[go-nuts] Re: 10x latency spikes during GC alloc assist phase

Thanks a ton! We (at Laserlike) gave this a try and it basically cured our problem in one fell swoop. The GC cycles now take longer to run, especially under load, but memory growth isn't too bad. I know that Go's philosophy is not to add knobs to the GC, but there really needs to be a way to

[go-nuts] Re: 10x latency spikes during GC alloc assist phase

Hi. We have made a simple one line patch to Go that turns off GC assist. Unfortunately, it was the only way. I feel your pain. $ git diff diff --git a/src/runtime/malloc.go b/src/runtime/malloc.go index 8850659748..a2c8697bf5 100644 --- a/src/runtime/malloc.go +++ b/src/runtime/malloc.go @@

[go-nuts] Re: 10x latency spikes during GC alloc assist phase

OK, I will add to that issue , but for the record here is our machine type and the env vars: This GCP hardware has 64 virtual CPUs (a hyperthread on a 2.6GHz Xeon E5), with 416GB memory. This is shared among other docker containers, of course, but as the trace file shows, our process uses 8

[go-nuts] Re: 10x latency spikes during GC alloc assist phase

I would add to 14812 . The report should include the environment variables, HW , and RAM. The report should indicate if any environment variables are not set to the