This evening my 3B running PiAware 3.8.0 has suddenly started displaying unusual Overall CPU Utilization in graphs1090, and it’s a bit laggy when SSH’d in. It appears to be a real effect that’s being measured, not a graph artifact.
The three gaps, in order, are:
- me dropping the gain a couple of notches and rebooting
- me rebooting again after noticing this behaviour
- me setting the gain back to what it was and rebooting again in case that was relevant (it wasn’t)
I use full-size graphs which makes it easier to align events vertically across all the graphs. This behaviour started at 19:10 UTC. In this case no other graph shows anything unusual. There’s nothing of note in
dmesg at that or any other time.
top reveal nothing unexpected. Mono and tmux weren’t running today so they’re not related.
Anyone else seen similar behaviour?