keptn Community, 20 May 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keptn Community Meeting - May 20th, 2020

Description

May 20, 2020
Noteworthy Issues & PRs (Review)
Florian:
Manual approval via CLI: https://github.com/keptn/keptn/pull/1800
Exposed mongodb-datastore via API: https://github.com/keptn/keptn/pull/1798
Ermin:
Show stage details in environments screen: https://github.com/keptn/keptn/pull/1832
Christian K.:
Issue https://github.com/keptn/keptn/issues/1777 (K8s 1.14 - 1.18 support)
Version Check in keptn install https://github.com/keptn/keptn/pull/1814
Increased test matrix for microk8s https://github.com/keptn/keptn/pull/1827 - see https://travis-ci.org/github/keptn/keptn/builds/689208935
Christian H.:
Johannes:
Roadmap → KEP / SPEC / Issues:
KEP 09 - Spec change of remediation action config
SPEC 31 - Implementation of spec change as proposed in KEP 09
#1616 - Draw event stream of a remediation scenario
#1816 - Keptn Contrib (unleash-service): Extract featuretoggle action from remediation-service into unleash-service
#1817 - Keptn Core (remediation-service, helm-service): Refactor remediation-service and move functionality of scaler to helm-service
#1818 - Keptn Contrib (dynatrace-service): Move posting Dynatrace problem comments to dynatrace-service
Additional issues: to-be-defined

SPEC 30 - Add ProblemURL as optional field to problem.open event
#1694 Add ProblemURL to the keptn problem events

Architecture for Remediation
Draw event stream of a remediation scenario · Issue #1616 · keptn/keptn
Working 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
Attention: new dates for upcoming community meetings
Wednesday, May 27
Wednesday, June 10
Please subscribe to the community calendar to stay up to date (ical)
Miscellaneous:
New blog out on: how to overcome scaling issues when using Prometheus in an organization

Learn more: https://keptn.sh
Get started with tutorials: https://tutorials.keptn.sh

A

And we alive cool welcome to this deaf meeting kept deaf meeting again. We have a couple of cool news to share and also a couple of PRS to review and I. Think I don't want to waste that much time. I want to just hand over to Florian because Lauren's the first one in the agenda and it's giving us an overview of his yards that he did this week.

B

Both are related once again through the manual story and, let's just jump right into it. So the first PR is for approving.

B

C

A DC and I and.

B

This works as follows: request or request I've added the captain get event approval that trigger command, which does exactly what the name tells you. So, if you remember, we know what those approval triggered. Events which is to name, suggests, trigger manual approval process, and now, if you like in the following into the steel I for example, yet having that event, approval that triggered for the project, structure and stage staging, you will get happy little approval that triggered event.

B

Where have all the required information like in this case tag and the image that's to be approved, that's relevant here, and what you will also need in order to further progress in this manual approach flow. Is the idea of exactly that event. That's also being returned here and now, if you want to approve this request, you can also use this in your eye and.

B

Say captain sent event approval finished.

B

Finished and again specified at a project in the stage and the ID of the approval of triggered event. Then it will send approval finished event, and after that, this will be received by the gatekeeper service, and this will trigger the further deployment of your artifact. What the gatekeeper will also do is it will check if the that finished event contains the required tech and image for the correlating approval that trigger event, all right.

A

Not about the feet, I have a question I'm, not about the feature itself, but about displaying the approval triggered event.

A

Here you are showing the the payload that is coming back. This is a kind of open question to the group. Would it make sense to display this information in our table? I mean we're just have see the ID in the first column and then and the second column the image and attack.

D

Then, as a follow-up issue, we can even think of.

E

An interactive.

D

D

Proof on this, but this was clearly not the scope of this issue. That's.

A

True, but that's true yep, it was not scope of the issue and I would also go with and follow-up issue on on that. One, yes, yeah address, as you mentioned, to provide an interactive way to approve it. Lauren.

D

Can you scroll to the right on the approval finished.

D

Yeah, so what is exactly checked, then? You mentioned you're checking the image. Can you explain this in what countries.

