Might be beneficial to test that implementations dont crash/segfault when experiencing very high throughput.
As a very simple (and possibly not helpful) example:
DiscordApp bots require a websocket conenction. Bots that are written in JS can use the ws package, or the uws package.
uws is significantly faster than ws, but when you are running a bot (or a shard) that is on a large number of guilds/servers (like, 1000+), uws starts segfaulting due to the throughput (amount and size of the traffic). The ws package does not have this issue at all, all the way to the guild cap per shard of 2000.
Might be beneficial to test that implementations dont crash/segfault when experiencing very high throughput.
As a very simple (and possibly not helpful) example:
DiscordApp bots require a websocket conenction. Bots that are written in JS can use the
wspackage, or theuwspackage.uwsis significantly faster thanws, but when you are running a bot (or a shard) that is on a large number of guilds/servers (like, 1000+),uwsstarts segfaulting due to the throughput (amount and size of the traffic). Thewspackage does not have this issue at all, all the way to the guild cap per shard of 2000.