keptn Community, 2 Jul 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keptn Community Meeting - July 2nd, 2020

Description

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

Keptn accepted for CNCF Sandbox 🎉
Webinar on Managing Keptn with a Slackbot next week: https://dynatrace.zoom.us/webinar/register/WN_Z5yJAp8fRyisFytxgQeBYg
Good first issues: https://goodfirstissue.dev/ or start here
Keptn mailing list / google group: https://groups.google.com/forum/#!forum/keptn
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
Ermin:
Show categories on xAxis in BarChart instead of timeline
https://github.com/keptn/keptn/pull/1995
Implement evaluation board to show Heatmap with deep link
https://github.com/keptn/keptn/pull/1993
Sandro:
Provide a “COPY JSON" button on the Bridge UI https://github.com/keptn/keptn/pull/1977
Andreas:
Introduce an “apiVersion” for Remediation files https://github.com/keptn/keptn/issues/1919
Spec: https://github.com/keptn/spec/blob/master/sre.md#remediation
go-utils: https://github.com/keptn/go-utils/blob/master/pkg/lib/v0_1_4/remediations.go
Openshift installation: https://github.com/keptn/keptn/issues/1951
Christian K.:
KEP 18 was merged (already discussed in previous meetings)
Improvement to curl get.keptn.sh https://github.com/keptn/get.keptn.sh/pull/12
Florian:
Integration test: Self healing with Prometheus: https://github.com/keptn/keptn/issues/1847
Integration test: delivery assistant: https://github.com/keptn/keptn/issues/1750
Integration test: self healing: https://github.com/keptn/keptn/issues/1846
Additional logging of kubectl commands during installation: https://github.com/keptn/keptn/pull/1970
Christian H.:
https://github.com/keptn/keptn/issues/1843 in progress
https://github.com/keptn/keptn/issues/1781 PR created
Andi Grabner (excused)
JMeter Extended Service https://github.com/keptn/keptn/pull/1959
Version info on k8s labels: https://github.com/keptn/examples/issues/103
https://github.com/keptn/examples/pull/106/commits/dc953649cb022ff75c367325ac47bc677bbb3c20
Roadmap → KEP / SPEC / Issues:
Work in progress:
https://github.com/keptn/keptn/projects/29
Backlog Refinement / Next Sprint:
see 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

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 welcome to this captain community meeting. We have a couple of folks already Johnny again. This meaning is recorded and will be available afterwards on YouTube and, as we are now a captain, not a captain by the CN CF sandbox project. So captain me the CN CF asset books project. We are also following the CN CF code of conduct so basically means be nice to each other and if not just please report to the panelists yeah, what do I want?

A

I want to start with a couple of announcements before I hand it over to our 15 first announcement. We have another webinar coming up next week on Wednesday next week. It's on how to interact with captain via a slack bot and our friends from Citrix will be joining. They are responsible for contributing the slack bot to captain it's right now, the sandbox project. So if you go to the captain sandbox, you can already check it out.

A

You can already test it, but if you want to get have like a introduction and the presentation, what it actually is, what you can do with it and please join us next week on Wednesday, we have a link provided here sign up for the webinar its next week, Wednesday 1:00 p.m. Central European Time, 7:00 a.m. Eastern Time feel free to join if you're new to this community meeting or if you're, a new captain developer or you want to get started.

A

Take a look at our good first issues, easiest ways to click on this start here button or also take a look at good, first issue, top death and finds captain and again I want to encourage you to join our Google group. We provides weekly updates on what's going on on captain and provide you with links directly to your inbox, so feel free to join us.

A

Also in our group and of course we have the usual ways of communication like slag and Twitter that you might already know- and you will find all of these links on the top of this document, so I think for announcements. We are good for today and so I'm handing over already to Airmen to show us a little bit about I guess the captain's bridge thank.

B

You very much Ian, yes, I plans to present to pull requests for the captain's bridge today. The first one is changing. The bar chart of the evaluation result comparison, so it's show categories on the x-axis instead of time line and maybe first show the issue. What was the issue here? The thing is, the bar chart was showing time line and if between two evaluations a long time passed, then you have a get in there and you yeah.

B

If you have many evaluations in a short time ranged and they kind of overlap, and the idea is that now that's changed and actually I will show the screenshots. After for requests.