B

Yeah sure, actually, this is hidden from the user in that place finish event by every single eye, because to CL I will use exactly the properties of crew without triggered event. So we'll take the image, that's being specified here and back as well and includes that into the approval that finished event. However, to reduce the API in order to directly sent the event to captain, you have to set the payload of the event by yourself, obviously, because you can specify anything there and the gatekeeper service will then check with those properties are set correctly.

B

So, for example, if you have a approval that triggered event with attack 0:10 one and then use the API to send an approval trigger event with the tag said to 0x that, while the gatekeeper will reject that, because the text imagine that face.

D

And don't one not a question you, my is the the service name not needed in the proof finished so eternally? Is it a collection containing all services or how.

B

Do you, the service is not needed in that case, because you refer to the idea of exactly this approval that trigger event.

B

D

And using the IDI I can find right collection.

B

B

Thank you and the other issue I want to show. You is actually also part of this story. It was a requirement for implementing the get approval triggered feature the finished feature of the CLI and that's the most of the MongoDB data store by.

A

B

B

We just open this pull request, so what I've done here is I've made the MongoDB datastore service that previously was only reachable from within the captive cluster I've, now made it available by the captain API.

B

So this is well I can use this API in order to retrieve all the relevant events of for death news case, because previously we had kind of limits down endpoints directly in the API, which allows you to fetch single events or a certain captive context, and that wasn't flexible and not for you in that case, because we go again to this I extension PR. We see that we here need to specify the project and the stage, and in that case we do not have to cap the context which was the.

E

Only parameter the.

B

Old API, most exacting.

B

All right and that's everything I wanted to show you today. Thank you.

A

Thanks Lauren, um as I mentioned I, will take it on my gender to to write the follow-up issue for the table of the CLI come on or for for it get approval and ante output in a table. Format.

A

Let's discuss this on on the issue that I will write down, yeah I, agree, I, think it's it's not a good user experience when someone has to copy/paste the ID and then use it to send their approval, finished event and I think we. We should provide her a more convenient way to do this. To this. We to do this with with one command and was an interactive approach.

F

C

F

You I think next turn is mine, since this was a short week and many changes had, but under the hood, I have only one issue that I only want for requested. I wanted to show you it's about environment screen that we introduced last week about details so showing stage details in the environment screen. The request was to enhance this view and I will show you a live demo when you click on the environment screen.

F

You now have an overview of your stages as you had before, but you can also click on a stage and then you can see the services that are deployed in this stage with information, what image which image is deployed and when it was deployed in this stage- and you also have if it's available the deployment, your I, the public deployment, your I will label here. So with a single click, you can access that service, that's nice, because now you can directly yeah jump into your staging or production stage and check your deployments there.

F

So you have to link to your yeah current deployment.

F

Yeah, that's actually from my side, since the other changes in the bridge have been regards that public, KPIs or the bridges not consuming the public API and that's something that yeah haven't changed anything in the bridge visually, but just under a hard I.

A

Mean can you go back to the bridge.

E

A

Yes, because I think we should mention that in the market or in the vial frame, we did not include the button to to the deployed service, but I really like this idea, because it directly forwards you to to the service itself. And additionally, you also added that deployed at timestamp question tutor to the group. What do you think about the deployed at time stamp? Does it make it? Does it make sense or- and is it a good idea to have it there I.

D

Think it's one of the most important information in a deployment tool to know. When is this version deployed so I, really like it I.

G

Think it works.

D

But maybe we can make it a little P so also add image column so that we have a common format.

G

Yeah, you don't always soft, it has the you, don't always have the image. Obviously, don't forget that it's one of the you haven't been there last week, Rob mentioned that, so we need to make sure this. The artifact that were deploying is as generic as possible. So for us it should really be a straight a string and not an image and.

D

Let's call it artifact yeah.

A

This is the most generic word for something to get deployed. Yeah.

F

We could also add a label there so that it yeah fits better, and yet those are information that we haven't considered in the mock-up, but once you start creating the view, you see all the data that you have available there and that's something I would I thought would be a good use case that you can directly jump into your stage and yeah click to your deployment to the public deployment you all right, yeah. Thank you very much for the feedback.

