keptn Community, 17 Sep 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keptn Community Meeting - September 17th, 2020

Description

September 17, 2020
Diary - what happened last week?
Preparation and some work for a 0.7.2 release
Bugfixes & Small improvements
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:
Backup Instructions for v0.6.2 and 0.7.x: https://github.com/keptn/keptn.github.io/issues/635
Lighthouse-Service: Default SLI Provider: https://github.com/keptn/keptn/issues/2317
Lighthouse-Service: Send get-sli event when no SLO file available yet:
https://github.com/keptn/keptn/issues/2318
Allow http upstream repositories: https://github.com/keptn/keptn/issues/2336
Andreas:
Allow to disable Bridge secret for basic auth: https://github.com/keptn/keptn/pull/2350
Correct version in swagger: https://github.com/keptn/keptn/pull/2349
Johannes:
CURL command that creates a DT-Dashboard for SLOs https://github.com/keptn-contrib/dynatrace-sli-service/issues/76
Ermin:
https://github.com/keptn/keptn/pull/2372
Show git upstream link in project overview or info how to update project
https://github.com/keptn/keptn/pull/2366
Feature toggle for version check and api token info
https://github.com/keptn/keptn/pull/2367
Hide api token and keptn auth command
Christian:
Bridge Download Link to CLI: https://github.com/keptn/keptn/pull/2333
Remove scheme from swagger.yaml: https://github.com/keptn/keptn/pull/2328
Community Contribution Guide and Issue Guide (bug first policy): https://github.com/keptn/keptn/pull/2321
Increase entropy for api-token
https://github.com/keptn/keptn/pull/2345

Roadmap → KEP / SPEC / Issues
KEP 06: The next generation of Shipyard: https://github.com/keptn/enhancement-proposals/pull/6
Keptn Quality Gate capabilities
Work in progress:
https://github.com/keptn/keptn/projects/29
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

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

A

Hi everyone and welcome uh to this captain community meeting september, 17 2020..

A

um As always, we are following the uh cncf tour of conduct, so that means uh please be nice to each other and we have quite a packed agenda for today, so I will be handing pretty soon over to uh to johannes or to florian or whoever wants to to start.

A

I just want to um mention a couple of things in the beginning. If you have some questions during the meeting, please indicate the questions in the q a section, so we can make sure to address them. We have a little bit of a more strict policy now of uh promoting people uh to panelists. So uh basically the policy is, if you have some prs to share. If uh you have basically shown yourself as a captain contributor, um you are invited as a panelist and you can um can contribute by screen sharing also to the meeting.

A

If not, then just please indicate.

B

A

Questions in the q a section we will be happy to address them during the community meeting yeah if you want to to stay up to date with captain. um I just want to announce. We still have our group groups uh where you can uh take part, and uh I think we already have a couple of people taking a look at good first issues, but still the links are provided here. So if you're new to this meeting, take a look at our good first issues. That might be a great start with that.

A

I think I'm heading over to johannes. Do you want to kick this part off.

C

No, uh you cannot directly go uh go to uh florian, okay,.

A

D

We're moving over to.

A

D

Maybe two words from from my side: uh the diary is something we started last week for people just to fill in maybe things they believe that significant, uh but just to summarize the work, but it's nothing. We want to speak about here, it's more. If you just want to read up. um This is like the teaser to read up yeah and another thing about the good first issues. We also have good first issues for captain's bridge.

D

So if you are an angular developer um or if you want to learn angular, even uh just take a look at captain's bridge.

A

Cool um so thanks so much and then I'm heading over to florian all right. Thank you.

E

E

Okay, today, I have four pull requests that I wanted to show you, so the first one is an extension to the documentation.

E

So let's have a look at the issue first, so basically, uh what we wanted to do here is to provide instructions on how you can back up your project related data. You have configured in captain in order to restore it afterwards in case an upgrade fails or if you want to reinstall a captain from on your cluster.

E

So basically what we have done here is. We have added a new section in the docs, for both the release 062 and for the current 071 release.

E

So in the 062 release you will find it here on the install cabin and backup and restore captain, and here you will find uh all the instructions that you need in order to to perform the backup.

