Closed jefflembeck closed 10 months ago
Need to look at the traffic shaping rules (netem for linux specifically) and make sure that the outbound rsyslog traffic is tagged in such a way that it doesn't go through the tc filters (and count against the test bandwidth when running a test)
Looks like you can add a rule to match the rsyslog traffic (and probably add ssh while you're in there) to pass
and give the netem rule a lower priority. Example here.
Need to look at the traffic shaping rules (netem for linux specifically) and make sure that the outbound rsyslog traffic is tagged in such a way that it doesn't go through the tc filters (and count against the test bandwidth when running a test)
Looks like you can add a rule to match the rsyslog traffic (and probably add ssh while you're in there) to
pass
and give the netem rule a lower priority. Example here.
@pmeenan ok, had to go read about tc
for a bit for this one, but I think I've got one in for syslog (give a pass for all traffic that has a destination port of 514 was the intention there, that oughta do it).
You think I should add one for port 22 as well?
Edit: done
@pmeenan I decided to go with saas_test_id
or Test ID
if the former does not exist
Closing as approach to logging will be default local. https://github.com/catchpoint/WebPageTest.agent/pull/638 @jefflembeck
This change allows a user to opt into logging via syslog, in case they want the agent to open a socket out to a service that consumes logs.
WARNING - This likely does not work on MacOS, since they changed the behavior of the syslog daemon