I'm running an older version (17.10) with around 1K users and three servers (all virtual running on VMWare). I've always seen ALARM_MONGODB_SLOW_READ warning at least a few times a day, for years, but they haven't caused an issue. Lately I've been seeing them more frequently and with bigger delays, and once the slow read seemed to cause sipXproxy to stop on my primary server (and I had to go in and restart it).
It doesn't seem like it's any one particular node complaining about slow reads either, most commonly it's imdb.entity, sometimes node.registrar, sometimes node.subscription.
Just wondering if anyone has any troubleshooting tips?
I'm wondering whether it's disk latency on the VMware side, something with my MongoDB, lack of memory, network latency between servers, etc. I do have 16GB RAM installed so I'd think that would be enough, but I could always add more.