I've just recently started having problems with a Zabbix setup with a Zabbix proxy on a remote site (connected by VPN). It's been working until recently. I did change things (upgraded router firmware and added IPv6 support), but as far as I can tell, things are still working OK generally after the change, so that may be coincidental.
I get problems where it appears that the Zabbix proxy stops sending results to the server, so all the server 'no data in the past 30 minutes' alerts will go off at once. They'll generally stay that way for another 30 minutes or so, then recover. This typically happens a couple of times a day, so most of the time it's working fine.
While the problem is happening, I can SSH onto the proxy over the VPN, and telnet between the proxy and server on ports 10050 and 10051 (both ways), so it's not that the network or VPN is failing. I can use zabbix-get from the proxy to the monitored hosts, and that works OK.
I've got level 4 logging on on the proxy, but can't see anything which looks significant in the logs there.
If I go into the PostgreSQL database, I can see that the number of items in the proxy_history table seems to be growing, with old entries not being deleted immediately. I don't know if this is expected, or a symptom.
I'm not sure what the columns in the table mean - does 'state' indicate whether it has been sent to the server (the state is always '0' here).
If I look at the 'zabbix[queue]' history graph, it is down at around 30 most of the time, then jumps up to about 2000 when it's having the problem, then drops back down to 30 or so when it starts working again.
The proxy is in 'active' mode.
ProxyLocalBuffer = 0
ProxyOfflineBuffer = 1
HeartbeatFrequency = 60
DataSenderFrequency = 1
zabbix_proxy is version 2.2.2
zabbix_server is version 2.2.9
I get problems where it appears that the Zabbix proxy stops sending results to the server, so all the server 'no data in the past 30 minutes' alerts will go off at once. They'll generally stay that way for another 30 minutes or so, then recover. This typically happens a couple of times a day, so most of the time it's working fine.
While the problem is happening, I can SSH onto the proxy over the VPN, and telnet between the proxy and server on ports 10050 and 10051 (both ways), so it's not that the network or VPN is failing. I can use zabbix-get from the proxy to the monitored hosts, and that works OK.
I've got level 4 logging on on the proxy, but can't see anything which looks significant in the logs there.
If I go into the PostgreSQL database, I can see that the number of items in the proxy_history table seems to be growing, with old entries not being deleted immediately. I don't know if this is expected, or a symptom.
I'm not sure what the columns in the table mean - does 'state' indicate whether it has been sent to the server (the state is always '0' here).
If I look at the 'zabbix[queue]' history graph, it is down at around 30 most of the time, then jumps up to about 2000 when it's having the problem, then drops back down to 30 or so when it starts working again.
The proxy is in 'active' mode.
ProxyLocalBuffer = 0
ProxyOfflineBuffer = 1
HeartbeatFrequency = 60
DataSenderFrequency = 1
zabbix_proxy is version 2.2.2
zabbix_server is version 2.2.9