Drew Fisher

So first, some background: we have two approaches we use to have servers monitor each other for liveness: ## Servers collection Each server is [periodically](https://github.com/deathandmayhem/jolly-roger/blob/d1e1b88ea771c5d58d6be78f1160bd08ae454609/imports/server/garbage-collection.ts#L75-L78) issuing a write to the...

Ahhhh, you're right; I'd forgotten about [`CLUSTER_WORKERS_COUNT=auto`](https://github.com/deathandmayhem/jolly-roger/blob/d1e1b88ea771c5d58d6be78f1160bd08ae454609/imports/server/loadBalance.ts#L21-L23), which attempts to run one child worker per CPU. And each worker gets its own `serverId` (and spawns its own mediasoup worker) and...

I'll also throw in a link to #1180 here since it's an open item around "what should our clustering strategy be"

Exactly that. Especially later on in the hunt, once people have chatted a lot, backend performance for the subscription used on that page suffers as it winds up fetching >10000...

Adding a warning next to the configuration options sounds imminently reasonable!

Yeah, I think the current behavior here is coincidental more so than intended. I've felt similarly about wanting successful submission at least to clear title and URL, though I never...

> The UX. Do we always want to show this checkbox as is done here? Or should it be a more interactive flow, e.g. a (nested) dialog indicating that the...