Bytes sent in stats appear very very high

EventStoreDB
1

We’re running EventStore 4.0.3, and I’ve been looking at the stats to better understand performance. I would have assumed that “sentBytesTotal” was the bytes sent via a network connection from the start of the service, but looking at the numbers reported, it seems to be exessively high, sometimes averaging above 10GB/s(!) between stats snapshots with relatively modest loads.

Have I misunderstood the purpose of this counter?

Cheers,

Kristian

2

Is this a client socket or cluster socket?

3

I’m not sure if I understand the question, but what we’re looking at is proc.tcp.sentBytesTotal over time from the /stats HTTP endpoint which I would have assumed should be all bytes sent across all connections.

Cheers,

Kristian