[Improvement] Suggestion for Handling Stale Metrics in Gravitino #3152
Assignees
Labels
improvement
Improvements on everything
Comments
|
thanks for reporting this issue. |
|
@TEOTEO520 , #3341 propose a way to compuate Pxx according to the time sliding window, you could check it when you are free. I'm not sure whether to make it configurable for now. |
jerryshao
pushed a commit
that referenced
this issue
May 27, 2024
…#3341) ### What changes were proposed in this pull request? using time sliding window to record timer metrics ### Why are the changes needed? Fix: #3152 ### Does this PR introduce _any_ user-facing change? no ### How was this patch tested? test in local env to check whether Pxx comes to 0 after a while
github-actions bot
pushed a commit
that referenced
this issue
May 27, 2024
…#3341) ### What changes were proposed in this pull request? using time sliding window to record timer metrics ### Why are the changes needed? Fix: #3152 ### Does this PR introduce _any_ user-facing change? no ### How was this patch tested? test in local env to check whether Pxx comes to 0 after a while
FANNG1
added a commit
that referenced
this issue
May 28, 2024
…#3591) ### What changes were proposed in this pull request? using time sliding window to record timer metrics ### Why are the changes needed? Fix: #3152 ### Does this PR introduce _any_ user-facing change? no ### How was this patch tested? test in local env to check whether Pxx comes to 0 after a while Co-authored-by: FANNG <[email protected]>
diqiu50
pushed a commit
to diqiu50/gravitino
that referenced
this issue
Jun 13, 2024
…etrics (apache#3341) ### What changes were proposed in this pull request? using time sliding window to record timer metrics ### Why are the changes needed? Fix: apache#3152 ### Does this PR introduce _any_ user-facing change? no ### How was this patch tested? test in local env to check whether Pxx comes to 0 after a while
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
What would you like to be improved?
Issue Description
There is a current limitation in the gravitino_server_http_request_duration_seconds{operation="xxx", quantile="0.xx"} metrics within Gravitino. If an operation is not called for a prolonged period after its last execution, the metric continues to display the duration of the last call. This behavior can lead to misinterpretations of the system's performance, as it does not accurately reflect the inactivity.
Suggested Improvement
It would be beneficial to modify the metric behavior to either reset to zero or drop the metric data when no calls are made to an operation within a predefined timeout period. This change would provide a more accurate representation of the system's current state and activity levels.
@FANNG1 +cc
How should we improve?
No response
The text was updated successfully, but these errors were encountered: