Zabbix simple check queue after adding proxy

[Gutts]

Hi all!
I'm using Zabbix 2.2.3. I had 1 zabbix-server and there were no problems. But our infrastructure is growing and there was a need to configure one zabbix proxy to monitor a certain range of production servers. At the moment about 40 servers are being monitored through our new zabbix-proxy. Everything works fine but there is a queue for simple checks.
I attach a couple of screenshots.

Zabbix-proxy configuration:

LogFile=/var/log/zabbix/zabbix_proxy.log
LogFileSize=0
DebugLevel=3
PidFile=/var/run/zabbix/zabbix_proxy.pid
DBSocket=/var/lib/mysql/mysql.sock
ProxyOfflineBuffer=12
StartPollers=50
StartPollersUnreachable=3
StartPingers=6
StartDiscoverers=7
StartVMwareCollectors=7
CacheSize=16M
HistoryCacheSize=16M
Timeout=30

What I already did:
-Syncronized time on both server and proxy using ntpd
-Played with configuration
-Disabled all items with "Not supported" status

The logfile doesn't seem to be very informative despite DebugLevel 3 is turned on:

20924:20150407:133246.849 proxy #32 started [poller #29]
20928:20150407:133246.850 proxy #36 started [poller #33]
20941:20150407:133246.850 proxy #49 started [poller #46]
20903:20150407:133246.854 proxy #11 started [poller #8]
20930:20150407:133246.854 proxy #38 started [poller #35]
20967:20150407:133246.858 proxy #70 started [discoverer #1]
20910:20150407:133246.860 proxy #18 started [poller #15]
20965:20150407:133247.368 housekeeper [deleted 80529 records in 1.440065 sec, idle 3598 sec]
20965:20150407:143245.369 executing housekeeper
20965:20150407:143246.733 housekeeper [deleted 110173 records in 1.362548 sec, idle 3599 sec]
20891:20150407:143249.414 received configuration data from server, datalen 561263

I would be really grateful if you, guys, give me some advice regarding getting rid of this queue or tuning zabbix-proxy configuration

[ingus.vilnis]

Hello and welcome to Zabbix forums!

In Administration -> Queue select Details and Overview by proxy. These pages will give you more clues how to track down the delayed items. Maybe they even are not from the proxy.

Also check Zabbix cache usage graph (and other graphs) on the proxy. Not likely but maybe you are out of memory with your amount of pollers. Check how busy are the internal processes as well.

Best Regards,
Ingus

[Gutts]

Hi!
Thank you for your answer!

I have checked all the usage/internal processes graphs both on server and proxy and they seem to be fine.
In Administration>Queue I see the same information (the monitored servers are always different)

[ingus.vilnis]

Hi,

What comes into my mind is that this might happen because of complicated ping item settings.

Could you please show a full screenshot with all settings from some of your ping items?

Best Regards,
Ingus

[Gutts]

Hi!
They are similar. There were no problems before we started to use proxy. Here is an example

[ingus.vilnis]

Hi,

Nothing wrong there. Standard item from ICMP template... Ok then.
Try adding more pingers in server configuration file and restart the server. Maybe that helps.

StartPingers=10

Best Regards,
Ingus

[Gutts]

The usage of pingers was 40% but I added some more. Didn't help

[alasdair]

Resolved?

@Gutts.

Currently experiencing the same issue. Simple check icmpping (fping) works fine on hosts the Zabbix server has direct access to but checks going through the proxy just stay in the queue with nothing in the logs.

Proxy config has correct path:

FpingLocation=/usr/bin/fping

and binary has suid bit set:

-rwsr-xr-x 1 root root 31496 Jun 25  2012 /usr/bin/fping

If you managed to resolve this it would be great if you could post the steps you took for the resolution of this issue.

Zabbix version: 2.2.9