Part of my job is analyzing and debugging existing scripts in my company.
One method is analyzing the logs creaded by all the
Log Message-activities in our script, including logged errors by errorhandlers (message + source)
Using REF, the logs stack up quite fast if multiple logs are created during execution, and we have multiple queueitems. Navigating though this can be quite tedious, unless you export it continiously to csv and process that, which is not preferable.
An improvement option could be:
Add log messageactivity gets an extra optional input argument, containing a string value for filter purposes.
- Orchestrator log overview would be enriched with an extra data column, containing this filter string.
- The search option in the orchestrator log overview can then be used to filter for these ‘filtered’ logs.
This could greatly speed up analysis, if we as developers update the log messages with the queueitem references, or ref-state names in case of init or end process. One can simply fetch all the relevant logs for a single queueitem instead of navigating through several pages.
An even beter bonus would be that from the queue overview, you could link through towards the specific logs, filtering based on this key, though I suspect that that requires additional database relations and constraints to be implemented. (But a guy can dream, right?)
Of course you can achieve the same by prefixing your log messages with the said values, but that makes leibility somewhat… annoying.
Love to hear your thoughts on this!