update performance tests in openmrs-core

Hi there!

noticed that the library JUnitBenchmarks: Code benchmarking in JUnit4 we use for our only performance test ConceptServicePT is deprecated in favor of JMH

Out of curiosity I ran the PT and it fails

@dkayiwa I see from history you were the last to change the test titled: Revert “Another attempt at fixing bamboo randomly failing tests”

Question: when is the PT run?

Goal 1: obviously I think we should replace the library that is EOL for some years already. Goal 2-x: might this be a good point to think about what it is that we want to measure? I can see that there was a current issue related to performance with concept [TRUNK-5071] Fix performance of get concepts by mapping - OpenMRS Issues

We could figure out some PT tests that are only run before releases or in a nightly build but that could mitigate this from happening again.

The PT is not run on schedule. I ran it manually for measuring performance when migrating to Lucene.

We should definitely drop it, if it doesn’t work and is EOLed. We need to come up with an approach to test api performance and/or webapp.

when was that?

I dont think it fails because of junit-benchmark but because of failure in obtaining a lock to delete all data from db after all tests are run.

So if this PT is not even run and fails would this be ok:

STEP 1 (free us from this EOL dependency): in ConceptServicePT

  • remove all references to junit-benchmark in the test
  • add @Ignore on class level so it does not get picked up in pom.xml’s:
  • remove junit-benchmark from dependencies

STEP 2 (establish performance tests):

  • gather requirements of performance tests/metrics we want to establish
  • find a library that can achieve that and put it in place
  • think about if/how we want to integrate that into the build/release process

A few years back…

I would remove ConceptServicePT entirely. I see no point in keeping it around.

Step 2 is much more interesting.

uuuh, that is a long time.

and I agree with both statements :slight_smile:

here is the issue https://issues.openmrs.org/browse/TRUNK-5126 if others also agree, I’d be happy to work on it

Am in full agreement. :slight_smile: