Zabbix, IPv6, DNS, and huge session spike?
Previously, we were using Zabbix 4.x. I just updated to 5.x today and still have the same problem as below.
Our hosts are setup in Zabbix (running Zabbix in Docker) using their FQDN for DNS instead of IP. Everything was working perfectly fine until....
We implemented IPv6 connectivity and all of a sudden, our session spikes shot through the roof. I can check that a DNS query resolves to hosts the correct IPv6 addresses. I've enabled IPv6 in Docker. Setting the agents up using their IPv4 addresses instead of DNS eliminates the session spikes. Trying to manually use the IPv6 address gives the error that Zabbix can't connect to the IP.
So I'm leaning toward something is messed up with how Zabbix is dealing with IPv6 clients. Using DNS causes a session spike since DNS is first returning the IPv6 addresses.
Is this a Zabbix issue or a Docker issue?
Our hosts are setup in Zabbix (running Zabbix in Docker) using their FQDN for DNS instead of IP. Everything was working perfectly fine until....
We implemented IPv6 connectivity and all of a sudden, our session spikes shot through the roof. I can check that a DNS query resolves to hosts the correct IPv6 addresses. I've enabled IPv6 in Docker. Setting the agents up using their IPv4 addresses instead of DNS eliminates the session spikes. Trying to manually use the IPv6 address gives the error that Zabbix can't connect to the IP.
So I'm leaning toward something is messed up with how Zabbix is dealing with IPv6 clients. Using DNS causes a session spike since DNS is first returning the IPv6 addresses.
Is this a Zabbix issue or a Docker issue?
Comment