Background propagation¶

Relays registered in the pool but disconnected because they are outside the connection budget can still receive public events through background propagation. The pool connects to them one at a time to propagate events on a best-effort basis, alongside normal pool activity.

Normal publication follows the acknowledgement policy for connected relays. A separate queue attempts standby relays one at a time.

Setting a positive budget enables this behavior automatically:

NostrPool pool = new NostrPool();
pool.setConnectionBudget(2);
pool.setBestEffortPropagationQueueCapacity(256);
pool.setBestEffortPropagationInterval(Duration.ofMillis(250));

pool.registerRelay(new NostrRelay("wss://relay.ngengine.org"));
pool.registerRelay(new NostrRelay("wss://relay2.ngengine.org"));
pool.registerRelay(new NostrRelay("wss://nostr.rblb.it"));

Publish as usual with pool.publish(signed). The pool sends to active relays and queues deliveries to the standby relays.

Use getPendingBestEffortPropagations() to check how many deliveries remain, including the current attempt.

Publication result

A successful publish follows your acknowledgement policy for the immediate sends. Background deliveries may still be pending.

Disable background propagation

Call pool.setBestEffortPropagation(false) to turn it off. Pass true to enable it again with a positive connection budget.

Best-effort background propagation

Background propagation is best effort: the pool tries to publish to as many standby relays as possible without holding up normal publication to connected relays. Background attempts have a time limit, a full queue drops its oldest unstarted delivery, and closing the pool clears pending background work.