Zabbix server crashes after adding SNMP devices

[Deviation]

I'm using the 2.4.6 appliance which I realize isn't meant to be in production. I used the VMware ovf template to deploy.

As part of my testing, I've added ~80 Windows hosts using the Zabbix agent. Everything has been running smooth. Recently I started added some Cisco switches. This when things started going wrong. After a period of time, the zabbix-server process crashes. I have about 10 SNMP devices currently configured using the SNMP Interfaces template.

I did make some adjustments to the default values. Below are the changes I made:
/etc/zabbix/zabbix-server.conf
StartPollers=50
CacheSize=1G

/etc/my.cnf
max_connections=256

Set the VM template to 8G RAM. Also configured a 4G swap volume.

Here's an excerpt from the zabbix-server.log:
12700:20160323:161333.033 [Z3005] query failed: [2006] MySQL server has gone away [select distinct i.itemid from items i,hosts h,hosts_groups hg,groups g where i.hostid=h.hostid and h.hostid=hg.hostid and hg.groupid=g.groupid and i.key_='perf_counter["\\KEY"]' and i.status=0 and i.state=0 and h.status=0 and g.name in ('SERVERS')]
12700:20160323:161333.074 [Z3001] connection to database 'zabbix' failed: [1135] Can't create a new thread (errno 11); if you are not out of available memory, you can consult the manual for a possible OS-dependent bug
12700:20160323:161333.075 Cannot connect to the database. Exiting...
12652:20160323:161333.269 One child process died (PID:12700,exitcode/signal:1). Exiting ...

MySQL stays up and running. Any ideas on where to start? I'm tempted to see if the issue is still present in the 3.x release.

[Joloch!]

I have the same problem, but I just added a single cisco switch and my Zabbix server goes down unless I delete that host. I don't know if that's caused by the number of triggers that are added to the host (almost 900 triggers). It just happens with Cisco devices.

I tried to use different templates but they doesn't several items and I can't get data for the graphs because of that.

I'm using Zabbix 3 running on a VM with 512MB of RAM.

[dsea]

You'll need to increase your connections by a lot more probably but it looks like this is a memory error. What is your current memory utilization at? For connections run: show status like '%conn%'; and see if you're still hitting your limit in the "max_used_connections" field. This is all still present in 3.0.1 I just upgraded to see if it would resolve a similar issue but I ended up having to increase the polling timeout and increase the max_connections to 5000 (not using nearly that though)