Average time to evaluate rule groups is higher when using streaming of chunks

[pstibrany]

When using streaming of chunks as implemented in #3889, average time to evaluate rule groups goes up.

[pracucci]

Is it because of the chunks decoding done in the ruler (instead of spreading it across all ingesters)? Do you have any ruler trace to investigate it further?

[pstibrany]

Is it because of the chunks decoding done in the ruler (instead of spreading it across all ingesters)?

I assume that is the case, but haven't investigated it further yet. What is worrying is that rule evaluation has doubled in average, but it could also be specific to rules.

[stale]

This issue has been automatically marked as stale because it has not had any activity in the past 60 days. It will be closed in 15 days if no further activity occurs. Thank you for your contributions.

[pstibrany]

Still valid

[bboreham]

Was this fixed by #4341 ?

[bboreham]

Answering my own question: no it was not fixed. That part improves, but mergeIterator seems to be so much slower that it defeats the improvement.

[stale]

This issue has been automatically marked as stale because it has not had any activity in the past 60 days. It will be closed in 15 days if no further activity occurs. Thank you for your contributions.