cloudspannerecosystem / autoscaler

Automatically scale the capacity of your Spanner instances based on their utilization.
Apache License 2.0
86 stars 33 forks source link

Best practice autoscaling threshold configuration? #343

Closed alexlo03 closed 2 months ago

alexlo03 commented 2 months ago

Hello,

Out of the box this project configures the high priority CPU threshold at 65% (regional), 45% (multi-regional) with a margin of 5%, which means that it only autoscales when it sees 70% (regional) or 50% (multi-regional)

Is this advised for Production use, or should we set the thresholds lower?

Refs

https://github.com/cloudspannerecosystem/autoscaler/blob/4248b1a1400baa06cd99b0d8ca25a37dc52e706f/src/poller/poller-core/index.js#L100-L101

https://cloud.google.com/spanner/docs/cpu-utilization#recommended-max

davidcueva commented 2 months ago

hi Alex,

these are the thresholds recommended by our Product Managers. They are a good starting point that can be used in production.

That being said, observe your resource utilization before and after, and adjust them if necessary.

David