Bug Report
Which version of the demo you are using? 1.10.0
Symptom
Enabling recommendationServiceCacheFailure is a noop
What is the expected behavior?
Enabling recommendationServiceCacheFailure results in cache misses and latency
What is the actual behavior?
Enabling recommendationServiceCacheFailure is a noop
Reproduce
Edit src/flagd/demo.flagd.json and change the defaultVariant of recommendationServiceCacheFailure to on. Look for cache misses or latency in Jaeger
Additional Context
This appears to be a regression from at least version 1.7.0. As part of the featureflag to flagd change, it appears the recommendationCache featureflag name was changed to recommendationServiceCacheFailure but src/recommendationservice/recommendation_server.py was not updated accordingly. Either the recommendation service needs to be renamed or the flagd parameter needs to be changed back to what it was previously set to.
I'm happy to make this change depending on which way the community wants to go.
Also, likely good to audit whether other feature flags were impacted.
Bug Report
Which version of the demo you are using? 1.10.0
Symptom
Enabling recommendationServiceCacheFailure is a noop
What is the expected behavior?
Enabling recommendationServiceCacheFailure results in cache misses and latency
What is the actual behavior?
Enabling recommendationServiceCacheFailure is a noop
Reproduce
Edit
src/flagd/demo.flagd.jsonand change thedefaultVariantofrecommendationServiceCacheFailuretoon. Look for cache misses or latency in JaegerAdditional Context
This appears to be a regression from at least version 1.7.0. As part of the featureflag to flagd change, it appears the
recommendationCachefeatureflag name was changed torecommendationServiceCacheFailurebutsrc/recommendationservice/recommendation_server.pywas not updated accordingly. Either the recommendation service needs to be renamed or the flagd parameter needs to be changed back to what it was previously set to.I'm happy to make this change depending on which way the community wants to go.
Also, likely good to audit whether other feature flags were impacted.