E

So in essence, it's uh first to get all the data from the configuration service that contains the git repository for your project and then this using this instructions. It will be stored on your local machine and then you also have to back up the mongodb data and also store it on your local machine. So this is done using the the mongodump tool. You actually don't have to install it uh on your local machine, so uh you just have to use cube cattle exact and on the mongodb pod there.

E

This tool is already installed and you will then only have to uh fetch the files from the part using the cube cutter cp command don. Then here we have some instructions to backup your git credentials in case you have configured upstream repositories for your projects and the next section basically describes how you can upload all these artifacts. You have stored on your local machine previously into the new captain, installation and yeah for the release. 071.

E

You will find uh this section in operate, captain and backup and restore pattern, and then essence uh the procedure is pretty much the same. There are some minor details that are different between those two versions, but that's covered in this documentation.

E

Are there any questions about that.

E

All right, if not, I will go on with the next pull request. So this one actually, the next two are related to the lighthouse service. So the first one, let me explain so basically, for example, a new feature in the dynatrace sli service allows the service to create a slo file from a dynatrace dashboard, with certain tiles that are made according to an convention and based on those tiles in the dashboard.

E

It's able to generate a slo config, and that means that if you use this feature, you don't necessarily have to upload a slo file anymore in order to do an evaluation.

E

But previously the logic of the lighthouse service was that that if there was no sno file available in the config repository for your project or your service, you just assumed that you do not want to perform evaluations for that service and then immediately send an evaluation done event with the status passed in order to continue in the deployment process or just to give you a response for for the quality gates, evaluation.

E

And now the logic is the following. So uh if no slo file is found, it doesn't send an evaluation done event out anymore immediately, but what it does is it will send a get sli event and then, in that case, for example, if you use the dynatrace sli service, this service will now be able to generate the slo file as I've previously described, and it will actually upload this generated slo file to your config repository and then the evaluation service can lighthouse service can use that one in order to perform the evaluation going forward.

E

F

E

Any questions about that.

E

All right, if not, I will continue with the next extension to the lighthouse service. So if you have used to use the service uh previously, you know that you have to define a sli provider for a certain project using the captain configure monitoring command and using this command. You had to either set your sli provider to prometheus or dynatrace or any other sli provider that that he wanted to take over the responsibility of fetching sli values, and now we have made a small extension to the lighter service.

E

So now it's uh possible to define a global default sli provider. So you don't have to to execute the configure monitoring command for each of the projects you have created anymore and the way this works is. uh We have introduced a lighthouse, config config map, and this contains the key sli provider and this can be set to the global default. You want to use going forward and, of course, if you uh execute the configure monitoring command for a certain project, it will then overwrite this global default.

E

Any questions about that.

B

Just a quick one from my end really cool. uh Is there also an option to configure this through the cli or through the api, or is this we just need to do it to the configmap.

E

Currently, this is done by creating the config map, for example using cubecuttle, but of course, because it would also be a good idea to provide a a cli command or an api command for that. But we can, of course,.

D

Discuss this thanks, yeah feel free to create an issue, and if people upload it.

E

All right and yeah, the fourth one is uh very little pr, but hopefully helps some people out, for example, if you want to use captain in a workshop setting, so what we have done here is we have allowed to not also not only allow upstream git repositories using https, but also using http.

E

So if you want to to use the repository that uses http, you can now do so.

D

Maybe uh just to add on this, because we have several people in slack comment on similar issues. If you have the need to use https, but with self-signed certificates, it's something we currently cannot support, uh or at least not easily support. So if that's your use case, please get in touch with us, write an issue or whatever and document your use cases, because we can only support what people report us. Obviously.

E

All right, so I think I'm through with all pr so now I will hand over to android.

G

Yes, cool. Thank you.

G

Oops, that was the wrong one.

D

Thank god, there was source code.

G

Better sounds good than something else.

G

What do you see now.

G

Meeting documents perfect, thank you now, I'm showing you a rather dangerous feature, but it can be hand handy.

G

So um what I've implemented in our hemp chart is that you can now disable the basic authentication for the bridge, and you can do this by setting the value bridge, dot secret, that enabled equals forms, and then there is no basic authentication.

G

So the use case behind this is, for example, let's consider you have a proxy which already does the authentication for you. Then, of course you don't need a second authentication um using the basic out, but but please be aware that if you disable the the secret here, the basic authentication um people will be able to access. Your api will be able to access your api token and so on.

G

The second issue I would like to show you is a small bug, so you reported it.

G

So in our expected documentation we always shown version 0.10, even if you, for example, have instant 0m2 and the reason for this is that the version was not properly set in our regular documentation. So inside our dockerfile we can then replace exactly those string. This version develops string now to the correct version. If you go into the dockerfile.

G

Then you can see here now the replacement of the version now works and we are setting the correct version, for example 0.72.

G

Okay, that's from my side any questions, otherwise, johannes you're. Next.

C

All right um thanks. I want to keep myself short because this week I was just working on on the dynadres integration part, um and this is actually related to the issue that florian was was talking about. It's the capability that the dynadres sli service can read: slis and slos from a dashboard from a dynadris dashboard and to help dynadrace users to create such a dashboard.

C

I provided a curl command, but I was just doing the call command and many thanks to andy who actually framed the dashboard and decided what we should put on, and I was then just taking the dashboard and putting everything into a curl command. But let me show that now to you.

C

Therefore, we jump over to the issue itself or no. We go directly the resource code.

C

Here then, on the master branch right now you find now a folder, that's called dashboard, and here you get the curl command with a couple of parameters that you have to set in before executing that command, and when you do that, you will end up in your dynadris, thailand. You will end up to get a dashboard that looks like that. One on this dashboard, you have a tile or multiple tiles for service performance indicators.

C

Then there are a couple ones for service errors and throughput and others are for process metrics and host based metrics and the dynadress sli service is now able to understand this dashboard and retrieves their slos and slis. Based on this configuration- and that's pretty neat because then someone who is using the dynades sli service does not need to write slos and slides by hand.

C

And when you want to learn more especially about this feature, please go to the issue and there I posted a link or yeah. I posted a link to a youtube video and, in this youtube, video the features and also the use cases around. This dashboarding mechanism are explained pretty well.

C

All right yeah, this was it from my side. Let me hand over to amin.

F

Thank you very much. um I'm going to present three pull requests regarding the bridge today. So the first one is the uh request, show git upstream link in project overview or info how to update project- and this is a rather old one- actually reported by andy- uh to show the guitar uh for the project and um yeah, since uh we now have, for example, basic uh basic uh authentication for the bridge. uh It's secured to show this upstream.

F

So in the overview of the projects, you can see either the the upstream if it's configured or if there is no upstream link, then you get the information that there is no upstream configured and also the link to the docs. How to update your project to set the github stream.

F

The second pull request is a feature toggle for version check and api token info. So we have now in the user menu on the right top uh the api token and the captain auth command, and we also have the version check, feature and.

F

Now it is, I will just show it in the code possible to enable or disable actually those features, just by setting an environment variable. For example, you can disable the version check just by setting this enable version check environment to false in case that, for example, you have a managed cluster and you don't want or you cannot uh do the the check or the bridge cannot do that check.

F

So you can disable this feature uh completely, and you can also disable um the information or this uh yeah api token and captain auth comment uh buttons and text fields in case that, for example, you disable the basic authentication for demo purposes, or you provide access to your bridge to several people for demo purposes. Then you can just hide those fields yeah um and the third one is also regarding this api token and captain auth comment.

F

So uh we had the enhancement that now you can yeah easily copy your caption api token, and also that auth comment from this user menu on the right top side, and now we have enhancement on that one.

F

For example, when you do a recording uh of your bridge, uh you don't have to worry that uh you will leak your api token since now, and I will do a live demo here, um you initially see only the label, but you can still copy the value or you can also show it, and here I can safely show it because you won't see the api token, but also copy the whole captain of comment uh so that I can easily uh authenticate to this captain um yeah with this bridgest um yeah.

F

That's the three pull requests I wanted to show you. Are there any questions? I.

C

Mean can you jump back to the screen or to the live team.

F

Yeah to the live demo, of course,.

C

uh What's the meaning behind captain art command that that's below? Oh.

F

There is the link to the docs since we had it in the old version, also so yeah that.

F

I don't know I left it there, because we had it.

D

F

In the previous version as well uh linked, I can also remove it or I thought it's, maybe a nice information just yeah to know what to do with or what is happening uh when you execute the captain. Auth comment.

B

A quick comment on my end: pretty cool, uh maybe an additional link up there would be the link to the swagger ui, or would that also make sense.

D

Yeah link to the api, basically to.

B

The api basically yeah, because that.

D

Redirects you to swagger.

B

Yeah because normally I would probably copy the api token, then click on the link and go to swagger and then I can try things out and.

F

True, actually, we have a separate issue on that. One to yeah maybe update uh those links, and I think that one is covered there cool, but it's um it's sure, easy implementation to provide here that the link to the api.

D

And I mean, as you uh are on that screen already and you seem to have merged my feature. Also in your branch, uh can you open the yeah? Oh yeah,.

B

D

One of my first pull requests: I'm going to present is the download captain cli link up there. If you can just click on it, there we go. It opens up the github releases page or if you have a certain version deployed. So it reads that deployed version of bridge and captain. It would redirect you to the correct version on github. So right now this was just a releases page, but if you deployed it, yeah.

F

The problem is that I have here the last latest version deployed, so it could not yeah actually determine my version.

D

And then I would take over the screen sharing, because then I can quickly show the pull request as well thanks for demoing my feature. um So this is the pr for the feature. Essentially, we have introduced uh environment variable.

D

Let me look it up there. We go called cli download link that you can override um or it's overwritten via the helm chart anyway, but you can enter the link if you want to or overwrite the link, if you have the use case tool or write it whatever and it's being, let me check yeah, it's being overwritten here in the hand, chart uh wire values yaml and it's defaulting to the app version that you have deployed so for zero, seven one.

D

It would default to zero seven one and four zero, seven, two, the next one and so on all right. Any questions regarding that bridge feature.

D

um Also, like I said before, we have good first issues for captain's bridge. That was one of it and I basically just took a look at it and I was able to easily do the good first issue and get in touch with the source code of captain's bridge. So I can only encourage all of you to take a look at the good first issues, all right. The next one is something that I noticed when I deployed my captain to google kubernetes engine and I did not have an ssl certificate.

D

I noticed that I need to open the issue. There is a drop down uh in swagger. Actually, it's not linking it. Is it whatever there's a drop down in swagger, where you can select https or http, and I was under the impression that swagger is able to automatically determine the correct scheme, and I was right so I basically just had to remove something because we had the scheme selected, sometimes some actually yeah.

D

We had it selected to http and in the docker file we extended it with https and I've just removed everything ready related to this, and now this regular ui is automatically determining whether it needs to be http or https.

A

Thank you so much for for this feature, because I always uh hit the problem of uh in my first try. I, uh it was always going wrong and then uh I realized it's the first. It's the wrong scheme.

D

And I changed it and it was, uh it was.

A

D

It's actually worse than that, because if you scroll down this drop down just vanishes, it's out of your out of your view and uh you never will know that there is such a drop down. So my first bike report was actually that we're selecting the wrong scheme, but then I noticed actually swagger ui is doing that uh yeah cool.

B

D

Then the next improvement I want to share with you is a little document called contributing.markdown. Okay,.

B

D

Is my pull request, so there we go so we actually have this contribution guide already, but I've extended it a little bit and like like, we added some guidelines on how we label issues, uh but we also try to make it a little bit clearer how our development process is working for people that are just looking at the project and don't know where to start.

D

So I encourage all of you to just read through it, and maybe one of the most important things that has been added is a bug first policy and we basically try to label our issues if you post any or if actually also, we post any. So anyone that posts an issue we try to label it with type critical bug, feature improvement or chore. um Basically, also in that order, critical issues need to be fixed right away.

D

It's a defect that is causing some sort of problem that people will lose money or value or time or whatever. So if you just have a typo, that's not a defect. Obviously, but if the if the issue is like uh my captain, installation deletes my whole kubernetes cluster, um then that would be critical and uh yeah similar for buck. That's something we've documented that should work, but it's not working as intended.

D

That would be a bug so either we need to fix the documentation or we need to fix a code issue that doesn't matter, but that's something, that's a bug and other than that. We have features for new features and improvements. If you want to improve on an existing feature and the chore is basically just something that doesn't have a direct value for end users, but has direct value for the development team, any comments or questions on that.

D

Otherwise, uh please just read the contributing guide and asking slack if you have any issues or comments.

D

Last but not least, we did a little bit of hardening again and we noticed that the api tokens that we generate and the admin passwords that we generated, I could benefit from a little more uh entropy.

D

So we've increased the admin password from 20 alpha numeric characters to 45 alphanumeric characters, as this is a password that you usually will not need. But if you have the password or if you guessed that password you get admin access to the mongodb, which is not good, that's if you have access to the cluster. Obviously, if you don't have access to the cluster, then everything is fine, but the api token is something that we saw with 20 characters is.

D

This is leading to an entropy of approximately 120 bits and that's not state of the art at the moment, and we increased it to 45 bit uh sorry to 45 random alphanumeric characters which leads to 256 bits a little bit more than that and that's considered state of the art.

D

If it's random enough and the randomness is actually coming from helm, not from us, so that should be fine yeah.

D

Any questions on that.

D

If not, I think I'll hand over to johannes.

C

All right: okay, um what's next well one or two sentences about the roadmap, we still are driven by the captain hansen proposal 0.6, which is about the new shipyard, controller and yeah. We are still working on that one, but I also want to mention.

C

We currently can observe that more and more folks out there are using the captain quality gates, and this is also a reason why more more feature, requests and improvements, especially in this regard are coming in, and this is also the reason why you will see that we are working on on improving this capability um yeah going forward, and now, let's jump over to the working items, and let's have a review on.

C

What's in there ready for review column, um all right, yeah yeah, you showed us actually that um the project tile has already um a link to the upstream git repository. Do you need support, support to get this pull request, merged or.

F

C

F

I I opened this per request like half an hour ago, so uh let's leave it in ready for review. Okay until the pull request is closed. um Maybe there will be some discussion still yeah. The same is for the disable. The version check feature um actually the waiting for a contribution by andreas to allow to set the uh environment. uh Why the helm chart.

C

Okay, got it yeah and leave it over open, and for that one uh can I do the pr review.

F

C

F

It's the same, the pull request is still open, so um maybe there will be some discussion or something to change and of course, if everything is okay and the pull request is merged, then I will move it to the done perfect.

C

Then, according to our bug first strategy, we pulled in this issue uh yeah. It was a a bug in the in the lighter service that has been introduced due to a change uh florent. Can you give us an update on that? One.

E

uh Yeah, basically, so what happened here is that when you send it an evaluation or a start, evaluation event for an unknown project or services stage that doesn't exist in your cabin installation.

E

The lighthouse service, basically yeah, didn't give appropriate feedback about this case and now what it does in case uh this uh happens, it will send an evaluation done event with uh status, failed and uh error message describing the error yeah. So the request has already been made one hour ago or something, and as soon as it's been reviewed and approved, we can merge it into the release branch. All right, perfect.

C

And then the next one is this: are you still working on that one, or is it also ready for review it's about uh getting the git commit into uh the response of the configuration service right.

E

uh I think andrea said uh one remark that I wanted to have a look at, but there were some issues for 0.72 coming in. Didn't have the time yet to to look through that.

C

Okay and and the last one um is that one finished.

E

uh This one is finished uh and can be closed. I just didn't want to show it here, because it's not really exciting. It's more an internal build pipeline.

F

C

All right yeah, then a quick update on the items that are in the in progress from my end, yeah. I want to improve a little bit the instructions to deploy the android service as it's misleading at some parts and not that consistent.

C

um Actually, I'm I'm done with that work and would lead to someone would need a review from someone, but I will reach out to the team. Then andreas you did refactoring of the.

G

Service, actually I'm working on this, so um we had the problem that we locked an error multiple times and the log messages were even yeah contradictory.

G

So I did here in refactoring- and I already have a draft here, but now I have to test it and for the next um issue. This is basically the same. The implementation is there, but 0.72 is most more important than this new help service. So I postponed it. Okay, yeah.

C

That's: okay, all right yeah! We don't need to discuss the assigned and committed issues as we did last time, and I will directly jump over to the back clock where we have the items for the next sprint for the next week and the first one uh I want to explain.

