[Bug]: Serenity bot crashes due to excessive logs

[jonaro00]

What happened?

Serenity projects emit way too many log lines since #864, and can lead a bot to crash (not 100% confirmed).

The excessive logs are due to functions annotated with #[instrument] (spans) are set to INFO by default.

I think we can compromise by having the minimum level increased in the Shuttle log consumer for spans,
but it would be awesome if this was separately configurable with #842 (set separate span and event levels, per crate)

Version

v0.17.0

Which operating system(s) are you seeing the problem on?

Linux, In deployment

Which CPU architectures are you seeing the problem on?

x86_64, In deployment

Relevant log output

2023-05-31T02:03:10.586481275Z  INFO {shard="Shard { shard_info: ShardInfo { id: 0, total: 1 }, .. }"} serenity::gateway::bridge::shard_runner: [span] recv
2023-05-31T02:03:10.586582375Z  INFO serenity::gateway::shard: [span] do_heartbeat
2023-05-31T02:03:10.586616675Z  INFO {shard="Shard { shard_info: ShardInfo { id: 0, total: 1 }, .. }"} serenity::gateway::bridge::shard_runner: [span] recv_event

Duplicate declaration

• I have searched the issues and there are none like this.

[techs-sus]

Same here. Even though my bot isn't crashing, I am concerned about the ever expanding log file. These "span" log messages take up hundreds of lines of space in the file.

Besides, it makes hosting shuttle more expensive 😔.

[oddgrd]

Thanks for getting to the bottom of this @jonaro00! If the #842 param takes a RUST_LOG string, would that resolve this? If we can't get that over the line for the next release we could also change the minimum level in the spans consumer as you suggested, since IIRC only tide users were affected by missing logs before #864. And we have very few tide users (I reported the initial bug).

[jonaro00]

From what I understand, the RUST_LOG only targets events by defaults, and not spans. Not sure how complex the difference is.

I'll try lowering the span log level.