B

For some reason, it's not oh yeah, good and now I take the time as a label, but we have similar. So we have same distances between all the evaluations. These evaluations actually have a score of zero because I don't have any configuration for this lighthouse service, but we can see that there is the same distance between those and if there are more, of course they will ya, be.

B

Distributed like that, and there is a new label or, let's say, I- consider the label with the key built ID as a label for this evaluation chart. So if the label with the key build ID, is provided on this evaluation done event, then I use that label instead of the time for displaying on the x-axis yeah. That's the first pull request.

B

I, don't have running bridge with with that already to show it, but I will demonstrate the next one, because that's a bigger one, I implemented a relation board to show the heatmap with a deep link. So the idea of this was to provide a deep link to the bridge, slash evaluation, slash captain context and to show only the evaluation of this specific captain context, or you can also provide the stage or the event ID of the evaluation even to have only that one.

B

Why is that nice? So I will just show you here actually on my local machine when you have the evaluation map here we provide now also the link or, if you know that captain context ID for example, of in your Jenkins pipeline, then you can ya, provide a link to this view and in this view, actually it provided me already the specification for the staging stage. If I take just the captain context, it will retrieve all the evaluations with this context.

B

So in this case, I have one on stage death and one on stage staging and showing the evaluation with the heatmap and, as you have seen, when I provide the specific stage that I want to see, then it will filter those evaluations to adjust that one stage yeah on top. There are some more informations like the service, the artifact that was deployed that captain ID. So it is context again and of course, in this case, the stage staging is only shown by you can easily check the evaluation and compare it with previous evaluations.

B

In this case, there are four evaluations, but my heatmap personal has yeah I have bad date on this demo. Environment so yeah, that's it basically from my side any questions.

B

Then, thank you very much and I'm handing over to Sandra.

C

So my issue was a very small one, a good first issue. It was about this dialogue, this event played of dialogue, and the request was to edit button, which would allow a user to copy the payload to chasten string to the clipboard and just paste it instead of manually selecting it.

C

That's how it looks like now: that's just a simple button on there and if you click it, you get a little message or feedback that says that it is copied and then you only have to paste it yeah. That's it is that okay, okay.

A

Well, thanks that looks, awesome, I think if Rob is here or as if he's watching this afterwards, he will love that feature because he requested it. He often uses the payload for developing his own services and that copy button just saves him a lot of time and a lot of trouble. Thank you again.

A

I would have one question: does it also work for the slo-mo file, or is it right now only for the for the event payloads that are in JSON right.

C

Now it's just on that payload I look here, but we could edit anywhere else. It's the metric is generic.

A

Yeah I mean we can create another issue- human, if that's a feature request, obviously, but it would make sense I'm just wondering if it would make sense that you can also copy the SLO file and you don't have to take a look in the kid repository because we already have the data here in the bridge. So I will go ahead and create an issue, but it looks really cool and I. Think it's very doubtful. Thank you for the game.

A

Alright, thanks. The next one is on here. I, see sorry I see that rock just joined so Rock take a look at the screen that you see right now in front of you, it's pretty cool. We already have the feature that you requested. We already have it implemented thanks to Sandra.

D

Okay, thanks yep just cut that thank you for doing that. That's great.

A

Okay, yeah: we can move on to advanced I.

E

Also from my side, I'm a spec change described in this issue. More precisely, we would like to introduce mechanism to precisely state the spec version which, for example, and remediation file, form follows and therefore I've introduced, a new field which we named API inversion, and here you can precisely see that this remediation ml now follows the spec version in 0.14.

E

Of course, we also provide the corresponding data structures in our coil terms, and what you hear is that I've introduced a new package which we named like with the speculation and inside this package. We can now connect data which is specific to this back version. So we can collect the types which are specific to this back version, and here you can see, of course, all those types which we need for the remediation.

E

So this was my first issue which I would like to present and the second bomb was regarding the shift in soda, so the install of failed because most a new ARPA cruise, we do not grant and admin roles to the insula anymore and therefore we cannot use the default nginx ingress anymore, because this would require privilege mode and what I did now is to use a new image and map it on what 8080 and now we do not need any privileged mode anymore. So this is basically the second issue.

E

Does anyone have any questions, because otherwise we can move on to Christian.

A

Well done, move on um so I was working more on other things. This week have to admit basically kept. 18 has been merged and I formulated issue that people are already working on and some of them will be presented today and I also just implemented a quick improvement of the descript that you can see in our QuickStart. So the curl of gab, captain Sh, that's being piped into bash I've, updated it a little bit. So you can now provide the captain version as a parameter. So, for instance, the script still works as usual.

A

If you call it like it is in the docks, but if you want to specify a different version so for whatever reason you want the CLI in version zero, six one you can use it like that and I also hope that this will help people with creating some scripts and automations. Because now you can just say, hey I want to use an older captain version or I want to define the captain version. I use in case there is an update later and my script breaks so yeah that should help just that quick heads up.

A

This is not life, yet I still have to deploy to our s3 bucket, but it should be live by tomorrow, I guess so. That should already work with all the versions. Then any questions regarding that I.

D

Guess Christian! So if if there are other command-line arguments to the Installer, then you wouldn't use that approach right, you'd, more, get it and then construct your full captain install that.

A

That's not the install it's really just to seal I. Actually, so, oh that's just vcli okay yeah, but the idea is obviously if you want to always have two sea lion versions: zero, six one! You can either go to github.com and probably well everything we do in that script. Obviously you can do manually, there's nothing stopping you from that.

A

It's really just a convenience feature to have the feature parity with, let's say: Kay, free s or sto, where you can provide the version that you want to download as a parameter and in all honesty that the pull request was just this change in. So if captain version is not said just used to the default, one that we provide and if it said, use this one, so it's really just a parameter that was already there but not accessible from the outside yeah.

D

Yeah I think all right, I think I think it'll be very useful. Thank you. No.

E

Question sorry, so it's just a step documented in our releasing steps and that we have to change this. This version number yeah.

A

That's documented, actually, if you have to change it in the version Chasen and in the in the script, so there there might be potential for optimizing this right now, but it's to two things we need to update and it's documented in our release checklist. If you're wondering our release, checklist can be found on github.com captain captain in the wiki right.

F

A

Alright then I'm heading over to Florian, you.

F

All right, thank you today, Anna for addressed and I want to show you, so three of them are actually integration tests for various use cases. We are going to support with captain, so the first one is integration test form itself evening with Prometheus. So let's have a look at the issue.

F

Basically, what this integration.

D

F

Does is it and follows the instructions of the tutorial that's already available, so basically it will on board the sub application configure for Prometheus monitoring, then deploy note generator that will generate the lot of traffic on carts micro-service in the production namespace. And finally, you to that problem that arises due to this heavy traffic. The remediation service should get in trigger a remediation action, and that should result in the replica account for carts.

F

Microcell is to be increased by one and if you are interested, we look at the code in this pull request. So basically, this is just a large shares. Shell script. We're basically just followed the instructions from the tutorial afterwards try to be nice and clean up the cluster, especially the leads, there's no generator and yeah see here at the beginning.

F

We create the project or actually we make sure that we don't have a pre-existing project that reiterated upload all there were twelve fires, so Jamie the health checks to make sure that the initial deployment makes its way through to production. Then we add the remediation file and deploy the service, wait until it's finished and deploy it in reduction and we configured Prometheus monitoring generate some load and afterwards we basically just use.

D

F

Captain API to retrieve all the relevant events for that use case. In that case, for example, the remediation trigger event or status changed event all the way through to the remediation finish the event and check if the relevant fields of those events are correctly calling me so hopefully that will help us during.

D

The next release, particularly.

F

Make test a little bit easier. All right, I would.

E

Have a question to this request: yeah can I directly ask it so you're checking the number of pods right and do you also check whether a relation result is again passing.

F

The nation was not part of this corpus in illicit ly, excluded it yeah for the sake of keeping that integration test focused on testing the self-healing, as a no validation can be excluded. Okay, I.

E

See this is just to see if the.

F

Mechanics of the remediation service are all working and the action goes through to the action executors and it does what it's intended to do thanks alright, next one, it's quite similar, but the next test is for the deliberate.

D

F

Or manual approval use case at was so basically here we.

D

F

Quite a large shipyard file containing a couple of stages that contained various combinations of approval strategies.

F

First stage was fully automatic, other one was manuals, approval for past evaluations and automatic forewarning, and so on so possible combinations of automatic and manual and again here, I first created project using the captain CLI with this shipyard and then I just sent.

