-
Notifications
You must be signed in to change notification settings - Fork 39
SQL signatures overwrite #86
Comments
Hey @stas Thanks for reaching out regarding this issue. I believe you're right and that this is a limitation in the current implementation. We have some different ideas about to better bring to light the slow traces that can get shadowed. I'll leave this open so you can track our progress. |
We applied a wild patch (https://github.com/alexanderad/opbeat_python/commit/ca950b27063d8420402122ff6c60af97d66aa602) to double check described issue. Perhaps results displayed in this way are not super fancy, but true performance report on multiple (but different) queries hitting same table is shown. Queries which are identical should be still grouped together, since they have same hash signature: |
@alexanderad I'm happy. Btw, the pic is no longer available. @roncohen it would be great to have your feedback on this too! Ty! |
Updated pic link. |
We were hunting down a slow endpoint and found this.
If you have two queries hitting the same table inside one call, only one query will report.
Looking at the way
extract_signature()
works, this theory sounds valid:https://github.com/opbeat/opbeat_python/blob/master/opbeat/instrumentation/packages/dbapi2.py#L125
In our case, a query that is OK would shadow a query that causes slowness.
I might be wrong, but if so, I can't explain otherwise why we would not see both queries.
The text was updated successfully, but these errors were encountered: