Werk #22515: Averaging checks: Fix ZeroDivisionError after time anomalies
| Component | Checks & agents | ||||
| Title | Averaging checks: Fix ZeroDivisionError after time anomalies | ||||
| Date | Sep 29, 2026 | ||||
| Level | Trivial Change | ||||
| Class | Bug Fix | ||||
| Compatibility | Compatible - no manual interaction needed | ||||
| Checkmk versions & editions |
|
Checks that average their values over time could crash with
ZeroDivisionError: float division by zero. This was seen with the
Proxmox VE CPU Utilization service when averaging is configured: it uses the
uptime of the VM as its clock, and if the reported uptime stepped back and then
returned to exactly the time at which averaging had started, the next check
crashed.
This case is now treated like other time anomalies, and the service keeps reporting the averaged value.