F

Ya, exactly using the CLI, we sent some evaluation done. Events also containing various combinations, so failed warning or past evaluations. Then we checked if the approval the trigger demands were sent out correctly, also using the CLI, because he would have the possibility to do some to the pre folder of Madras world. So this checked, if all the expected events were retrieved and then I pretty much again used the CLI to send the correlating approval that finished on events and afterwards again checked using the get event approval that triggered command.

F

If the open approvals have been closed as a consequence of sending the approval that finished events we're a little bit low on time, so I'm not going too much into detail in the Shiva script, but of course you are invited to to have a closer look at it if you are interested, but it's really similar to the previous requests using a shell script, I created the project, send the events, use the C&I to retrieve their approvals and sent approval instruments all right. Are there any questions about that integration.

A

F

All right, perfect and the last one I want to show you is actually really similar to the self healing with prometheus use case. This is this is very much focused on testing and new possibility of chaining multiple actions one after another. So in essence, here again.

C

F

A project then sent a problem that open.

D

F

And here in this remediation file for this problem type that was specified in the problem of event. We have multiple actions, and here we basically just tested various scenarios. So, for example, the first one was if the service was not available at all so much onboard yet or created yet. So, in that case, we of course expect remediation to be finished immediately and with.

F

Appropriate error message then the next one was the service was on boarded, but the remediation file was not available yet so that, of course, should also be reflected accordingly and then from there on we tested if the actions have been executed and also tested the case. If, for example, one of the action executors was not available or.

F

Already or if it was available and the action couldn't be executed for some other reason, for example, person, maybe I am time that that occurred and yeah and finally, we tested if all the actions were executed, one after another and the correct order, and if we eventually received a remediation initiative end and of course, if you want to have a look at the shell script, you can look let the corresponding poor request for that issue.

F

All right, then, just one very, very small addition, but that is actually very helpful, I hope so I extended not even capabilities of the CLI quite a bit. So now, if you especially for the install command. So now, if you not use the verbose flag, it will give you a lot more insights into what is going on. So the biggest change here was that we know print out all the cute cutter commands that are executed and their exact outputs.

F

This one even got a little party, a emoji from Kristen's. That, of course, was a nice feedback and I hope it will help folks to propositions in case there are some issues: yeah.

A

It certainly will help, but I've looked through a couple of locks so far and I think in the future we might have to introduce some level of velocity as well as a parameter, because I think some output is really there's like a lot of output. That's obviously as defined in a ticket but I think there might be users that might be overwhelmed with the output, but yeah.

F

A

So forth, for now we have as much information as possible that time we can trim it down, but.

E

This is I already. It supports the poles mode, so changing this would be yeah a really really small issue.

A

All right, his Christian acumen here, yes he's in yep yeah.

G

I heard you call last time on the YouTube video that I should join this time. Thank you. Thank you. No problem, no, the weather is bad, so it's it's not so bad to to stay in the office oops. This was my video. Let me presenting Sheamus King.

G

So yeah this is already merged right, so the captain in Samantha's was yeah a good first issue, so it wasn't so so hard to exclude the Oakland shift namespaces from the yeah recommendations. So what I did here is just- and it don't happen anymore, right now, setup so I'm, just yeah checking. If the prefix from the namespace is cube or open shift, then it should skip the recruited recommended action to delete this and as well. What I have removed is because this captain, sir, is 7.

G

We have this killer, not anymore right, so this is obsolete at the future and what I'm working right now is the meter data retrieval so yeah I'm, yeah I thought it would be a little bit more easy for me, but it isn't so and ondrea's. Thank you for all your help.

G

So what I did here so far is to. Let me quickly go and code is to add the add, an extra endpoint for metadata, generating force, a response model yeah the response model generating the API endpoint and now I, have to add a new epoch role. To actually be a list.

G

Deployments was in the Kuban eros cluster, so I have to adjust our body and now I'm able to yet to get the bridge deployment reading the bridge, what it's called image tag, and also getting from the swagger llamo, the captain version and last but not least, what I've done protrusion vigor version and for the namespace to get the namespace we're at the API service is running or captain is running.

G

I added to the service yeonil, an extra environment variable API, pod namespace, which is getting this meter data from from Cuban Eddie's to print this out and the eyes can see I already executed it once so. We're getting the bridge version, captain version and the running namespace of captain at the moment and what needs to be done is to add these stuff, as requested as well to the twosies, create support package functionality, but I. Don't necessarily.

G

Perhaps we need an extra discussion here, because most of the information is already in the support package, like the deployment information from the bridge and all the deployed services within the captain, namespace so yeah.

A

We can discuss this on the issue, I guess, if it's already there, the idea probably was to replace it with that API called. On the other hand, what if the API is not available, then the support archive that's being generated, uses, cube, cuddle, I, think so. Yeah yeah, let's discuss it on the issue. Okay,.

G

Right, so that's basically from my side and again I want to thank you all for the t-shirts right. So I want to get me hold. I want to get that whole captain ego from now on.

D

Clap clap, clap.

A

You have pretty cool and it should have on the on the side of your shoulders. It's because it's very special teacher, the first time we tried it glad you liked.

F

A

I would like just to mention Andy karna is not here today, but I was just going to mention that the Qi meter extended service, which was in captain Kahn trip, has been merged into captain core. We don't really need to show the pull request, because it's essentially a copy-paste of the code from captain Kahn trip to our captain repository, but one quick thing I would like to share is the examples repository.

A

We had this request to add the humanity's recommended label app humanities, Iowa Version, which we've left out so far and Andy has done a great job and added its there's. Actually, a quite interesting command. I haven't seen that in helm, but you can kind of extract the tag name of the image with that command and then just added into your community's manifest. You know too, into the helm chart actually. So that seems to be working nicely and he's also, as you know, working with dynaTrace a lot.

A

So he is mister dynaTrace himself sometimes, and he's tried this well in that race and it seems to be recognized by dynaTrace as well. In this case, it's a service that he deployed in three different versions, so it's showing up all three versions but yeah. Essentially, the point was that monitoring tools can use this information and yeah. So if you're interested in a pull request, it's linked in the document as well as here, and you can take a look at it.

A

So if you want to add that to your manifests yeah, it should be as simple as shown in the pull request in the last commit I. Think he and this one I'll actually add that commit link in the doc.

A

Okay, so in the document itself the next thing would be kept or spec issues. If you have anything to discuss, there's nothing been added, so I guess we can skip it today.

A

Also, looking at and yes yeah he's nothing, then we can go over to the work-in-progress issues.

A

We have quite a lot of issues and are ready for review column. So if you have presented one of the issues today and they already have been merged, please move into the done column. Once you have time, then the things that are in progress, this one is from johannes. He is going to work on it once he's back, then this one is the one that Christian has just told us about. It is still working on and I think that one florian you also just have started working today.

F

A

Nice then, we still have two stock issue: that's also they weren't going to be done by Johannes ones. He's back about MongoDB issues when starting up captain on was it mini, cube, I think so it's a dock issue, and then we have one issue that we have some somebody has told us basically that they would like to take this. One I guess, there's a question that I need to answer. Then twenty two minutes ago, in the middle of the community meeting he's not in a community meeting.

A

Is he guess not, but he will probably do that. So that's also nice, so we've assigned him and with that we can actually go into our backlog.

A

Work is any. Does anyone have anything to add to the issues assigned or in progress?

A

Is anyone still waiting for reviews? I guess I mean you're waiting for for review, so I'm.

B

Waiting for two reviews on my pull requests and I will move them to done as soon as they are merged. Yeah.

A

We'll need to get them merged and quick all.

E

Right you like to do this.

C

B

The two that I presented today I, am.

A

Not sure haven't they already been reviewed. I have two already been reviewed. Maybe, and one of us needs to give the final approval. No, no, not yet. Okay,.

C

A

You son, oh yeah, cool, all right, then, let's go through the issues here. The first one is a real, a minor, but it's really nice. At the moment. In the directory service, we have fixed operator version zero, six zero because of Cuban leaders, restrictions of of captain, but we've captain zero. Seven that were working on those restrictions are no longer necessary. So we only need to remove that restriction again, basically and should be like I think one or two lines of change.

F

A

Wants to do that.

A

A

Then the next one I had to put back in a backlog. It's the version. Checker interrupts you so when there is an error, basically think yeah seshu has described obvious, or he has described this when he creates a project. There is some problem. If the internet connection or as a viable rule, he gets this error message shown for the version check of the captain COI, and I think the idea was I. Think under yes, you, you commented on this quite a lot. The idea was to make this error message a little bit.

A

Look less like an error and more like a warning about to not show it at all.

A

F

A

It's also probably a rather easy issue. It's it's marked as a good first issue. So if anyways.

G

I can I can take this one I'm.

A

G

Will comment on the.

A

Issue yeah I can still this is this is kind of a github issue. We cannot assign an issue to someone unless he is really an administrator of the project or he's commented on that issue but yeah. Thank you.

A

Then this one is from Andrea's, maybe Andrea's. You can talk a little bit about it sure so.

E

Let's kept eating, we will not install SQL anymore by default and before we also installed so-called public gateway in our installer, and now we cannot assume that this public it way is available and after implementing cap 18, therefore we have to make, and this gateway, configure, able and I would propose to use an environment variable for this and, as I permitted it, I can take this issue.

A

That's awesome one question: we should offer a way to install this public gateway because I tried this and I was able to just apply the manifest for the public gateway. It's quite easy, so we should probably have that option available in our documentation.

A

E

Is actually one task described in the talks? It.

A

E

We can reuse, public or gateway which it is coming with. The installation of is.

A

There actually one coming, because what I've looked at there is. You have to create one by yourself: okay, garbage.

E

A

We'll check this one by Thomas, I'd like to move it to the bottom of the backlog, because we haven't discussed this internally yet. Finally, so then we have one for captain's: bridge alloys has donated us some nice screenshots that we can use to make the bridge look a little bit different and yeah. So basically, it's a change of the background color on the left side and some changes in the font sizes I require.

A

Who would like to take this one I.

B

Will do later so, okay,.

A

So I mean thank you.

A

Then we've kept 18. We also have a need to have a little bit better troubleshooting with support archives regarding ingress options. I'll actually open this issue in a new tab, because it's yeah there's quite a lot of info in there. So maybe this is also something that the the issue that Andrea says just described with the public gateway. Actually it's already there yeah has to do so in we should edit. We should adapt the generate support archive command such that we check.

A

If it's your system is available, if a public gateways available, so I guess this is now.

A

A different one.

A

Do you know the issue number Andreas of your issue? Out of your help? Okay, I'll. Take a look.

A

There was 86.

A

So basically, we need some information in the support archive that helps us to troubleshoot on whether the installation should work or not, or what is wrong with the installation and public Gateway SEO system would be one of the things to check also to list all the services with the type and the internal external IP addresses in the East, your system, namespace, or maybe we can extend that list to some other common known namespaces, where we would expect you in I know it's of arrival you can configure, but there are some best practices, I guess also if there should be an option to detect whether an ingress controller is installed.

A

So, for instance, cubanía should know which ingress classes are allowed. I'm, not sure if this is easily possible. So we should just try it out and if it doesn't work, we can comment it on the issue and we should find out what IP addresses the cluster nodes that were working on have so either internal and/or. An external IP address, if available of the cluster, should also help for people that use node port instead of load. Balancer.

A

Right any questions regarding that issue or any volunteers for that issue. Oh.

F

A

Right. Thank you.

A

There was this one good, then here comes the big one of one of the big ones of kept 18. It's a very long issue, I'm not going to go to read all of it, but essentially it is the changes in the CLI for captain install and it has to do with removing anything related to East, U or nginx ingress, and introducing some new command-line options.

A

Don't need to decide who does it now, I guess if you want to read through the issue carefully and ask questions afterwards: that's fine, but yeah actually.

E

I would like to do one of this big issues. Either this one or the other one you're removing it from, but I don't care which yeah.

A

I mean we have here the installer issues, so that's the the same issue. That's linked to the CLI, but I've we've split it into two issues so choose whichever you want, which.

F

I prefer other one.

F

Feel free to choose which.

E

One would you like.

F

Also one of those but I'll leave the choice up to you, which you want then.

E

We don't have to decide it all.

A

Right you just take each of them like take one.

E

Just the sense and that's okay for me: I'll.

A

Put it into that column or, though, because it's kind of its kind of assigned and committed, but it's not really assigned but yeah, all right, the last one or almost the last one is also one from actually last week. I think Johannes has edited as a comment. I formulated an issue now, so that's also a reason why I've shown the pull request Matakana with the recommended label version, because we should also introduce this Cuban leaders, recommended label version and other Cuban leaders recommended labels to our captain core and control, plane and etc.

A

Installation, because that's really information, that's helping monitoring tools and administrators to find out what is going on in the cluster and what is there I'm, not sure if we need all recommended labels, but obviously the version we can use the name. We can use instance I'm, not sure if we need to because that's more something for stateful sets component I think we can use two because we can set in a database or API or or web UI, or something part of we can probably also set to just captain.

A

So this issue is really just like adapting a lot of manifest files in the captain. Installer.

A

Any volunteers.

A

If not we'll leave it in a backlog and we can just take it if any one of us finds the time and I think the last one like I said we haven't discussed it, but I think Tomas is in the meeting and he can tell us like the idea behind this issue, and maybe we can have a quick discussion. Whether we yeah should do this right now. I mean we can't decide it probably now, but let's just have a discussion anyway.

A

H

Yes, currently, we are using multiple, multiple host names for exposing captain, the captain API and the bridge to the outside world, and we are often using CPOE by os DNS provider to generate host names for them. But, to be honest, these two services are not really really stable any time and we talked a bit about it and I think it would be a good idea to make the API ant cave and the bridge accessible via some host name or the IP address itself and slash API were slash bridge.

H

This would reduce the complexity of this of this topic in my opinion, and we yes and it would be easier to configure a nice domain name for for captain currently I think we should make this. We should also let the let the old path there until I think captain syrota, 800, 0.9 and afterwards deprecated, yes and currently, I. Don't think that um we have two possibilities to achieve this. We could rewrite everything which which ends with Slash API and send each other to the current endpoints.

H

While we duplicate the ten points and do it this way. So yeah me go.

A

E

See lots of benefits, for example, we only need one node balancer in the future, and this can save lots of money for customers. The second one is: we exactly need one one SSL certificate and it will probably be easier to use use such an certificate.

H

A

Yeah I mean III, see the benefits of it. It was also in back in my head when I did write cap 18 to do that. It was more of let's maybe postpone this after a 0.7 release, because it has some implications. Like Yuri said you would have to leave the old URL available as well as the new endpoint available, and that creates some complications in what we do, because right now, this whole thing is managed. I can show you why our and nginx I think I need to open it. Any installer manifests.

A

Think no yeah I was in captain there we go to API gateway nginx. It's right now managed in an engine X image itself with a configuration. That's the one that andreas also has updated, that they have shown us to the unprivileged nginx version. And right now we have like, like the location, slash is being passed to API service and I agree. We can just write, slash API here and pass slash API to the API service and then pass slash or slash bridge to the bridge. But there's a lot more to this.

A

So we'd have to rewrite all the other end points and and again we need to provide the old end points as well. So we have to duplicate a lot of code for this. If you want to provide the old end points and I'm, not sure if we want to do it, I think we should break the functionality. So there's no deprecation, we just break it.

A

We have a zero in our version at the beginning, we can break things but then again, I'm not sure, if that's really what we want right now, so I'm open for discussion, basically, basically.

E

If we would like to do it sometime, it's better to break it now and earlier than later, and for the mum we would have a one here here in this meeting who can start with this issue? Of course, everybody of us can help you.

A

Yeah I mean, like I, said I'm open for discussion. We can do it and, like I, said better break it now than in two months. So yeah.

E

Crystal for you, what do you.

G

So, for me, it's it's beneficial when, when we don't have these extra subdomains, because then you can use, for instance, we have my court SSL Certificates right, so we can add for the load balancer or for the ingress know. Whatever data will be, we can reuse our certificates in half yeah. Well, it's certificate in front of our bridge and in front of the API.

G

So yeah I like this idea and to be honest, I'm with you better break it now than later, because after I think 0 7 there will be a lot of wins in through 8. It's it's there. Now then later right.

A

Ok, so is anyone against it? Basically, let's put it like this, because if not we'll can we can leave it in there and if anyone wanting to help with that issue, yes I would do it right. Then let me assign it to you and thanks for bringing it up and thank you, everyone for the discussion.

E

G

Three hours of traffic configuration last night.

A

All right, I think we're done with our backlog and our planning. Thank you, everybody for your participation and well at one one, quick announcement from my side.

A

I will be on paternity leave for two months, so you will not see me for two months, but next week you and Johannes will go through most of the meeting again and obviously all the usual suspects that are in this meeting I hope to see you again well in the YouTube recording, maybe and obviously when I'm back in two months, I hope to see all of you again and even more people and yeah. So thanks for joining this meeting and see you soon, all right see.