Include 'instanceName' and 'schedulerBackendName' in CTA logs once multiple 'cta-taped' services are enabled
Introduction
As discussed in https://gitlab.cern.ch/cta/operations/-/issues/1254 and #576 (closed), we are support multiple cta-taped services running on the same host, each one for a separate drive.
Each one of them will have a different configuration file, which identifies it's instance name and scheduler backend name (amongst other configs).
Problem
If we implement this as described above, our monitoring systems and operators will have a difficult time differentiating between different services running on the same host.
Apart from the log file name, the contents of the logs will have no info that allows us to distinguish between various cta-taped services.
Solution
(copied from https://gitlab.cern.ch/cta/operations/-/issues/1254#note_7523136):
- all
cta-tapedconfiguration options must be listed:
- in the logs at boot time
- in
driveConfigjson keyinstanceNameandschedulerBackendNamemust be on every log line forcta-tapedandcta-frontend
- important for operators: to prevent having to look into multiple log files
- important for our new log forwarding system as
rsyslogneeds to understand the context
Edited by Joao Afonso