Skip to content

Adding the application Flower #31

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 164 commits into
base: main
Choose a base branch
from

Conversation

adityapgupta
Copy link

Current progress for adding Flower to AIOpsLab using a k8s cluster

@adityapgupta
Copy link
Author

@microsoft-github-policy-service agree company="Indian Institute of Science"

@HacksonClark HacksonClark marked this pull request as draft March 10, 2025 23:05
@HacksonClark
Copy link
Collaborator

Converted to a draft since this is still a WIP. When ready for review just mark as 'Ready for Review' and add me as a reviewer :)

@HacksonClark
Copy link
Collaborator

@adityapgupta Is this still a draft?

@adityapgupta
Copy link
Author

adityapgupta commented May 22, 2025

This is almost complete, I just need to add the readmes and see if I can work around the sleep issue. I have fast forwarded the aiopslab-applications change too. I will remove it from draft as soon as I am done with this

@adityapgupta
Copy link
Author

Hey @HacksonClark , I have removed the use of sleeps in the fault injection (sorry for the delay). Need to make a few small changes now, would appreciate ur advice here:

  1. Flower doesnt use stuff like Prometheus and OpenEBS. Turning all of them off everywhere for Flower is a bit inconvenient. Is it ok if we expect the users to have the kind cluster running for this?
  2. I have added a LLaMa client for my testing which uses a Groq client. Do you want it to stay this way or should I swap it out for something else?
  3. Where should I update the relevant requirements?
  4. Which readmes should I update for deployment and testing details? I have already added a small one about the Flower application itself.

@adityapgupta adityapgupta marked this pull request as ready for review June 19, 2025 17:28
@HacksonClark
Copy link
Collaborator

@adityapgupta

  1. Sure, if you want to be kind-specific you can. I feel if you need to be kind-specific though something is very wrong. Kind is just a k8s emulator, so if it works on kind but not on a real cluster something is incorrect.
  2. Any clients you want are fine. We don't force anything too specific in terms of clients, as those are meant to be implemented by an AIOpsLab user (in this case, you!).
  3. We use poetry, so you'll have to update pyproject.toml for the new dependencies.
  4. Only README that may be relevant is the main one.

@adityapgupta
Copy link
Author

@HacksonClark

  1. I dont mean kind specifically, I dont even use k8s here. Its just that I need to write an if condition everywhere to not launch/search for the k8s related stuff while using a docker application. But considering the main readme asks users to setup kind anyway, I thought I can skip this check for docker applications
  2. Ok Ill keep the Groq client in that case for diversity (and for people like me who don't have a paid OpenAI key)
  3. Ill update pyproject then
  4. Will make a mention in the main readme about docker and Flower

@HacksonClark
Copy link
Collaborator

Yeah, flower is kind of its own thing.

@adityapgupta
Copy link
Author

I have added a check for docker problems by including a new field in registry. However I think its better that some k8s cluster runs in the background to prevent the code from breaking at multiple points. Its anyway the first part of the setup guide, so I expect users would do it.
I have updated the requirements. On a second thought, I also realised that the readme is not problem specific, so I dont think we need to make any changes to that.

That wraps up everything I wanted to add. Please go through it once and let me know if there is something you want me to change.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

10 participants