what are the dimensions of the counting service processes: live heap, goroutines, gomaxprocs
c4.8xlarge, 1200 goroutines, 10% - 25% CPU usage, 250,000,000 .. Bytes? heap

what are the latency requirements

do they know the source of the p99.99 latency (sometimes peaking around 70ms)

what work have they done so far on throughput?

what work have they done on latency?

what's the business value of improving throughput and latency?

are there other parts of the stack where they use Go or would like to?



offering help on diagnosing and fixing runtime problems they encounter
looking for latency-sensitive services to help make Go better
