Sink process abruptly terminates with an error

[let4be]

ERRO (sink-postgres) substreams encountered an error {"error": "receive stream next message: rpc error: code = Internal desc = subscription channel at max capacity"}
My substream is running in a development mode and can become throttled by an RPC(RPS limit) on dense blocks

Shouldn't this error be handled more gracefully?

[let4be]

Usually when channel buffer becomes full producer should back off, here it affects consumer's side which is ... unexpected

[maoueh]

This is an issue within the Substreams server directly, seen by rpc error: code ... message which tells that the error is coming from the server. Seems like the server close the connection and reported this error.

It's more asubstreams issue than a substreams-sink-postgres issue.

[let4be]

We can move it, it's just originated when using substreams-sink-postgres
It also "could be probably" reported at substreams-sink, he-he

[maoueh]

Definitely an error on substreams server, might be more visible when using a sink that is connected to a database. There is probably a wrong handling in substreams server when client don't consume message(s) fast enough.