All app.bsky.* NSIDs are on the blocklist (in the env var) in production. Because I'm afraid that ingesting Bluesky data will take a lot of resources. There are some many events.  But, jetstreams subscriptions can actually be filtered on `wantedCollections` and `wantedDids`. We could force automation on `app.bsky.*` to have a conditions on `event.did equals {{self}}` (and eventually other DIDs, as we scale the infra). That would result in only a few events already filters on the repo DID. It's a different kind of conditions than the one we currently have, since it doesn't filter on received events, it directly filters the events we subscribe to. One potential limitation is URL length and the number of DIDs we can put in the websocket URL. In the future, we'll eventually shard jetstream subscriptions in different workers to mitigate this Relevant docs: - https://docs.bsky.app/blog/jetstream - https://atproto.com/guides/streaming-data#jetstream