BUG: For large databases from old instances, logs stop being ingested and cpu usage keeps constantly at 100%
We noticed on a production instance with a ~6GB logs.db database and about 8 months of data, that when the routine that removes old log entries (than three months), logs seem to stop being ingested and the CPU usage keeps constantly at 100%.
The proposed fix, as discussed with @njean42 is to have the cleaning routine being executed more often, deleting small batches, to avoid blocking the main database writer thread.
The issue was introduced on implementing #204 (closed).
Version number of controlcenter
Version: 1.9
Postfix version
Use postconf mail_version
to get postfix version
Version: 3.4.1
Golang version
Version: 1.17
Which Operating System and version are you using?
OS: Manjaro
Edited by Leandro Santiago