If it wasn’t for graphs1090 I wouldn’t have known that I had a break of about an hour in my recent collected data:
Looking at the log I can see that the system decided after 3707 seconds that something was amiss and kindy restarted dump1090
All was well subsequently
Now I’m well aware that I can configure the “Outage Alerts Threshold” to give an email notification after (say) 6 hour outage, however what I would like to be able to do is get the automatic dump1090 restart to occur much sooner when no messages are received. Waiting over an hour seems a little excessive in this part of the world.
Is there a paramater that I can tweak and if so, where would I find it?
Any advice much appreciated!
This really is a minor issue and rarely occurs so it’s once in a blue moon (or in this case, perhaps a super moon, purely coincidentally)
Hi @foxhunter thanks for the suggestion about using monit.
I was really not contemplating adding additional monitoring tasks as the code already recognises when no messages have been received.
So there is already a suitable “watchdog” task baked into the code and it is very likely that the period it uses is configured somewhere.
Perhaps a thread wakes up every now and then and if nothing counted for > 3600 secs then assume the worst and sensibly restart dump1090.
The existing functionality is fine, I just want to do a fine tune to match the traffic in my location.
Personally I’d be happier with 300 rather than 3600, even allowing for the nighttime traffic. It’s extremely rare to see zero messages here in uk.
OK, i did not know that there is already something in to restart the service only.
I am using Monit since months for my devices and it worked, that’s why i did not think about a different solution.
The only device where it’s not installed is my Jetvision AirSquitter because they put this kind of monitoring into their tools, it restarts if after a few minutes if there is an unusal drop of messages.
Seeing zero messages for more than a few minutes is also very unlikely here mid of germany. Had it only a few times during the low traffic last year.
Yes, I agree, I myself would like to see an option of 1 or 2 hours, usually a reboot takes care of my issues. But waiting 12 hours seems way too long. Maybe the software should be updated to simple tasks like that as to help many people in the same situation.
