jenny1371 commented on PR #6682: URL: https://github.com/apache/fineract/pull/6682#issuecomment-6254487068
For the record, the measurements behind closing this (k6, deposits spread over 10 accounts, 60 s per run, same image, only the jitter setting differs, one run per setting): | Load | jitter 0 (today) | jitter 0.5 | |---|---|---| | 50 deposits/s | 87.8 % failed, p95 13.5 s | 89.1 % failed, p95 15.1 s | | 15 deposits/s | 8.5 % failed (75/885), p95 5.6 s | 7.5 % failed (66/885), p95 4.7 s | All failures were `OptimisticLockException` answered as 409 after the 3 attempts were used up. At 50/s the accounts are saturated (about 5 deposits/s per account) and randomized back-off cannot help; at 15/s the difference is within what one run can show. So the hypothesis that jitter reduces contention failures is not supported on this setup. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
