HAProxy active checks as Manager LB

[Jure Sah T2]

Hello,

I am using HAProxy as a TCP loadbalancer for my Wazuh manager cluster (for agent connections).

I have configured active health checks:
https://www.haproxy.com/documentation/aloha/latest/load-balancing/health-checks/tcp/

I have noticed that sometimes some of my worker nodes are being detected as unreachable. However the node appears to be working properly (wazuh-manager systemd service is active and there are no errors) and I can exclude any network issues as all nodes are on the same network with no firewalls between them.

What I'm thinking maybe happens is that maybe Wazuh starts declining TCP connections having detected them to be bogus or something. Do you think this is possible?

What is the recommended health check method for this type of setup?

LP,
Jure