Sometime the issue is all inside the monitoring tool
retained text
Hi all I wanted to share a strange issue that bothered me for a while, just in case anyone else finds it useful (or funny). My lab has a couple of Proxmox hosts running several VMs, one of which is a TrueNAS VM with a dedicated SAS card. I also have a Raspberry Pi 3 running Uptime Kuma and Mailrise to monitor all my services. A few weeks ago, I started receiving alerts that my TrueNAS VM did not respond: "timeout of 30000ms exceeded". These were always immediately followed by a "recovery" alert. No other VMs showed any issues. I started digging into TrueNAS: checking connectivity cables, resource usage, and even increasing the timeout period in Uptime Kuma to 45 seconds. But the issue continued; I was getting 6-10 false alerts every day. The weird part? No other service was affected. There were no errors in the logs of the VMs using that NAS, and no errors within TrueNAS itself. I even checked the Uptime Kuma access events on the TrueNas access logs, and it didn't miss a beat there was an access log entry every 60 seconds, even for the timestamps that Uptime Kuma reported as "failed." I checked the switch for port flapping, but saw nothing (and besides, why would it only affect TrueNAS and not the other services on the same host?). Finally, I decided to simply restart the Uptime Kuma container. Just a basic restart, no version changes, no config changes. And since then I haven't had a single alert since. I’m still not sure what the cause was, and since I've since updated to the latest version, I didn't get a chance to see if it was a specific bug in that release. The moral of the story: sometimes "tapping the gauges" actually works. Bye K.
document fetched 2026-11-02 00:32 · canonical https://reddit.com/r/homelab/comments/1xkzx6p/sometime_the_issue_is_all_inside_the_monitoring