F

If that's all I'm handing over to Christian thank.

G

You so I've been working on issue 1 7, 7 7, which was to basically check the community support matrix, starting from 114 to 118 for our current master branch of captain, and so far a couple of things already have happened. Like we've added testing for cheeky 1.16 last week, we also have added testing for the captain control plane, so the install optional use case, quality gates on various versions of mini cube and I'm still working on that issue. But there is some smaller artifacts that I can show you again in terms of pull requests.

G

The first one is that we've increased the sorry that we've removed or adapted the version check in the Installer or in the install CLI, to now only check for platforms that are not equal to open shift. But if the platform is not equal to open shift, then we just check our version requirements here.

G

This is different from what it used to be before because before we did check this requirement only when we were installing they're all use cases options, so the default option of the Installer now we're checking is for all of them, because yeah captain seems to be working on all versions of communities of our support matrix right now.

G

The second pull request is also regarding the matrix I've increased the test matrix from Michael Cuban. It is just to show some code, essentially I've used the same approach that I used with mini cube. I think I've shown it last week. I had I have a template here now in our Travis CI Yama and by default we're testing micro, kubernetes version, one point 16, and by inheriting from that template or by using the template that can override the version and just use different micro quantities versions.

G

If you're wondering if the version syntax here, I'll scroll down that you might, you might see why this is happening. We're using that in the channel of our snap, install command and the channel of micro kubernetes basically have added a link here, but you can go there and you can see which channels are available and 1.17 stable is basically the last 1.17 version. That's marked as stable by michael kubernetes.

G

All of that results in a built matrix, I know there is an error in that, so I'll have to dig into that later, but basically, we now have a built matrix that tests mini shifts, which is only available in one version as far as I know, or at least in one that we support, because that one is mini, shift to open shift 3.11.

G

Then we have micro kubernetes in the free mentioned versions, and then we have our mini cubed tests in all of these versions and with all of that basically, we now have a very nice build scenario for our master branch that keeps testing whether we have whether we still are supporting the correct community's versions. But what is still missing and I've documented that an issue is testing piu, Anita's, 1.17 and communities 1.18 with the full installation so with Helmand is. Do I haven't managed to do that. Yet that's still on my list.

D

Christian I have a question to Travis, so is there no limit on parent chops? There.

G

Is a limit on parallel chops? Sometimes you don't hit that limit. For instance like here this whole built for micro communities took 20 minutes and these were executed in parallel, but for the mini cube standalone, these took 40 minutes, whereas each of them is around 20 minutes. So I guess the limit is about 3 to 4 chops in parallel.

G

So at some point we will have to ditch some of the tests. I guess because well, let's just face it, we don't need to test every single version every time, but we can select, maybe free, maybe the lowest one, the highest one and one in the middle to verify that we are still good, but with three tests in parallel. This seems to be working. Fine.

G

Also, just to be fair, I mean before anyone like is saying: hey kept misusing all the resources and Travis when we are calling tests here. This is not like being 100% CPU and hundred-percent memory usage. This is really lightweight CPU and memory usage. It just takes some time because there's a lot of network traffic in wolf there's a lot of checks involved. There are a lot of sleep statements involved, so yeah.

G

Alright, it's also just been a short week for me, so I don't have anything more than that and I'll I think as Christian acumen is not yellow head back to Johannes. Alright.

A

Ok, I have no issue to discuss to this week, but I will give you an update on the roadmap. I mean every second or third week. I will do that so that everybody is aligned and knows what the team is currently working on and just let me use this slide.

A

Yeah I showed you this slide of two or three weeks ago, and we, the decor team, is doing really good progress on working on the use cases that drive to kept noted, seven version and, as we saw today, we now can we can get their approval by using the CLI and also approving them by sending the approval event, and we also making good progress in regard to the captain bridge for supporting the terminal approval scenario, and the next use case that we are tackling for old at seven is the use case around a self-healing with custom remediation providers, and this use case is currently in this data of backlog refinement.

A

And what does this mean?

