How To Shift-Left Stateful Kubernetes Applications Management

[@DevOpsToolkit]

Does the combination I presented make sense? What do you think? Which combination would you use?

[@dirien]

Nice video Viktor. I did not find the gist to follow along with the video. Maybe I look at the wrong place.

[@DevOpsToolkit]

It should be in the description in the "Additional Info" section. Can you double-check? Sometimes I see things others do not have the privilege to see.

[@dirien]

@@DevOpsToolkit Ahh perfect. I see it now too! Perfect thanks! 👍

[@ioannisgko]

Great video as always, with new ideas to research! Just a question, shouldn't the title of the video be "stateful" instead of "stateless"?

[@DevOpsToolkit]

Oh my! That was such a silly mistake. Thanks for letting me know. Corrected!

[@АлександрЗверев-й1х]

Awesome! Thanks a lot.

[@mohamedsamet4539]

Thank you for the video. So helpful as usual :).. I have a question about crossplane: Is it possible to trigger/execute an action just after the k8s cluster created by crossplane become ready?

[@DevOpsToolkit]

It is possible to apply a manifest that depends on some other resource. Such a manifest could contain resources that execute whichever logic you need.

[@mohamedsamet4539]

@@DevOpsToolkit In that case I may create a job that executes any action when the resource aws cluster become "ready"

[@DevOpsToolkit]

@@mohamedsamet4539 Yes, even though it might not be intuitive. You need to patch that job with a value that will be available after the AWS cluster is created.

[@javisartdesign]

Thanks Vicktor! That was really interesting. There is a way to launch a sidecar container with devcontainer or GitPod to have an IDE for coding?

[@DevOpsToolkit]

I don't think okteto has that option. You should probably go with gitpod for that.

[@vn7057]

Question IAC vs Stateful Somehow I feel it will be painful for rebuild/DR need if only stateful Ofcox something ELK we can iac provision cluster but we only able to have stateful design for the cluster But if something Grafana and its dashboard

[@DevOpsToolkit]

The state is (almost) inevitable. Almost everyone deals with data in some form or another. So, the question is only how we deal with the state rather than whether we have the state or not. Now, as you said, managing the state is painful so it's always better to keep the state in third-party apps designed to manage it (e.g., DBs, brokers, etc.) than inside our own apps.

[@vn7057]

@@DevOpsToolkit Ya~ Kafka and DB ship to vendor forsure 😛

[@JaydeepDave12]

Sometimes developers ask me to rebase their test branch from prod branch!!!! (1800 changes 10 conflicts)

[@PeterLunderbye]

Nice video.... Realize i also need to check-out acorn

[@DevOpsToolkit]

kzbin.info/www/bejne/q5uZeJSkZpWKn8k

[@mateuszszczecinski8241]

I think it is not convienient to use yaml file to menage db schema, when almost every body is used to use sql. Meybe it's better to replace SchemaHero with for example kubernetes job running tools like flyway, db up, etc.

[@DevOpsToolkit]

In this particular case, I prefer SchemaHero mostly because it fits well into GitOps story (e.g., Argo CD, Flux, etc.). Nevertheless, the argument is indeed valid, and I as well prefer writing DB schemas using SQL.

[@onemanops]

Me neither