reduce severity for 'pebble: compaction cancelled by a concurrent operation, will retry compaction'

[tbg]

Have been seeing these regularly recently:

teamcity-16884024-1726292891-120-n7cpu4-0005> E240914 12:20:44.318336 17340 3@pebble/event.go:764 ⋮ [T1,n5,s5,pebble] 255 background error: pebble: compaction cancelled by a concurrent operation, will retry compaction

Apparently they are expected¹, so they should likely be logged as informational, if at all.

Jira issue: PEBBLE-257

Footnotes

1. https://cockroachlabs.slack.com/archives/CAC6K3SLU/p1726482169839989 ↩

[RaduBerinde]

We should instead expose some metrics about how much work was wasted (eg bytes written in canceled compactions).