|
Develop JMeter benchmark tests for the edge-api. These will measure performance from the standpoint of the end-user (harvester).
- Happy path probably good enough for now.
- Checked into folio-org/folio-perf-test
- Can expect pre-existing data already exists
- Can expect that throttling has been disabled in the target edge api. These tests may help inform a decision on what the throttling configuration values should be.
- for the non-List* verbs, tests that simply repeatedly hit that verb are fine.
- for the List* verbs, having each thread perform a harvest makes sense... e.g.
- make an initial call with from/until
- make additional calls w/ the resumption token returned until completion
- Running these tests w/ analysis of the results is outside the scope of this story. Running the tests to ensure they're functioning as expected is clearly necessary.
|