Last modified: 2014-10-29 11:24:29 UTC
The parsoid service on the beta cluster runs on deployment-parsoid04.eqiad.wmflabs. I have made it to log to /data/project/parsoid/parsoid.log, I guess that is simply the standard/error outputs being redirected there. It would be very helpful to prefix each message with a human readable timestamp. For now we get: master shutting down, killing workers Worker 32039 shutting down Exiting master master(32050) initializing 5 workers - worker(32060) loading... .. The timestamp should probably only be added when run in daemon mode, i.e. when the standard/error outputs are not ttys.
Any idea to get timestamps in log files ? :-]
All events sent to logstash now have timestamps. But, default logging (in development) doesn't. Is this useful for dev?
(In reply to ssastry from comment #2) > All events sent to logstash now have timestamps. But, default logging (in > development) doesn't. Is this useful for dev? The purpose was to figure out when a Parsoid daemon related event occurred on beta. Since they are now sent to logstash (bug 57711) that fulfills my original need :-) Thank you!