canonical / canonical/postgresql-operator
logging scheme makes no sense
- Dominant language
- Python
- Stars
- 20
- Forks
- 36
- Avg merge
- 4d 15h
- Merged PRs (30d)
- 30
Description
## Steps to reproduce
1. Deploy postgresql VM charm from `14/stable`
## Expected behavior
Log files are created and consumed in the standard way. You have your active log file as `postgresql.log` and rotated/archived log files as `postgresql.log.1.gz`, `postgresql.log.2.gz`, etc... with possibly configurable retention/rotation.
## Actual behavior
Log files are generated once per minute which makes searching through it extremely difficult and cumbersome. Furthermore, it does not actually work or make any sense. I see in an environment right now,
- `postgresql-5_0245.log` is today, Feb 21,2025 0245 am UTC
- `postgresql-5_0247.log` which you expect to be 2 minutes later is from Feb 14, 2025, 0247 am UTC
This is extremely strange and completely unnecessary IMHO. The standard way of creating and rotating log files that almost every service out there uses works really good and I don't see why the wheel has to be re-invented here.
Contributor guide
Assessment
This issue has not been assessed yet.