A

Just? Let me jump over going forward. I will also maintain and I will maintain the roadmap of captain and a roadmap Kanban board this one. You will find when you go to captain organization, then there is a project tab. When you click on that one, then you see the roadmap, it's a public project, and here you see a Kanban board for the captain releases and I yeah.

A

The best way to read this Kanban board is from the right from the right to the left, because here we can see that this issue- or this use case four out of seven- is done. It's the use case about exposing the captain's bridge and using the CLI and also providing basic authentication, correct.

D

Me if I'm wrong, this would be six punched.

A

Correct yeah, we delivered that one with all that sixth, okay.

A

Currently in progress is the captain Hoffman proposal, ten, which is all about the sea lion, hands mints for retrieving, caped, captain status information and the second one is the delivery assistant. Please don't get confused and this is a kind of renaming of the manual approval scenario, and here you see the benefit that the user gets out of that use case and also to expecting respect change.

A

That goes along with this user story, and let me now come back to the self feeling use case, as I mentioned, or as I said, this one is in the status of backlog, refinement, and here again you see a note with a title of the use case and then on the benefit and the advantage for an end user and further down you oops.

A

What did I do? What happened? No I moved it there that's back further down. You see two references, one is to the captain enhancement proposal for this use case and the second one is for the spec change that is required for implementing this use case. Yeah we as the captain team.

A

We just still use this way to provide an public place to to discuss, use cases and and enhancements, and please feel free to take a look at the cap enhancement proposal, which is a kind of explanation of the use case from an end user perspective, and there is also the respect change that then corresponds to the cloud events or to the specifications that are required to implement this use case, and when you want to take a look into the future, there is also there are also two use cases that will be prepared for the captain ordered eight release, but this is not content of today.

A

Now let me go back to the community meeting notes, because here I have written it down the cap. Oh nine is about the self healing use case and providing a new way of specifying remediation actions. This is the cap. Then we have the spec that specifies everything, and we invested this week a lot of time into breaking this specification, or this use case down into concrete tasks.

A

The first one is drawing the event stream of the remediation scenario, and then we have already identified tasks that can be done in order to fulfill this use case.

A

All right, yeah.

A

This this task drawing the event stream. We will also mention in a couple of minutes, but for now it's okay and next to this spec change. There is also a very nice example where the community provided us feedback and and proposed a spec change on the problem event. Actually it was Adam, Gardner and and and drop who we're working on a use case.

A

We already required the problem URL here in their service, the implementer and they proposed to add or to add the problem URL to a problem event, so that they then can have already get a backlink to the source of the problem itself, whether it was identified by an address or Croesus, they can then go back to to the source and based on this spec change, we also derived a concrete task that is already yeah that can be implemented because it's it's it's clear.

A

What need to be changed and it's also laid out what things need to be modified and I really would like to get this back change into into the master branch, because it is a very nice use case to get the problem ul. Also for for the action providers, because they can then show this URL and then also provide a backlink who can I ask for giving me a feedback on or a review on, this spec change.

G

Yeah, one of them would be me and I'm, not sure, I think the repository set up so Alice right power would be the second one to approve it. I think I will double-check that I.

A

Think I need at least one approval from Eurosonic Christian and a second approval from from someone else. One approval from Co.

G

Donor, so it's enough if I approve it: okay, okay,.

A

Yep as I have written down here in the in the description of this spec change, it proposes the new optional field of a problem ul, and it also addresses an issue that was identified by them, because we provided a wrong example of the dynaTrace problem. Please don't get confused when you see multiple lines that have changed, because also the example has been adapted.

A

G

I'll take a look at it. Thanks thank.

A

You very much cool yep when this one is merged into the master, and then we are ready to implement the spec change in this this task. Alright, then, let me go to the next meeting point in the agenda. It's about the architecture for for a remediation, workflow big, shout out to andreas who also worked on that and gave me good feedback on how we are going to implement the remediation workflow going forward, and this one is actually.

A

A task in the working items.

A

It's that one that is under review, ready for review and.

A

It is here, draw the event stream of a remediation scenario and when I open it, then um you will get. You will get here a payload or not repeated its specification of an event flow. You can open this event flow by using the sequence, diagram, dot, o-r-g, just copy and paste it and at the end- and you see the event flow that is shown by this PNG, and it explains how captain works when a problem is received by captain.

