UCCX HA issue Finesse

[Karen Caulfield]

Hi,

We’re running UCCX 12.5 SU2 integrated with CUCM 12.5,
and we're experiencing an issue with our high-availability (HA) deployment.

The problem is that when we do a manual failover (shutting down the PUB node),
agents connected to Finesse on the PUB are not redirected properly to the SUB node.
Some agents get disconnected, some experience partial Finesse UI load,
and CTI seems flaky during the transition.

Is there a known issue with HA failover handling in Finesse,
and how can we make this transition smoother for the agents during a failover?

[Mohammadreza Hadi · Accepted solution]

There is some steps you can check (to see if your problem will fix or not ?)

Ensure that DNS A record TTL is set very low (e.g., 30 seconds or less)
for the FQDN resolving to both nodes.
This allows quicker re-resolution during failover.
on the reverse proxy, implement health checks that automatically detect PUB failure and reroute traffic to the SUB without relying solely on DNS..


make sure your agents are logging in via the HA FQDN,
and that both UCCX nodes are registered as A records under that FQDN..

Additionally, there’s a known caveat that Finesse client doesn’t truly "failover" mid-session—rather,
it requires the agent’s browser to reinitiate the session against the new active node,
So instructing agents to refresh their browser..

[Karen Caulfield]

thanks @Mohammadreza Hadi 

with your guidance, we have now better user experience on our finesse agents