I was seeing agent execution getting stuck but without time-out occurring. I.e. the poller was connecting to the port but not streaming data which should result in timeout.
This adds a stream time-out to fix that.
Add timeout on unix agent execution after connect.
Auto-Deploy finished, Test PR at http://1319.ci.librenms.org or https://1319.ci.librenms.org
One question that's whether to continue processing the data if the agent had timed out during fetch. Means we have partial data. At the moment it is processing whatever data has been fetched. So if it's usable it will be used.
I'd say grab what data we can.
Prob worth setting this to be a user controlled variable so people can increase / decrease the timeout.
The existing connection time-out is 10s hard coded. Just stuck with that, but can change if you want.
Add config settings for unix-agent time-outs.
OK, I've added config settings for both time-outs.