Does anyone know which one is the best load-balancing method for two identical icinga-master nodes in HA?
not sure if all API requests should be forwarded to a single master, or if it is ok to have a round robin distribution (or a least loaded, shall the balancer permit).
Master actively connect to satellites, so only connections from ApiUser accounts are expected for passive checks, icinga-web interaction, and Integrations with other monitoring systems.
Thanks @stevie-sy , I will forward all api requests to one of the masters, yet it would be nice to know if a round robin algorithm would work equally well .
Best
you’re welcome. I don’t know if keepalived is able to do this. My knowledge is enough first that the failover works. I haven’t dealt with anything else yet. Others from the community may be able to provide better information