Problem: the uptake-release/error-rate check compares successes vs failures. However if a client succeeds in syncing a large collection, we will see 1 success for the client. If it fails, we could see thousands of failures. Meaning this isn't an apples:apples comparison.
Possible solutions:
Adjust the check's logic to only count 1 status per client. (probably be easy to do with aggregation)
Exclude download-error messages as this is where the duplicate errors come from. This would then require a separate check for attachment downloads.
Problem: the
uptake-release/error-ratecheck compares successes vs failures. However if a client succeeds in syncing a large collection, we will see 1 success for the client. If it fails, we could see thousands of failures. Meaning this isn't an apples:apples comparison.Possible solutions:
download-errormessages as this is where the duplicate errors come from. This would then require a separate check for attachment downloads.