We sometimes see `grpc: failed to set
# spicedb
t
We sometimes see
grpc: failed to set send compressor: transport: set send compressor called after headers sent or stream done
inside
dispatch.v1.DispatchService/DispatchCheck
. The request overall seems successful. Is this something to worry about, or something to be expected?
y
i haven't seen that before. i don't think we're manually managing compressors on our internal API clients, and it sounds like it's retrying or otherwise dealing with it. I think this might be an implementation detail of the grpc-go client?
v
spicedb uses snappy compression in the dispatch API
what version are you running?
e
I'm seeing same issues as well. I am using 1.39.1
v
but where do you see this, in the spicedb logs? or is it returned as an error to the client?
t
It's form an otel span,
dispatch.v1.DispatchService/DispatchCheck
. We're also on 1.39.1, but i'm pretty sure i've seen it before on older versions.
v
you are saying that is shows in an otel span?
v
interesting
mind opening an issue with the image?
dispatch requests are canceled often, which shows as an error
that's part of how SpiceDB works - I wish there was a way to not show those as failed in the trace
but that is new
v
thank you!
8 Views