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
I am curious how the throughput (txn/sec) changes with different number of partitions.
Sure! We can test that out.
Also I am curious what the latency (ms/txn) will look like with a more realistic interval between transactions.
Will test with different interval (current interval == 0)
We are measuring time from the creation of transaction context to the completion of write
Correct!
The congestion at TransactionMonitor may be making the latency look worse. Maybe we should start another timer when transaction is execute to get the time excluding the waiting time on TransactionMonitor.
Sure!
As a summary, three things to do:
Add performance test with traffic in multiple partitions (need change in PerformanceCli)
Add performance test with different intervals
Add another timer that excludes waiting time on TransactionMonitor. (need change in PerformanceCli)
We need to add benchmark test for waltz with ducktape. So that we can
Expecting performance report, like:
The text was updated successfully, but these errors were encountered: