HTTP 500 errors should be retried

[symphony-elias]

Bug Report

During maintenance windows, when pod, km or agent is down, HTTP calls can result in HTTP 500. Such calls need to be retried. For instance, on one of our bots, we got:

2021-04-20 18:26:26.546  INFO 1 --- [_DatafeedThread] c.s.b.c.s.datafeed.impl.DatafeedLoopV1   : Recreate a new datafeed and try again
2021-04-20 18:26:26.684  INFO 1 --- [_DatafeedThread] .s.b.c.r.r.Resilience4jRetryWithRecovery : Retry in 64.0s...
2021-04-20 22:01:45.235  INFO 1 --- [_DatafeedThread] c.s.b.c.s.datafeed.impl.DatafeedLoopV1   : Recreate a new datafeed and try again
2021-04-20 22:01:45.389 ERROR 1 --- [_DatafeedThread] c.s.b.s.s.DatafeedAsyncLauncherService   : An API error has been received while starting the Datafeed loop in a separate thread, please check error below:

com.symphony.bdk.http.api.ApiException: {"code":500,"message":"Received an error when calling a pod endpoint"}
        at com.symphony.bdk.http.jersey2.ApiClientJersey2.invokeAPI(ApiClientJersey2.java:168) ~[symphony-bdk-http-jersey2-2.1.3.jar:2.1.3]
        at com.symphony.bdk.gen.api.DatafeedApi.v4DatafeedCreatePostWithHttpInfo(DatafeedApi.java:479) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.gen.api.DatafeedApi.v4DatafeedCreatePost(DatafeedApi.java:414) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.core.service.datafeed.impl.DatafeedLoopV1.createDatafeedAndPersist(DatafeedLoopV1.java:144) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.core.retry.RetryWithRecovery.executeOnce(RetryWithRecovery.java:105) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at io.github.resilience4j.retry.Retry.lambda$decorateCheckedSupplier$3f69f149$1(Retry.java:137) ~[resilience4j-retry-1.6.1.jar:1.6.1]
        at io.github.resilience4j.retry.Retry.executeCheckedSupplier(Retry.java:419) ~[resilience4j-retry-1.6.1.jar:1.6.1]
        at com.symphony.bdk.core.retry.resilience4j.Resilience4jRetryWithRecovery.execute(Resilience4jRetryWithRecovery.java:65) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.core.service.datafeed.impl.DatafeedLoopV1.createDatafeed(DatafeedLoopV1.java:139) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.core.service.datafeed.impl.DatafeedLoopV1.recreateDatafeed(DatafeedLoopV1.java:127) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.core.retry.RecoveryStrategy.runRecovery(RecoveryStrategy.java:48) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.core.retry.RetryWithRecovery.handleRecovery(RetryWithRecovery.java:159) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.core.retry.RetryWithRecovery.executeOnce(RetryWithRecovery.java:112) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at io.github.resilience4j.retry.Retry.lambda$decorateCheckedSupplier$3f69f149$1(Retry.java:137) ~[resilience4j-retry-1.6.1.jar:1.6.1]
        at io.github.resilience4j.retry.Retry.executeCheckedSupplier(Retry.java:419) ~[resilience4j-retry-1.6.1.jar:1.6.1]
        at com.symphony.bdk.core.retry.resilience4j.Resilience4jRetryWithRecovery.execute(Resilience4jRetryWithRecovery.java:65) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.core.service.datafeed.impl.DatafeedLoopV1.readDatafeed(DatafeedLoopV1.java:121) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.core.service.datafeed.impl.DatafeedLoopV1.start(DatafeedLoopV1.java:86) ~[symphony-bdk-core-2.1.3.jar:2.1.3]
        at com.symphony.bdk.spring.service.DatafeedAsyncLauncherService.uncheckedStart(DatafeedAsyncLauncherService.java:88) ~[symphony-bdk-core-spring-boot-starter-2.1.3.jar:2.1.3]
        at com.symphony.bdk.http.api.tracing.MDCUtils$MdcRunnable.run(MDCUtils.java:59) ~[symphony-bdk-http-api-2.1.3.jar:2.1.3]
        at java.base/java.lang.Thread.run(Unknown Source) ~[na:na]

Expected Result:

HTTP 500 errors should be retried, see: https://github.com/finos/symphony-bdk-java/blob/main/symphony-bdk-http/symphony-bdk-http-api/src/main/java/com/symphony/bdk/http/api/ApiException.java#L81

Actual Result:

HTTP 500 errors lead to bot failing.

[SivaTharun]

@symphony-elias i am a newbie to open source contribution, can i start working on this issue.

[thibauult]

No problem @SivaTharun, your contribution will be very welcome!

Do you need any help before starting?

[SivaTharun]

Thanks @symphony-thibault for following up, i just need to know how to reproduce the above scenario, if you can point me to a unit test, which asserts for the 500 error from DatafeedAsyncLauncherService class (or) you can point to me to the resource for reproducing the above exception.