cmsdaq / DAQExpert

New expert system processing data model produced by DAQAggregator
1 stars 2 forks source link

logic module to catch unidentified problems #59

Closed andreh12 closed 7 years ago

andreh12 commented 7 years ago

discussing with @gladky we propose the introduction of a logic module which fires when

This will allow to find such cases in the notification manager archive more easily and help to improve the software to identify not yet caught cases in the future.

More specifically, this new logic module will check whether NoRateWhenExpected has fired but no logic module inheriting from KnownFailure has identified a problem.

gladky commented 7 years ago

We have the test case of unidentified-problem which was caught on development instance:

http://daq-expert-dev.cms/nm/archive?source%5B%5D=UnidentifiedFailure&page=1

I suggest we change the suggested messages from

Reason for no trigger rate could not be identified

  1. make an elog entry

to

Trigger rate is zero. The problem could not be identified.

  1. recover the system according to instructions from shifter bulletin board
  2. make an elog entry
gladky commented 7 years ago

Introduced with 2.4.0