I’m experiencing a rather slow decrease of performance. I feel I have to restart openhab every ~3 months. Is there an easy way to get my hands on some indicators that I e.g could plot out using rrd or Influx?
Ideally it should covery openhab incl. bindings, rules DSL and jython.
So far I only found topics on specific performance questions - but nothing for measuring in general…
New Add-on bundle for Prometheus health Metrics. But really, taking a quick snapshot of what’s going on using htop before restarting it will probably be sufficient. Is the CPU pegged? Is it using a lot more RAM than when it first starts up? Those are easily answered without needing a three-months collection of metrics. From there you can start to focus in on certain specific areas.
Regarding a quick look - I have not saved anything or investigated in detail. But I can say that CPU load did not change significantly from before to after the restart. Memory usage might have decreased by ~1GB in the course of the restart, but I only see a host machine wide aggregated number…
The Prometheus thread feels a bit deserted. Last feedback was that it was not running with OH 2.5.2 anymore. The github projects also shows almost no activity in the last 2 years. Are you using it - you were interested.
As a quick hack, I’m storing a
ps -L -o comm,start,stat,bsdtime,pcpu,maj_flt,min_flt,lwp -p 578
top -n 1
away now and hope that I will remember to grab another one before the next restart (hmmm - maybe I put it into a cron job).