A Performance test was configured to run for 9 hours. The test ran to completion but the test reports only show data for a little over 1 hour.
Messages similar to the following are recorded in the jetpack.log file on the engines of the test:
2021-01-05 21:16:30,086 [parallel-1] WARN c.blazemeter.jetpack.data.DataSender - Got an unexpected error while submitting
java.net.SocketTimeoutException: Read timed out
2021-01-05 21:18:08,136 [parallel-7] DEBUG reactor.retry.DefaultRetry - Retries exhausted, retry context: iteration=9 exception=java.net.SocketTimeoutException: Read timed out backoff={EXHAUSTED}
The kpi.jtl files on the engines of the test show that metrics were successfully collect for all 9 hours of the test.
There were no outages recorded during this time and no internal alarms were raised indicating that the data service was having any issues.
Release : SAAS
Component : BLAZEMETER PERFORMANCE TESTING
Unless the behavior can be reproduced, the root cause of this cannot be definitively determined. There may have been a communication blip to the data service or something else may have occurred which caused this issue.
On the engines, jetlag attempts to send metric data in a request to the data service 10 times. After 10 failed attempts it will stop trying to do so. This is by design. The test will continue to run, but the data will no longer be uploaded.