C

Well, I was using the captain generate support archive command um and I was- and I have a captain installation with where I install captain in a specific namespace, not in the default one in another one, and then I executed this command and that did not get there on a reason or a hint why the the archive was empty and the improvement would be that when there is no default captain namespace, then the command should tell you that you have to specify the namespace to also retrieve the logs from from the captain installation.

C

This is the first improvement. It's actually not a bug. It was my mistake, but still I want to leave it like that and there's also a second improvement um addressed by this issue, and it's about telling the user that this command also requires um a config as a cube, control configuration or a cube context, because we need directly access to the kubernetes caster to get all the logs, um which are then bundled into um support archive.

C

And this is what we want to add to the command. And we also want to double check to verify or to ask the user if he or she is still confident that he has the correct queue context. Configured.

C

This one is is labeled as improvement, but I would also label it as a good first issue. It's all about the kept cli and just improving the error messages and the user experience.

D

F

D

Takes it uh from the community? Actually, I would like to label it just as good first issue, but I'll do it. If nobody else does it, um but we need it rather sooner than later right. True yeah, then I'll just take it. If anyone else still wants to take it, I can give it away afterwards. So.

C

All right, then, uh yeah.

E

In case you don't have the time to do it, I can do it as well so yeah, let's.

D

Align will sync if we need to switch okay.

C

uh Question I okay, when we do it one one by one.

D

Yeah, but uh I think the next uh prepare release. 072 is just a meta issue. There's like no true yeah. I think it's more a reminder and the summary.

C

Issue is it yeah correct? Actually, it's not. There are not tasks in there. It's just about getting the documentation for for the next release done and with a double check of a couple of things, and also there should be an explanation of the new features that were implemented uh for the bridge, but yeah.

D

Maybe we can make the issue yeah, but maybe we should move it down like uh that should be when we are ready to release. I think and I'll talk about the next one.

B

D

Documentation around captain qualitygate there we go so uh we've noticed that the the documentation for captain qualitygate is a little bit unpolished. Let's put it like this, and we have some tasks that we should do like define, what's needed to set up a quality gate and how can we integrate such a quality gate in an existing pipeline?

D

So it shouldn't be like a tutorial per se, but like a reference document and yeah, we need to just need to enhance on that information, and uh I think uh we have it split into two uh into two uh sections, like the concept, quality, gate and qualitygate here, I'm I'm not sure did you specify where we should put the new information, johannes.

C

um I did not specify, but the concept should go into concepts.

B

C

Gates- and they actually did the how to's and how to do certain things in this regard, they should be placed into this section.

D

Yeah, okay, yeah. It makes.

C

Sense right now we are just explaining what is an slo and sli and how you can deploy an sli provider, but there is no guidance how you can create a quality gate, how you can delete one or update one and so on.

C

D

So yeah so who's up for that one.

C

Yep then I will take it and I will um pull all the information together that we have for this. For this task and we'll come up with improvements of the documentation.

C

All right, okay, the next one.

C

Yeah, as we explained today, two times or mentioned two times, uh the dynares sli service is now capable to understand the dashboard and to to get all the informations from a dashboard and extract it into saline slos.

C

But there we are missing a couple of unit tests to to validate that all the the functions are working properly, and this is what this feature is about. This issue is about improving the test coverage of of the dynades sli service, and here I added a couple of chore tasks, because what I also figured out is that the dynadres sli service does not specify the spec version in the slo and sli.

C

Then it's dealing with a dynadress.com.yaml file, which is actually not there, and this should also be uh double checked, and here it needs some refactoring. This function because, as I was testing out today, the filter uh filtering mechanism, I figured out that this parameter is still is not needed, anymore um yeah and it's obsolete.

C

All right yeah, this is to some extent, are refactoring and and increasing the test coverage I'll. Try to take this.

B

As I guess, I'm responsible mainly for the poor test coverage of the writing code and not.

C

Creating tests? Okay, thank you. Thank you very much andy ford for taking it.

D

All right, the next one is refactor error, handling and logging. It's similar to the one ted andreas has already talked about. Andreas has taken this issue for the data tree service already, and he's also posted a very similar issue for the danatris slide service. So essentially, we also need to slim down the error logging a little bit so errors are in the same line and not split over multiple lines and basically yeah log errors and lock. Debug messages were appropriate and yeah.

D

Who is who wants to take this? One.

E

uh I would like to do that.

D

Thank you, florian.

C

Okay, okay, then, then, what we also want to implement in the dynadres sli service is the is the flag that allows us to enable or disable the the feature of the dash of extracting the dashboard and the flag should be um named.

C

Like that um proposal, where I say it's a slow slide generation flag and it allows to configure it as enabled disabled and initial initial enabled means that every time the dynamics sli service gets triggered, it also fetches the information from the dashboard disabled means that this feature is turned off, and there is no interaction with the dynadress api and initial means that when get sli event is coming in, that has no indicators in the payload. Then the dynadres sli service should do an outreach to the dynadres api and should get the information from the dashboard.

C

But just once and when the slis are there, then um it should not do this step again.

C

Are there ideas on this feature on this proposal?.

D

I'm wondering if we should rename the flag to automatic, slo slide generation because or fetch slo sli from dashboard. That would be even yeah. It gets quite long, but it's it's as low as the light generation. This is a little bit vague in my opinion, so like fetch as a low sli from dashboards.

B

But my question is: why is there such a big overhead with the api? It's not it's not querying the individual dashboards, because the current implementation gets the full list of dashboards with a single call and then iterates and only picks the dashboard. With that name.

D

um If we made the problem is uh with another change that we have, that we always fire the get sli event regardless. If there is an slo file or not, this will have an effect on all onboarded services and all services that have evaluation, get get the sli events, because now cardstb, for instance, in our in our examples, cardsdp will also trigger such and get sli event, and then this will also fetch the dashboard data and for every get this live, and you have basically an api call or maybe more than one.

D

At some point uh I mean nice that you implemented it in such a way that it's only one, that's good, but still this could get out of bounds really quickly. If you think about huge deployments, and then it makes sense to make this configurable.

B

Yeah, I'm still not sure about the feature. Flick though.

D

We can have it enabled by default. uh It's it's really just the fact that we are able to turn it off. If, if somebody wants to turn it.

B

Off yeah, I would rather bind this with the dynatraceconf.dml file to be honest with you and put it there, have it off by default, maybe, and if somebody decides that they want to have this anyway, then upload the dyna choice, convo, where they say search for it. They either say uh reference a dashboard by its name, which I've already implemented or say. Query I think, there's I don't know a discuss this, maybe offline yeah.

D

Because this will go into.

B

D

Can you edit on the issue, then maybe and we'll just uh postpone the implementation for now.

G

D

I don't know a couple of days or until the next meeting, because the functionality is basically there, it's just a feature flag to turn it on all right. Let.

C

Me move it down.

D

C

The next one area.

D

Captain cli updates, okay, so uh basically the cli should provide some information, uh not just that there is a new version available of the cli all right there. It should provide more information uh how to update the cli. So right now it just says: hey, there's a new captain, cli version available, and we just want to extend it with a little text that says yeah you can get it by visiting captain.sh or maybe even captain.h, docs, quickstart and yeah. It's just a text change and that's a good first issue.

D

So if anyone just wants to contribute uh with a text, change.

D

I don't know, don't know, should we just leave it open in the backlog, because there's no point in assigning this right now and if.

B

Somebody wants.

D

To take it, or we just do it quickly before the next community meeting feel free.

C

Then, let's go to the next one: it's now we are uh yeah today, issues related to the bridge and this one is enabling highlighting in the heat map. In other words, when I take a look at the heat map, there should be some indication that is telling me on which evolution result I'm currently looking at and in this in my proposal, it's a border and also some shadowing with a yellow, but also feel free to come up with it, with a better way of doing the highlighting and based on our quick research on this topic.

C

We also figured out that the underlaying component that is used for this feature is already capable to to provide this highlighted, highlighting mechanism.

C

Okay, who wants to take that one.

F

uh Yeah you can assign it to me.

F

Actually. I would uh like to point out something to the the working items, because there are uh two issues that are not zero: seven, two related now in the assigned and committed, maybe just to move them to the backlog and yeah, take the ones that are now four zero. Seven, two okay uh yeah with higher priority agree, but actually the the nxf5, which is for zero. Eight. We will not uh merge that one two yep.

G

F

This yeah actually those three refactoring tasks. We will move it back.

C

D

All right, okay, the next one, is a rather big one, also for captain's bridge, so our evolution done screen uh needs a little bit of reworking, um but this is for zero, eight zero. um So do we want to.

C

Discuss it now, no, actually, it's for uh for all right: okay, seven, two yeah.

D

So the idea is that we do some little bit reworking of the screen uh to have this a little bit uh better from like top down and you get the slo result on top. So, like the orange bar, you can see there's a low overall result. Then you can see the breakdown with the heat map, and after that you can see the real breakdown based on the sli values, with a nice table that you can sort and some criteria in there actually is sorting part of no.

C

To keep this or to keep the scope um to not make it that big from the scope, I excluded the part of of showing the sli breakdown in a table format. It's the focus of this issue is mainly restructuring and providing proper information uh where we have the information and it's just a way of, or just an improvement in visualization, because there are a couple of of um ideas that were brought into the backlog, and I just scrolled through them and took a look on each of them and then pulled them.

C

Those together into this issue and yeah here are the tasks explained what to do.

D

All right, it's still a rather big one yeah who wants to take it. Who has got some time.

D

If there is no time we.

F

uh Yeah you can assign it to me, as I already mentioned, the ones that are 072 uh related, of course, have higher priority. You know, uh so I would take that.

C

Thanks very much.

G

And that's for zero, seven: two! That's correct! Okay!.

C

Yep as more people are asking for quality gates features, um we also want to get better there still.

G

But yeah, if you have a deadline, then I guess the deadline is more important.

C

Yep this brings me to the next one: um it's not for the services screen, it's more for the environment screen where we have a stage that has more than five services, then there should be.

C

um Then then the services should not be here um should not show up in this way, but just each each service should be in its own uh row yeah, and they should also be ordered according to their to their last root event. That happened to this particular service.

C

Plus, when there are more than five services for a particular stage, then they show more button or show more label should allow you to to pull in the next section of our next part of services. Actually, this one is already available. This feature in this tile, all right yeah, but um has someone time to work on that.

F

One yeah, if not just, leave it in the backlog and.

C

When there's time left.

F

C

All right yeah, we discussed this enable disable the feature.

D

And meta issue.

C

Met the issue and this one, we also moved to the backlog the last time. Then we are here at the point where we again want to to mention that we always have a couple of good first issues and help wanted issues which should be outreach to the community to take a look at those pick. One and just leave a comment there and we will assign it to you when you want to get when you want to get started. Working on that. One.

C

All right, then uh yeah we are done with the with the working items and with the backlog requirement and also discussed the work for the next week.

A

Are you honest, we had a question from the community about the the next release, since we saw it as an item uh in the in the spring planning. Is there any uh exact date or is it like? Should it be in the next sprint or in the next two sprints or do? Can you give some some update on this.

C

A

C

I'm sorry, I cannot give you an update on an exact date, but you can expect a release of 0.72 in about two weeks, but not on a fixed deadline. Right now,.

D

But you can you can look in the issue list and see what's in the issue list, obviously for milestone072 that should give you a good estimate on whether the release is almost done or not almost done.

C

Let's, let's always take a look into this milestone, and here you get an overview of the features that are relevant for 0.72. Thank you.

D

All right, I think, we're done. um Are there any questions uh from the community still.

A

um If there are no questions or if there are any questions, please uh put them in the chat. uh I just want to announce that for next week um we will have in the first couple of minutes a special um celebration for our new community rockstar. So um you might know that we are quarterly awarding community rockstars. um We already awarded the first community rockstar.

A

Those are community members that show excellent in a couple of different categories uh like contributing to captain talking about captain uh and uh for for this quarter uh we are awarding the community rockstar next week. um So if you're also interested in contributing to captain you've seen with a lot of angular good first issues, we have golang.

A

We have a couple of other issues open if you want to start contributing, please feel free to do so uh and also join us in our different communication channels, uh so that we also know about you joining us in making captain better.

A

um If there is no more question, is there anything left from the team you want to share.

A

Okay, then, thanks so much for joining this week and see you all next week in the community meeting and in the meantime, also see you on slack slack.captain.sh or in the cnc.

A

Please feel free to reach out and uh bye.

F

F

F