SpotiByeAds

SpotiByeAds copied to clipboard

Hi,

I read the below message at your 2020 roadmap.

Kubernetes helm chart based onboarding and setup. (In-Progress, under analysis/design.... help needed!)

But where's in-progress helm chart?

I want to try it and make PRs if needed.

Thanks,

We are yet to start on the effort, if you are able to contribute back that would be really great!

We have the following components running in kubernetes using helm and helm file:

• bitnami mongo db
• old hygieia-api (looks like the spring 5 version is not actively developed)
• new ui
• bitbucket collector (old version as it looks like the spring 5 version is not actively developed)
• jira collector (old version as it looks like the spring 5 version is not actively developed)
• jenkins collector (old version as it looks like the spring 5 version is not actively developed)
• helm collector (old version as it looks like the spring 5 version is not actively developed)

Everything is running fine but the new UI does not work very well.

• does not configure code collectors correctly so no data is pulled.

We are willing to push our changes back but first we would like a running setup.

I would be willing to do a video call to show u what we have and if you would like a k8s/helm contribution back.

Please get back to me if this is of interest to you guys

Here's a very naive first pass at a Helmfile+Helm deployment of Hygieia.

I see this has been sitting around for 2 years.

@balnbibarbi - good first start. I don't mind taking a secondary stab at this and think of a helm chart that has options to turn on collectors in the values file.

Hygeia team let me know if you want a helm chart deployment pattern.

I see this has been sitting around for 2 years.

@balnbibarbi - good first start. I don't mind taking a secondary stab at this and think of a helm chart that has options to turn on collectors in the values file.

Hygeia team let me know if you want a helm chart deployment pattern.

That would be really awesome, please go ahead with your contribution. We would greatly appreciate your contribution

I can start it but it will take a bit as we may need some pre-req work done. This will take a bit however ¯_(ツ)_/¯. I see images out on dockerhub but they may be stale. I only found this... https://hub.docker.com/u/capitalone - let me know if there is another container repo you may be using.

Helm will be the mechanism to get this deployed in k8s, but container design/management will be the key to success.

Think of do we make containers for each collector or do we allow the collector to be added in a bootstrap way? Does that question even matter? Should we just make containers for all repos in hyegia? Either way how is the hyegia group going to maintain the containers (build, publish, maintain)...

First things first though is just getting a vanilla hyegia running. I will start there an iterate. I will link back in comments when I get something working.

Hi Doug, if we have a separate container image per collector, anyone who wanted multiple collectors could always bundle the separate containers into a single pod, or have multiple Deployments, each for a single collector. On the other hand, if there is a single image containing all collectors, then it can't really be split apart, so everyone gets all collectors even if they only needed one of them, which would be bloat. In general, it's best to err on the side of too-fine granularity of container images.

Can you elaborate on the "added in a bootstrap way" option? I'm not sure what you mean.

Ultimately, we need a CI pipeline to build container images automatically from the git repos upon commit. It's possible to do that using GitHub Actions, but there are many other ways too. Let's discuss CI elsewhere, and keep this ticket about Helm.

Also I'm very aware that all I did was a very early first pass. I just wanted to start the ball rolling. Please share whatever you end up making very early, even if it's very rough. I'd like to iterate on it with you.

OK I think granular containers are better as well. Therefore, we need to start thinking about where you want to build these... what CI tool and so on and so forth.

Bootstrap - look at grafana helm chart or sonarqube helm chart with plugins. But ignore it since we are going containers. At least for now.

grafana example -> https://github.com/grafana/helm-charts/blob/5b2675d6b5b8458367e9f18340cda7b1166d453b/charts/grafana/values.yaml#L742