Hi, In the documentation under the HA section, it mentions: "It's important not to load balance traffic between Prometheus and its Alertmanagers, but instead, point Prometheus to a list of all Alertmanagers."
I'm curious if this is strictly for high availability and network partitioning concerns, or if there is a more functional reason that every Alertmanager member needs to receive the alerts. What prompted this question, is that in our three member HA alertmanager cluster that we've been sending alerts to via a load balancer (from multiple prometheus instances), we've observed that alerts stored on each cluster member may have pretty drastically different endsAt times for a single given alert (one to two minutes). We believe that this may be contributing to random flapping alerts, that prometheus indicates has been firing the entire duration. Thanks. -- You received this message because you are subscribed to the Google Groups "Prometheus Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/4b4e8a88-1243-4591-b2d7-43bd550738een%40googlegroups.com.
