-
Type:
Problem report
-
Resolution: Unresolved
-
Priority:
Major
-
None
-
Affects Version/s: 7.0.32, 7.4.16, 8.0.0
-
Component/s: Server (S)
Environment
- Zabbix server 7.0.32, PostgreSQL 16
Steps to reproduce
- Create host "Backup server 01" with trapper items "Backup duration" ("backup.duration") and "Restore test duration" ("restore.duration"), type "Numeric (float)".
- Create triggers from block 1 for both items.
- At 20:01, send "150" to both items with API "history.push".
- At 21:02, send "150" again, only to "Restore test duration".
- At 21:20, open Monitoring > Problems.
1. Trigger expressions for "backup.duration"
trendavg(/Backup server 01/backup.duration,1h:now/h)>100 trendmax(/Backup server 01/backup.duration,1h:now/h)>100 trendmin(/Backup server 01/backup.duration,1h:now/h)>100 trendsum(/Backup server 01/backup.duration,1h:now/h)>100 trendcount(/Backup server 01/backup.duration,1h:now/h)>0
Result
| Check | Result |
|---|---|
| "Backup duration" at 21:10 | no problem; avg, max, min: unknown, "not enough data"; sum, count: return 0, trigger OK |
| "Backup duration" trends row for 20:00 | observed at 21:55, still no problem |
| "Restore test duration" | 5 problems at 21:02 |
| Same with "1d:now/d", value at 23:40 | no problem; evaluated at 00:10, row for 23:00 written at 00:55 |
Expected
Finished period is evaluated with all values received for it, without a later value: problem for 20:00-21:00 is created at its scheduled evaluation, 21:10-21:20 (documentation: evaluated once per period).