WebSocket ping interval (500 ms) intentional?

[0xsimulacra]

Hello,

On the current main branch, the client pings the WebSocket and waits for a pong every 500 ms.

1 - Is the 500 ms ping interval intentional for main?

Context / Rationale

• In a WebSocket client, I typically set the ping interval to about 50–55 s (most servers send a ping every ~60 s and expect a pong anyway, else the server will close the connection).
• Even though the ping payload is empty on main, a 500 ms interval across many nodes could look like constant traffic and potentially burden proxies.

[haardikk21]

Need to reconnect ASAP if node stops receiving flashblocks for some reason. Perhaps a backoff could be good here if it fails to reconnect but we're okay with doing 500ms right now and its implications on the proxy

[0xsimulacra]

Thanks for the reply—much appreciated.

There’s already a 10-second max backoff implemented from what I can tell (it’s “hidden” in the sleep function, so it isn’t obvious at first).

I was running a modification of the current branch with a 50-second ping interval. It’s good to know I can change it to a 500 ms interval now.

If it’s useful, I can open a PR to log the RTT between the node and the proxy with each response as well ?

[haardikk21]

maybe a debug level log should be fine, yeah. anything higher is probably too verbose

[0xsimulacra]

@haardikk21

Apologize for the delay but I just managed to make the PR regarding the RTT today.