on uh the next issue that patrick was working on was the copy button in the bridge um yeah. It's now pretty handy that you get a button uh when you open the bridge, and this button allows you to copy the content of the slo file, as he shows here in this pull request, you have now recopy in the close yeah. It's it's pretty neat when you want to get the content.

B

And patrick did a second pr in this week because he did an update of of the readme dot md, where he now reference to the kubernete to the k3s tutorial to get instructions for for getting started. When you go to captain captain.

A

Yeah, that's a nice addition because it helps if you want to develop and you don't have a kubernetes cluster at hand. Then this is just copy paste and it should work at least on linux and mac os, and I guess it also works on on windows. If you have the uh subsystem for linux but yeah, um maybe the next two I can do just a quick uh info. So uh we've had several contributions.

A

Several contributions for our make file um thanks for those they will help us in making the process of testing things and trying out things a little bit easier. So we have uh supratik. I think he's demoed this before he did some changes and we merged it. So that's fine and the second pull request. uh I don't remember the name: can you hover over it.

B

A

It's yeah, it was a new contributor. I think he isn't here, um but um he provided us exactly.

A

Thanks juan um he provided us a make file target to build all the docker files at once, which is quite nice because, right now, if you just want to build docker files and push them to somewhere, um you don't have a way of doing that, because you would need to dig in our travis ci scripts and, let's just say, that's an unpleasant experience for everyone involved and the makefile just helps in centralizing this, and one of the next steps will most likely be to move everything from travis ci towards the make file, so the developers can use the same tools locally than what we use in the ci.

A

So thanks for that and.

B

Perfect then, let's go real, really really quick over those that are in progress um andreas. Can you give us an update on the first two issues.

G

I'm working on this so, as already said, we have postponed 0.8 a little bit and those are two issues regarding 0.8 all right.

B

Then we have here one from andy, but he's not on the call. So pratik is not on the call. Then here we have an improvement of the documentation.

A

Yeah I've shown that.

B

You showed that and is it already merged.

A

No, it's not much that I. I still need some feedback and I need to copy it for 0.8. um I was waiting for feedback yeah.

B

All right and then we have here their release, 0.72.

A

B

Issue, that's in.

A

Progress in progress.

B

Then ankit is not here, but he wants to work on improving the post installation steps.

A

Yes, there is a pr already open with some comments and I think he's updated it. So I think it's it's our uh task again to review it.

B

Okay, okay, then it's on our end and then finally, here is one uh for it's a it's a bug. It's the incorrect mime types in the bridge um I mean: do you need help there.

E

um Actually, this time I yeah I had this week, I had the time to check that one and I'm working on it. It's the case with uh when you support forward on the bridge and you re-load the bridge on a specific sub page. The base url is set wrong and therefore the javascript and the style sheets cannot be loaded correctly. I'm working on that one and yeah.

B

We'll fix that okay perfect, then we skip the next column as they were already discussed uh last week and let's go to the backlog yeah here I want to present you one issue: it's about not cleaning the quality gate evaluation results, it's more about invalidating, invalidating evaluation gate results, because there was the user request coming in if it's possible that we can kind of mark.

B

One evaluation, as not valid like shown in this example here- is one that is an outlier compared to the others, and this user wants to ignore that particular evaluation result from from the process of doing the quality gate evaluation and, as you can see here or can read here on this issue, we came up with a conclusion or with an idea of sending an sh.captain.event slash, evaluation invalidated, and this is the payload of this event and with this event, you can then are mark it as not valid, and then it will be excluded from the comparison strategy that you specified in your slo file yeah- and there was a lot of discussion going on to came to this conclusion to this idea.

B

But uh it's now really their approach that we also want to provide an additional endpoint. That's that is added to the slash events endpoint followed by the type and then the the actual actual type of the event itself.

B

And under this end point you then, can use a query parameter to exclude invalidated evaluation results so that you don't have to fetch all evaluation results that are valid and those that are invalided and then do the match. This will be implemented by this endpoint and it's not already a long story, but we also need to reflect this in the bridge, meaning that you will then get the possibility to select one particular evaluation result that you want to mark as invalid and therefore the wireframes are still missing.

B

Yeah quite an uh a long task, a long story, but is someone up to work on that one? The next week, yep.

C

So I will do the implementation in the back end, all right, cool thanks.

B

E

And you can assign me also on that issue, since I have to provide some uh suggestions on the ui.

B

Perfect in the break, when you need help for the wireframe, please let me know, of course, okay yeah and then uh we are now getting back on track working on the captain 0.8 release and I now move those items to the backlog again that were already started a couple weeks ago.

B

It's the nx. If I have the captain's bridge um and then also we already discussed that the project board, which is now the main page in the bridge, should be splitted up into uh smaller components, and this is also considered as a refactoring task.

B

I mean uh um I.

E

Have already been working on the an xfi uh issue, um yeah I'm already assigned, and you can move it to the assign and commit it. Okay.

B

Thanks a lot I mean, I will also move that one over as we already assigned you to that issue.

E

Yeah, of course,.

B

Okay and and the same is true for um the cli support of the new shipyard format. uh Florian, you raised your hand a couple weeks ago to work on that one. Are you still okay doing that.

C

Yeah sure, looking.

B

Forward to perfect.

B

And then, finally, I have here one issue that is about providing the documentation of the services that are now possible to run in the execution plan, as we are yeah now supporting the new shipyard format. Also, there are changes required on the services that execute the new shipyard, and this should be the the first documentation for that change.

B

Here, referenced the content that um refers to the external services, and then here are the tasks or the things that need to be considered in the.

B

Documentation has someone time or wants to work on that.

B

Issue: okay, then uh let me just leave it where it is and um when someone finds time in the next week, then please feel free. Please assign you or leave a comment there, and then we will assign you to that one all right.

A

B

A

Having any good first issues or any other issues they want to work on, I mean you can always write on the issue and we'll assign you. But if you want to say it now, then.

B

It's also a perfect time for me.

A

All right, yeah.

D

We're also trying.

A

We're trying hard to add some new issues, especially good first issues, but if you want to take another one and you're not sure about it, just ask questions on it. It also helps us to understand what we are missing in our descriptions for the issues.

B

All right, yeah and and to close this community meeting for this week. I have to announce two two announcements uh that came in during the meeting, because rob did an update of the video for the alexa skill he this one is you can find here. Please follow this link to see the new, the new or the updated version, and rob and jung will also present captain tomorrow at the devops stage.com, which I think is a meet up or let's just go there.

B

It's actually a conference, an online conference and there rob and captain will present rob and and urine will present. Captain register it's for free and you can learn more about captain when you want all right. Are there any questions, feedbacks feedback that you want to bring in this week.

A

I guess we're in over the hour as well. So uh let's conclude it.

B

Let's conclude it all right, then, uh thanks for joining this week have a nice day and see you next thursday, bye, bye.

F

F