unreachable agent flapping

[pridetechdesign]

Hi, I occasionally get false alerts about agents being unavailable. This has increased in frequency when I moved my database to a different host from the zabbix server.

Are there steps I can take to reduce these false positives? I tried checking the logs but I dont see anything about agent failures.

UPDATE: I've updated the server from 2.4.1 to 2.4.5, so far have not seen any flapping yet. Hopefully this has solved the issue, and maybe some other issues I was experiencing.

[pridetechdesign]

agent flapping is still a problem.

I just experienced a false alert where every agent being monitored suddenly went "unavailable for 5 minutes", but there are no communication issues between the zabbix server or any of the agents or the database.

Server and agents are 2.4.5

[pridetechdesign]

As of yesterday, this problem has escalated to the point where it requires a server restart. All of the agents suddenly reported being unreachable, even tho there are no connection issues between agents and server, and they did not automatically recover from the state until I restarted the server.

However, following a server restart they all returned to green status. Server resources are not maxing out, and the log isn't displaying any errors..

[Palmertree]

Check your Zabbix Que under Administration to see if things are queuing up. This can be caused by different factors: not enough pollers, something taking to long and timeout in conf file set high or network connectivity. Zabbix agent ping loss is usually monitor by the "nodata" function and if you get are not getting data, it will look like your agents are not responding. If you are seeing more than one agent go offline at the same time more like poller resources or network connectivity issues.