MQE: Panic when memory isn't managed as we expect #9121
Conversation
For use in tests
We would get a new slice when we already had an existing slice we were keeping around for use.
| // Increase the memory tracker for 2 FPoints, and 1 HPoint | ||
| memoryConsumptionTracker.IncreaseMemoryConsumption(types.FPointSize*2 + types.HPointSize*1) |
There was a problem hiding this comment.
Why is this necessary? Can't we just assert that CurrentEstimatedMemoryConsumptionBytes is 0 below?
(same comment applies to the other similar changes in this file)
There was a problem hiding this comment.
Because the functions we're testing actually call DecreaseMemoryConsumption at some point, which requires the current estimated memory to be larger than what is returned. I figured may as well put in real values if that path is being executed.
There was a problem hiding this comment.
Sorry, I'm missing something: where is DecreaseMemoryConsumption called in this test? floatTransformationFunc reuses the slice it's passed.
(I can see this is necessary for TestFloatTransformationDropHistogramsFunc though)
There was a problem hiding this comment.
oh, sorry, I see your point. I'm doing this for consistency with TestFloatTransformationDropHistogramsFunc
What this PR does
Which issue(s) this PR fixes or relates to
Related to https://github.com/grafana/mimir-squad/issues/2280
Checklist
CHANGELOG.mdupdated - the order of entries should be[CHANGE],[FEATURE],[ENHANCEMENT],[BUGFIX].about-versioning.mdupdated with experimental features.