alt Hacker NewsGenerally the more throughput the system supports the better. In this case we were hitting limits (btw each operation is many queries of different sizes) and the customer observed higher latencies which is typical if the system can't sustain the throughput required.
After this change latencies are back to normal and throughput increased.
Ahh, so it was a customer pain point of higher latency so they were happy to see latency go down and throughput go up. Good to hear.
Great write up, cheers to the people involved.