A

This can be a problem from from users or dynaTrace and taking care of the new way of specifying remediation actions. Then we can- or this PNG shows you how. Then the services work together and how do you yeah? How do you working or how do workflow is, is going on going through this event? Stream is I.

A

Think it's too much for for a community meeting and I'm I would ask you, when you're interested in this workflow, to take a look on that one and to reach out to to myself or to someone else from the core team to get feedback. When you have questions on regarding the event stream.

A

All right yeah, but this is the fundamental base and to continue working on the self-healing use case, and we are also taking this one as the truth and this the basis for going further. Regarding to your self-healing tasks,.

A

A

Okay, then um any questions so far.

A

If not, then I can go back to the working items and yeah.

G

Can you promote Tomas, he just joined as an attendee.

G

A

Epitome Isis here.

A

Okay, then yeah: let's go back to the working items and.

A

Like we did last week, we are going through the Kanban board of the working items from the right side to the left and discussing what's still in progress, but other blockers were not and then finally talking about issues that we can assign to someone I.

G

Guess the drive end stream. We can close that the 1616 yep direct yeah cuz, all the others. Okay, there's a couple of issues on the bottom that are still open. Okay, let's see this.

A

A

Exposing the event endpoint this we saw today.

A

And this is a PR you wanna, discuss and show us the next time. I.

D

Mean we can all see the result when we are using the CIA. So that's fine for me.

A

Rekt on an approval, finished event.

G

Yeah I think you don't need to discuss it. Okay, but.

A

G

Don't need to move the items now, just one wondering because of the two open issues or non closed. Three is three non closed issues. If they are done or not, I have.

F

Two in the bridge that are not yet merged, so I will okay.

G

F

As soon as the pull request is merged, okay, then, let's leave.

A

It open as long these two guys to use the public API yeah.

D

A

Then click on a stage: okay, yeah.

F

One of those has a broken build, so I have to check on that. First.

F

G

The captain core, the 16:58 16:50.

A

G

C

Close okay, good.

A

All right and the up creator for christen this was on an issue that documentation issue that you will get.

G

This week, yeah.

A

I think we can move it to there. Yeah.

G

Call him it's not worth mentioning it.

A

Then we have to to open and the up creator floor, and you want to give us a review about the upgrader next next week.

B

A

Then I'll leave it in the review column and manage open approval. Events in a collection.

A

Nice thanks, then, let's go over to the issues that are in progress if first one CLI implement get project service stage and metadata. This is this: one is from Christian hacker man. He has an open pull request.

A

And the thing are there already discussions going on regarding this his pull request? Oh, he did not link it. Yeah.

G

Florian has commented on it.

B

I had some remarks and currently in exchange with he's going to provide him with some.

B

G

Yeah Cristina said he can't join today. So let's hope he can be there next week and he can present it next week.

A

I'm, just linking it's.

G

Not always working the linking oh.

A

G

Let's, let's deal with that later yeah, the next one is an open issue. Maybe, as Thomas is not here, I can ask him Thomas. That was one that I did by your request. Removing the API road captain header in go utils and I. Think you did it manually in the CLI. Is that still necessary? Should we go forward with that.

E

I think we'll need it. Yes,.

G

Okay, so there is a couple of things we still need to do and I will need to run the tests afterwards, but I think that I will get done on Monday perfect. Thank you.

G

The community support I've talked about it. Also working progress should be done until the next community meeting cool.

A

F

Last one is yeah in the bridge. There are still some unit tests that I'm working on which I want to finish yeah it's ongoing, but yeah since I had not the time to look into that this week it will be next week. I guess.

A

Cool then we have two that are assigned it's the first one with the filter mechanism in the list of fruit events, and the second one is bridge shows wrong duration date from revelation. Ten events I think we discussed these two guys last time and I think we can skip it okay. This brings me now to the last column, which is the backlog in the black backlog. We have three box. When we follow the park first approach, we should address them as soon as possible.

