Skip to content

Iterate on WAL rate db health check indicator

Extracted from !128365 (comment 1505253631)

Why this issue?

!128365 (merged) added WAL rate db health check indicator with 70MB as the threshold. As mentioned by @rhenchen.gitlab in this comment, setting threshold for WAL rate wouldn't be the desired option.

This issue aims to iterate on the WAL rate indicator metrics to make it better.

Proposed Solutions:

  • Using standard deviation to detect abnormal spikes.
  • Consider replication lag along with WAL rate to get a better picture.