Help with collecting bandwidth stats using BandwidthTransport in Rust-libp2p

[Hediii-coder]

Hi all,

I’m working on a Rust project using rust-libp2p and trying to instrument my network with bandwidth metrics. I’m using the BandwidthTransport wrapper around my transport and have set up a prometheus_client::Registry to collect metrics.

Here is what I have so far:

I create my base transport (TCP + Noise + Yamux).

I wrap it with BandwidthTransport::new(transport, &mut registry).

I then create a Swarm using this bandwidth-wrapped transport.

My goal is to periodically read bandwidth stats (sent/received bytes, rates, etc.) from the Metrics associated with the BandwidthTransport and output them (e.g., log or expose via Prometheus).

I’m struggling with:

How exactly to access the bandwidth stats from Metrics or BandwidthTransport after the swarm is running.

How to properly integrate this with SwarmBuilder::with_bandwidth_metrics() if I want to use it.

Any examples or idiomatic ways to expose or print bandwidth usage in runtime.

I haven’t found clear examples or docs on how to read and output these bandwidth metrics at runtime.

Could anyone provide guidance, example snippets, or best practices for:

Accessing current bandwidth stats from the swarm or metrics instance.

Integrating BandwidthTransport with SwarmBuilder::with_bandwidth_metrics().

Exposing or logging bandwidth data during the swarm event loop.

I can share part of my code if it is possible

Thanks in advance for any help!

[jxs]

Hi, have you looked into the metrics example?

[Hediii-coder]

Hi jxs,
Thanks for your help!
I switched to using with_bandwidth_metrics in the SwarmBuilder instead of BandwidthTransport. Here's the relevant code:

let mut swarm = libp2p::SwarmBuilder::with_existing_identity(localkey)
.with_tokio()
.with_tcp(
libp2p::tcp::Config::default(),
noise::Config::new,
yamux::Config::default,
)?
.with_bandwidth_metrics(&mut registry)
.with_behaviour(|key| {
// Gossipsub config...
})?
.build();

However, I’m not sure how to access the bandwidth stats (e.g., bytes sent/received) from the Swarm at runtime. How can I read and log these metrics periodically?

Thanks for any advice!

[jxs]

see https://github.com/libp2p/rust-libp2p/tree/master/examples/metrics#opentelemetry

[Hediii-coder]

Hi jxs,

I was finally able to collect some metrics and wanted to clarify exactly what they represent. Here’s a snapchot of my log file:

/ip4/tcp,Outbound,1572
/ip4/tcp/p2p,Outbound,3204
/ip4/tcp/p2p,Inbound,1044
/ip4/tcp,Inbound,2484

I found it a bit strange that /ip4/tcp/p2p,Outbound can exceed /ip4/tcp,Outbound. Could you help me understand why this might happen?

[jxs]

can you elaborate on what do you mean by snapshot of your log file?

[Hediii-coder]

I'm actually logging the data that is generated from the use of with_bandwidth_metrics to a csv file. By “snapshot of my log file,” I meant that I took a short extract or sample of the log data at a particular moment in time to illustrate the metrics I was seeing. It’s not the full log just a few lines that summarize the inbound and outbound traffic counts for different protocols.
Specifically, I was showing counts for /ip4/tcp and /ip4/tcp/p2p traffic in both directions, to clarify why I noticed that /ip4/tcp/p2p,Outbound seemed higher than /ip4/tcp,Outbound

[jxs]

ok I see, it probably is related to the specific setup or context of your project. I don't know which Behaviour you are using, it seems you are trying to communicate with IPFS nodes, but I really can't help further