Describe the bug
Query-frontend shows in logs query performance stats and bloom-filters related metrics ("index_bloom_filter_ratio" and "index_post_bloom_filter_chunks") are always equal to zero, even if bloom-gateway metrics show filtering.
Making query like {job="gcplog"} |= ip("10.16.0.0/16") I can see
Describe the bug Query-frontend shows in logs query performance stats and bloom-filters related metrics ("index_bloom_filter_ratio" and "index_post_bloom_filter_chunks") are always equal to zero, even if bloom-gateway metrics show filtering. Making query like
{job="gcplog"} |= ip("10.16.0.0/16")
I can seein bloow-gateway logs something like:
so here filter_ratio varies from 0 to 1
in index-gateway logs I can see
in Grafana for this request I can see filtering ratio
BUT in query-frontend stats I receive:
and here I always see
I guess it's a bug of query-frontend, because basing on logs of index-gateway and bloom-gateway configuration is fine.
To Reproduce Steps to reproduce the behavior:
{your_label="blabla"} |= "any_long_string_with_a_lot_of_characters"
Expected behavior Query-frontend should show in stats actual filter_ratio value and amount of chunks to process after filtering
Environment:
Screenshots, Promtail config, or terminal output attached in problem description