A

The first one is captain: exposing the bridge failed when there is an ingress namespace that is not maintained by captain. Actually I brought this one in by a request from a from a captain user and the situation is that one, this captain user is using a goonie Tascosa there. He installed on nginx, but not used for captain for another purpose, and then he installed captain in a full-fledged manner.

A

And finally, when he wanted to expose the captain's bridge, then there was the problem that that the implementation of the captain API takes a look, whether the nginx namespace is there or not, and in case it is. Then we are assuming that there is a captain ingress.

A

But in this particular case the namespace is there, but the captain ingress not, and this resulted in a failed command and the user was not able to expose the captain's bridge.

A

The fix is pretty pretty straightforward referenced if the function here that takes a look into which type of ingress is installed and used, and that the fix itself is just checking whether the namespace contains the resource that we are looking for.

A

And it's someone has someone time to do that.

B

A

B

G

Please, or actually for everyone to remember the next community meeting is on Wednesday again, not first day so being that tomorrow is a public holiday and Friday is probably vacation time. For many of us make sure that you have the time until annex community meeting.

A

All right I mean we can assign it, but when we, when someone, cannot make it or cannot implement this till the next meeting, it's it's. It's okay, yeah.

G

I mean this issue: I, don't I, don't see a problem with I'm sure Florian has this done in five minutes, but.

E

G

To to have this in as a yeah to remind everyone that there's a public holiday and probably vacation, and it's a short week tomorrow for this yeah backlog, all.

A

Right Christian, can you explain this one because yeah.

C

A

Identified it and yeah I'll.

G

Actually, request access, I think that's easier.

G

Yeah I can scroll so identified an issue with the captain's bridge, essentially, I have added to a solos into metal. Oh yeah, Moe so number calls other services a number of time calls of other services and those two SLS I've added, but I did not add them into my SLI provider. So my SLI provider sends back this information unsupported as Ally metric for number calls at a service and for the other one.

G

However, in captain's bridge, this information is not shown and the reason behind this I've also troubleshoot. This a little bit is because I've added this SLO as status info, rather than having it as having it with criteria with targets whatever in my SLO. So this is not shown only for status info. It is, however, shown if the status or if there would be criteria behind it. So this is what it would look like with criteria behind it.

G

If there's an error message- and this is what it obviously should look like for those two info metrics as well- although I would not use red I, would use, yellow or maybe don't even use the color just use the exclamation mark yeah. So essentially the functionality is already there. The color might need to be changed and some if-else statements need to be changed.

G

Any questions regarding that issue or well, anyone opened for it to do it.

F

Yes, it's in the bridge you can assign that one to me.

G

Johannes, can you do the assigning I think that's easier because you have your screen. Alright,.

A

Thanks for giving me control.

G

Now you have to scroll down it's in that list, but it's okay. You can do it here too.

A

A

Then we have one about an environment. Variable configuration service is not evaluated in Health Service Thomas. You brought that wanna up right.

E

Yes, there's the problem that configuration variable for an environment variable for the configuration service, URL exists, but this are the value which is passed. There is not a path through to the method which uses the configuration service and, furthermore, the service should throw an error if there is no configuration service variable set- and this also doesn't work.

E

Yes, any questions about this.

A

It's interesting that we did not orotate that it came up now or in which it I mean. The problem occurs when the pastor actually problem occur.

D

Variable and assume that this environment variable can be used to define the configuration.

A

D

I mean it's not a critical part, but something B has to be fixed.

D

Thomas, would you like to do it? You can reshoot I'm.

E

Not sure if this will, if I get through it next week, but I can try to do this cool. Thank you.

A

Okay and then we have I, think we discussed this one discussed this one also last time it's about the environment scream.

A

The next step is to show which services are out of sync, meaning that when we compare, for example, at the F environment, with a staging environment and def has already a new version of a service deployed, then there should be an indicator showing that showing that the staging stage is out of sync compared to the DEF environment.

A

I'm gonna have actually a question regarding this issue: I've written it down here and I think we can also discuss it now in this meeting. What do you think and is it that much work to also provide the indicators for the problem and the failed evaluations?

A

Because and to be honest, when we just show the out of sync symbol and indicator there.

E

I see yeah, it.

