Zabbix proxy: Utilization of housekeeper processes over 100%

[XxSoulDarkxX]

I have a Zabbix Server and two Zabbix Proxies running version 5.4 where the housekeeper processes have been at 100% utilization for several months.



What's strange is that I have other proxies with the same version, the same number of hosts, and the same templates, but they don’t have this issue.

The two problematic proxies handle 200 and 80 hosts respectively, and the history for item data doesn’t exceed 7 days. However, in the proxy logs, I constantly see entries like the following:

2549372:20250731:134257.288 executing housekeeper
2549372:20250731:134355.679 housekeeper [deleted 780071 records in 58.380930 sec, idle for 1 hour(s)]
2549372:20250731:144356.364 executing housekeeper
2549372:20250731:144712.654 housekeeper [deleted 833834 records in 196.282909 sec, idle for 1 hour(s)]
2549372:20250731:154713.386 executing housekeeper
2549372:20250731:154923.588 housekeeper [deleted 856421 records in 130.192289 sec, idle for 1 hour(s)]
3675268:20250731:160921.423 proxy #88 started [housekeeper #1]
3675268:20250731:163921.800 executing housekeeper
3675268:20250731:163947.481 housekeeper [deleted 682788 records in 25.677738 sec, idle for 1 hour(s)]
3675268:20250731:174002.307 executing housekeeper
3675268:20250731:174038.625 housekeeper [deleted 854497 records in 36.314982 sec, idle for 1 hour(s)]
3690968:20250731:175228.398 proxy #88 started [housekeeper #1]
3690968:20250731:182228.739 executing housekeeper
3690968:20250731:182244.557 housekeeper [deleted 528018 records in 15.813154 sec, idle for 1 hour(s)]
3698839:20250731:183354.323 proxy #105 started [housekeeper #1]
3698839:20250731:190354.618 executing housekeeper
3698839:20250731:190407.677 housekeeper [deleted 469288 records in 13.054415 sec, idle for 1 hour(s)]

The configuration of one of these proxies is:

Server=10.73.14.76
Hostname=m-10.251.0.135-10050-standalone
ListenPort=10050
LogFile=/var/log/zabbix/zabbix_proxy.log
LogFileSize=0
EnableRemoteCommands=1
LogRemoteCommands=1
PidFile=/var/run/zabbix/zabbix_proxy.pid
SocketDir=/var/run/zabbix
DBName=zabbix_proxy
xxxxxx=xxxxxxx
xxxxxx=xxxxxxxx
StartPollers=250
StartPreprocessors=50
StartPollersUnreachable=130
StartTrappers=50
StartPingers=50
StartHTTPPollers=5
StartVMwareCollectors=15
VMwareFrequency=60
VMwarePerfFrequency=60
VMwareCacheSize=128M
VMwareTimeout=120
SNMPTrapperFile=/var/log/snmptrap/snmptrap.log
CacheSize=5G
StartDBSyncers=5
HistoryCacheSize=512M
HistoryIndexCacheSize=512M
Timeout=30
LogSlowQueries=3000
AllowRoot=1
StatsAllowedIP=127.0.0.1,10.73.14.76

The proxy databases use MySQL, and log bin is disabled in /etc/my.cnf to avoid unnecessary disk usage:

[mysqld]
disable_log_bin

I need help since the behavior is very unusual.

[LenR]

Proxy databases are basically disposable. I've just stopped the proxy, dropped the database, reloaded the proxy schema, then restarted the proxy in the past, at least for space issues.

I wonder if the proxy data is somehow incorrect. It may be a false report.

5.4 is pretty old and out of support too.