-
Problem report
-
Resolution: Unresolved
-
Major
-
None
-
None
-
None
We got a report of an interesting issue - the alert syncer process starts to display title "queuing alerts", but it seems to hang, as it is stuck calling zbx_ipc_socket_read():
3017528:20240108:163607.413 zbx_setproctitle() title:'alert syncer [queuing alerts]' 3017528:20240108:163607.413 In am_db_get_alerts() 3017528:20240108:163607.413 query [txnlev:1] [begin;] 3017528:20240108:163607.414 query [txnlev:1] [select a.alertid,a.mediatypeid,a.sendto,a.subject,a.message,a.status,a.retries,e.source,e.object,e.objectid,a.parameters,a.eventid,a.p_eventid from alerts a left join events e on a.eventid=e.eventid where alerttype=0 and a.status=3 order by a.alertid] 3017528:20240108:163607.418 query [txnlev:1] [commit;] 3017528:20240108:163607.419 End of am_db_get_alerts():SUCCEED alerts:0 3017528:20240108:163607.419 In am_db_flush_results() 3017528:20240108:163607.419 In zbx_ipc_socket_write() 3017528:20240108:163607.419 End of zbx_ipc_socket_write():SUCCEED 3017528:20240108:163607.419 In zbx_ipc_socket_read() ... 3017528:20240108:163658.870 End of zbx_ipc_socket_read():SUCCEED 3017528:20240108:163658.871 End of am_db_flush_results() flushed:0 3017528:20240108:163658.871 zbx_setproctitle() title:'alert syncer [queued 0 alerts(s), flushed 0 result(s) in 51.458100 sec, idle 0 sec]'
This in turn creates high utilization for alert syncer and alert manager processes, and also creates a huge load on the installation as a whole - the UI becomes slow, and there also was a case where users weren't even able to SSH into the Zabbix Server to collect debug data.
This issue seems to happen completely at random, at one point we had closed the case as it hadn't re-appeared for a month, but then it showed up again and we are not sure why.
Please let us know what information you need to investigate this.
Zabbix Server version - 6.0.23