We recently upgraded a single-machine Enterprise deployment from 10.9.1 to 11.2 (not 11.3, for reasons not relevant here). Prior to the upgrade, the Server instance logging level was set to WARNING. Following the upgrade, the setting still says WARNING (and we changed it to INFO, saved, changed back to WARNING, and re-saved). However, the actual logging is being done at the VERBOSE level, which as the documentation mentions is causing a major performance hit on a busy production server.
I looked for environment variables that might be overriding the logging setting, and did not find anything. I am at a loss for how to get it to actually follow the logging level it is configured with. This is an issue not only because of the performance issues but also we have clients using certain services (e.g. our geocoder) who don't want us logging their queries and we want to honor that request.
Solved! Go to Solution.
A reboot of the system fixed the logging and charting issues. Memory usage is still pretty high, but the system is stable.
Another possibly related issue is that in the statistics reports/charts, it doesn't show any activity on the geocoding service. However, the logs show the service very actively used (averaging around 3k/hour). It may just be a side effect of the out-of-memory condition caused by the stuck logging.
A reboot of the system fixed the logging and charting issues. Memory usage is still pretty high, but the system is stable.