Queue outgoing events when relay connection is down #1

New issue

Open

opened 2026-04-27 17:33:45 +00:00 by padreug · 0 comments

padreug commented

2026-04-27 17:33:45 +00:00

Owner

Problem

When the relay connection is temporarily lost, nostrclient drops any events published during the disconnection window. Extensions (nostrmarket, events) publish via nostrclient's internal WebSocket, which queues the ["EVENT", ...] message — but nostrclient's relay connection may be down at that moment, and the event is silently lost.

Observed: events extension published a NIP-52 calendar event at 17:21:58, but the nostrclient relay connection was lost at 17:18:43 and didn't reconnect until 17:22:53. The event never reached the relay.

Expected behavior

Outgoing EVENT messages should be queued and retried when the relay reconnects, not dropped.

Proposed solution

Add a bounded outgoing event queue (e.g., 100 events max)
When relay.send() fails or relay is disconnected, queue the event
On reconnection, flush the queue before accepting new events
Drop oldest events if queue is full (bounded to prevent memory issues)

Impact

This affects all extensions that publish via nostrclient:

nostrmarket (stall/product/DM publishing)
events (NIP-52 calendar event publishing)

Workaround

Currently, extensions can re-publish events manually (e.g., re-approve an event). But this isn't automatic.

## Problem When the relay connection is temporarily lost, nostrclient drops any events published during the disconnection window. Extensions (nostrmarket, events) publish via nostrclient's internal WebSocket, which queues the `["EVENT", ...]` message — but nostrclient's relay connection may be down at that moment, and the event is silently lost. Observed: events extension published a NIP-52 calendar event at 17:21:58, but the nostrclient relay connection was lost at 17:18:43 and didn't reconnect until 17:22:53. The event never reached the relay. ## Expected behavior Outgoing EVENT messages should be queued and retried when the relay reconnects, not dropped. ## Proposed solution - Add a bounded outgoing event queue (e.g., 100 events max) - When `relay.send()` fails or relay is disconnected, queue the event - On reconnection, flush the queue before accepting new events - Drop oldest events if queue is full (bounded to prevent memory issues) ## Impact This affects all extensions that publish via nostrclient: - nostrmarket (stall/product/DM publishing) - events (NIP-52 calendar event publishing) ## Workaround Currently, extensions can re-publish events manually (e.g., re-approve an event). But this isn't automatic.

padreug referenced this issue from a commit

2026-04-27 18:09:59 +00:00

fix: queue outgoing events when relay connection is down

padreug referenced this issue from a pull request that will close it,

2026-04-27 18:38:28 +00:00

fix: queue outgoing events when relay connection is down #2