Persistent Hourly Authentication and CPU problems when ET_PROBLEMS are disabled
Closed this issue · 2 comments
loadgen:
container_name: loadgen
image: "dynatrace/easytravel-loadgen"
links:
- backend
- frontend
- www
environment:
ET_APM_SERVER_DEFAULT: "${ET_APM_SERVER_DEFAULT}"
ET_FRONTEND_URL: "http://easytravel-www:80"
ET_BACKEND_URL: "http://easytravel-www:8080"
ET_PROBLEMS: ""
ET_PROBLEMS: ""
Every hour there are authentication problems and CPU problems.
Based on this yml config no problems should be present in our easyTravel deployment.
This is conflicting with problems we are manually simulating and remediating
I am not sure, but maybe the "use default if empty" line from https://github.com/Dynatrace/easyTravel-Docker/blob/master/build/loadgen/build/scripts/run-problem-patterns.sh causes this:
ET_PROBLEMS=${ET_PROBLEMS:-'BadCacheSynchronization,CPULoad,DatabaseCleanup,DatabaseSlowdown,FetchSizeTooSmall,JourneySearchError404,JourneySearchError500,LoginProblems,MobileErrors,TravellersOptionBox'}
the ":-" syntax means "use the content after if the var before is empty", so you might trigger this default here.
Can you try with at least one problem pattern which does not affect you in the list?
That default if empty seems like an odd design choice.
I have now set a problem pattern to use:
ET_PROBLEMS: "DatabaseCleanup"
So far so good, the CPU and Auth problems have not been raised this morning.
Thanks for your help Dominik.