Increase scaler stream interval to 200ms

[tommy351]

Currently the scaler uses a lot of memory when there're a lot of HTTPScaledObjects (We have 300+). Based on the heap snapshots, it seems that the StreamIsActive handler consumes most of memory. I adjusted the interval from 5ms to 200ms to fix this issue.

Part of heap graph, you can open heap snapshot files below for details.

Heap snapshots:

• Before
• After

Memory usage:

Before:

After:

Checklist

• Commits are signed with Developer Certificate of Origin (DCO)
• Changelog has been updated and is aligned with our changelog requirements
• Any necessary documentation is added, such as:
  • README.md
  • The docs/ directory
  • The docs repo

[JorTurFer]

This is a nice improvement! The resources saving is awesome, but I'm afraid about if this can impact in current applications, I don't think so because 5 checks per second sounds enough, but IDK if it makes sense adding a parameter for setting this value from users side (with a default value).
WDYT? @t0rr3sp3dr0 @tomkerkhove

[t0rr3sp3dr0]

I like the idea of making this configurable and also think that 200ms is a good default value for it.

[JorTurFer]

@tommy351 , could you add a parameter for that value (and document it) ? 🙏

[JorTurFer]

Hi @tommy351
We will cut a release probably this month (we are coming back from the summer break and we are still a bit slow). Do you have any update about this?

[JorTurFer]

Hello @tommy351
Any update? We plan a release soon and this'd be a nice improvement

[tommy351]

Hi @JorTurFer, is the parameter you mentioned application-wide or just for a single ScaledObject?

[JorTurFer]

I think that we can start with a global parameter because the improvement will be huge. If someone has a use case where it's required at HTTPScaledObject, we can just iterate.
From changes pov, a simple parameter within scaler startup is quite easier than adding it via CRD

[JorTurFer]

I add the config value in another PR because we want to cut a release tomorrow