A

Looks like there is something missing and in the future, we want to show also these two indicators there. And what do you think? What about your estimation, oh yeah,.

F

I think that should be possible to display those as well since we already have problems highlighted in the services screen.

C

F

Yeah I would also do that in one task. So altogether.

A

Okay: okay yep, my thought about that cool.

F

Yeah- let's, let's have that in that issue, but since I already have four issues for the next week and as we already heard, the next community meeting is on Wednesday again I.

F

Don't think that I will have time or I don't think that I will get so far to do that. One.

A

Okay, yep, that's that's fine aren't just then we leave it in the backlog, yeah and I. Think the same is true for the next one: that's about highlighting there. They selected SLO in the heat map. It's also preach issue and I think we will leave it there as well right.

A

Okay, then us till to discuss to two issues two or three. The next one is on testing and then platform integration test for manual approval use case, I.

A

Have written I have here provided a test metrics of all the combinations of the approval strategy and I've also written down a concrete test scenario, but I think I explained it last time. I mean right here as someone time to tackle that one.

A

It's silent, I, take it for a No yeah assess the next week is, is really short, I think yeah. We leave it where it is, and still I want to bring up two issues that are related to the self-healing use case. The first one is about extracting the implementation of the feature. Toggling mechanism that is currently part of the remediation service is extracting it from the remediation service and implementing it in a in an unleash service, which will then be part of the captain contrib organization yep. This is this is the task for that issue.

A

D

I would like to mention that we don't really have to find the events for the actions in order to trigger this unleash service. So I will also add here a comment on the issue to the create here. Yep.

A

That's that that would be create just reference. The specification and yeah here it's written down and the tasks itself would be to use the captain service template, which is provided for the go programming, language and based on this template and set captain unleash project all the ways with Travis CI integration and then actually, it's all about moving one function from the captain remediation service to the new service.

A

Additionally, there is also environment variable in the remediation service that can be removed, because there is no need for this environment variable anymore, and here I have written down, also on which you whence this service should react on and how to respond.

A

The ultimate goal of this task is to have an unleash service which is available and released in version or dot or dot one and can be deployed on any captain installation.

A

This would also be very nice task for getting familiar with writing your own captain service as the functionality is there and the template is there, as well as the events I have declared, but it is actually a very nice task for getting started. Writing your own service.

D

D

Would like to mention that we already have a repository which is called.

E

D

Contained in the captain contract, so I guess here we have to do some refactoring to I.

G

Don't think there is Kota net repository. I think this is completely empty. Only has a readme.

D

G

Don't branch ok not sure, what's in there, but we should not reflect or that one to be honest, I think that code is really.

B

Next vacation, until Tuesday I should get pretty far at least thanks foreign.

A

A

Has someone son at some time left over otherwise I would keep explaining and the next issue I think.

G

The time is very at the end of the meeting, so me.

A

And you yeah and yeah every there's, a lot of things currently assigned I think we are good to go for the next week. That's it's perfectly. Fine brings me now back to we did the backlog refinement for the next sprint and also the.

E

A

Part then we have two bullet points: open opened them. There is the announcement that the next community meeting is also on a Wednesday. It's it's on May 27th. We moved it also from a Thursday OH a Wednesday. The same is true for the community meeting on June 11, and this has been moved to June, 10th and last but not least, one additional info that udin wrote a blog post about using promises or scaling issues when you're using promises in an organization. This one is X, is already published on medium.

A

Take a look on that on the blog post, read it and learn more about from users and how captain can tackle the situation's were Prometheus is used and yeah Cap'n also plays an important Impala, important role there, all right some additional remarks or some open questions.

G

If you're watching this on YouTube afterwards, there is always the good first issues. I just want to mention that again, it's on the top of the community meeting notes. If you can scroll up, there is a link to the good first issues. So if you just want to get familiar with captain or with writing coding, captain just look at these and yeah all contributions are obviously welcome. Welcome right.

A

Okay, when there are no further questions, then I will now stop recording and say. Thank you very much for participating and contributing and see you next time. Don't forget it's on the Wednesday next time. All.

G

Right thanks right have a nice evening.

G