first pass at disk-based scale-up and down

[radu-gheorghe]

One-line summary

Issue : #10

Description

This is a first draft of disk-based scaling. I've tested it and it works if you add this to the scaling spec:

    scaleUpDiskUsagePercentBoundary: 10
    scaleDownDiskUsagePercentBoundary: 5

I've added the scale-up logic in the scaleUpOrDown function, so that we enforce it. I've added the scale-down logic to the scalingHint function, so that it's only a hint (e.g. if we have high CPU, we won't want to scale down).

Types of Changes

• New feature (non-breaking change which adds functionality)
• Configuration change

This is yet to be done, but I'd like to get some feedback early:

• Documentation / non-code

Tasks

• Initial PR
• Iterate over feedback
• Add docs
• Add tests?

Review

• Check the code
• CHANGELOG

Deployment Notes

None that I'm aware of

[otrosien]

@radu-gheorghe thank you very much for your interest in our project, and the effort you have put into starting this PR. Here are my first 2cts (haven't looked deeply into the code yet): First, we are looking into rewriting and binding the autoscaling code into the HPA framework - unfortunately this is not reflected in any ticket yet; I'll take care of that next week. Therefore, any addition that adds complexity to the current manual scaling effort may be in conflict with that idea. On the other hand, if the code doesn't get too convoluted, it should still be a good addition for the meantime. I'll start looking into the PR next week.

[radu-gheorghe]

Thanks @otrosien! The HPA framework idea sounds nice for sure, thanks for sharing the plan.

This PR shouldn't be too complex - if you think it fits, please let me know what I should change.