Live data from Hacker News

Tracing Goroutines in Realtime with eBPF

sazak.io

1–10 of 12 posts

Re: Tracing Goroutines in Realtime with eBPF

#3
Interesting to see an ebpf approach to this idea, some time ago I did the same with Dtrace: https://gaultier.github.io/blog/detecting_goroutine_leaks_wi...

However I did not think of observing the ‘castgstatus’ runtime function.

Always nice to be able to compare the DTrace and Ebpf approaches, they both have different strong points.

Re: Tracing Goroutines in Realtime with eBPF

#5
that is actually sick.

how common is it for go devs to experience leaking goroutines ? id like to think go is a lot less shoot yourself in the foot here since they provide a framework for concurrency/parallelism rather than you working with the tiny pieces of it and building out the architecture yourself, but ive only needed to use goroutines once and it was a pretty problem-free experience.

Re: Tracing Goroutines in Realtime with eBPF

#7

Cool! is there something similar for async rust ?

https://docs.rs/tokio-console/latest/tokio_console/ comes to mind.

I'm not aware of eBPF-based dynamic instrumentation for either tokio or other async runtimes, but I'm also not (in need to be) debugging this on a daily basis.

Re: Tracing Goroutines in Realtime with eBPF

#8
post #5

that is actually sick. how common is it for go devs to experience leaking goroutines ? id like to think go is a lot less shoot yourself in the foot here since they provide a framework for concurrency/parallelism rather than you working with the tiny pieces of it and building out the architecture yourself, but ive only needed to use goroutines once and it was a pretty problem-free experience.

> how common is it for go devs to experience leaking goroutines

About as often as leaking memory in C++

Re: Tracing Goroutines in Realtime with eBPF

#10
post #4

So cool! Does it support multiple Go versions?

I would say that this is limited to a targeted Go version and architecture. For example, the filter checks for the goroutine pointer on `r28` which is correct for arm64 but not universally true.

Any changes to the struct layouts, stack, or heap layouts would also cause failures in these lookups. E.g., in Go 1.17, many functions now use direct register mappings for arguments rather than always placing arguments in the stack.

Would need to thoroughly vet compatability with each new Go version before using something like this in production.

Post reply on HN