You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Description
We have observed several CI failures where 38 log events are expected, but only 37 are captured/forwarded. Initially we thought that this must be due to a timing issue with a log failing to be captured at the beginning of the test run, or a log failing to be sent up due to timing at the end of a test run. After investigating, the missing log is usually from the middle of the test run and logs before/after the missing log were forwarded without issue.
This indicates that there may be an intermittent timing issue with log events being lost (potentially related to harvest cycle).
Failures like this typically show up in the form of:
Here is an excerpt of the ConsoleMF test app showing 5 log messages of the form AsyncNoAwaitInTransaction{Level}LogMessage being requested, and output to the console by log4net:
If you search the sent JSON payload you will not find AsyncNoAwaitInTransactionErrorLogMessage, but all of the surrounding logs (before and after) are in the JSON payload. Interestingly, the seen metric is missing this log as well:
Initially I thought this might be a race condition during log event harvest, but after investigating more, I am not sure that is what is happening. The counts should be incremented regardless of any race condition.
This warrants more investigation and appears to be an intermittent agent bug instead of a CI flicker.
Expected Behavior
The agent captures and forwards all logs within the sampling limits.
For Maintainers Only or Hero Triaging this bug Suggested Priority (P1,P2,P3,P4,P5): P4 Suggested T-Shirt size (S, M, L, XL, Unknown): S
The text was updated successfully, but these errors were encountered:
Description
We have observed several CI failures where 38 log events are expected, but only 37 are captured/forwarded. Initially we thought that this must be due to a timing issue with a log failing to be captured at the beginning of the test run, or a log failing to be sent up due to timing at the end of a test run. After investigating, the missing log is usually from the middle of the test run and logs before/after the missing log were forwarded without issue.
This indicates that there may be an intermittent timing issue with log events being lost (potentially related to harvest cycle).
Failures like this typically show up in the form of:
Here is an excerpt of the ConsoleMF test app showing 5 log messages of the form
AsyncNoAwaitInTransaction{Level}LogMessage
being requested, and output to the console by log4net:Here is the
log_event_data
payloads sent to the collector:If you search the sent JSON payload you will not find
AsyncNoAwaitInTransactionErrorLogMessage
, but all of the surrounding logs (before and after) are in the JSON payload. Interestingly, the seen metric is missing this log as well:Initially I thought this might be a race condition during log event harvest, but after investigating more, I am not sure that is what is happening. The counts should be incremented regardless of any race condition.
This warrants more investigation and appears to be an intermittent agent bug instead of a CI flicker.
Expected Behavior
The agent captures and forwards all logs within the sampling limits.
For Maintainers Only or Hero Triaging this bug
Suggested Priority (P1,P2,P3,P4,P5): P4
Suggested T-Shirt size (S, M, L, XL, Unknown): S
The text was updated successfully, but these errors were encountered: