Skip to content

Latest commit

 

History

History
51 lines (34 loc) · 2.08 KB

configuration.md

File metadata and controls

51 lines (34 loc) · 2.08 KB

Configuration

Tune WebSocket behavior and retries for your environment.

Purpose

  • Adjust timeouts for send operations.
  • Understand retry behavior for transient I/O failures.
  • Configure WebSocket buffer sizes for large events.

WebSocket client settings

NostrRelayClient reads the following properties (with defaults):

  • nostr.websocket.await-timeout-ms (default: 60000) — Max time to await a response after send.
  • nostr.websocket.max-idle-timeout-ms (default: 3600000) — Max idle timeout for WebSocket sessions. Set to 0 for no timeout. Prevents premature connection closures when relays have periods of inactivity.
  • nostr.websocket.max-text-message-buffer-size (default: 1048576) — WebSocket text message buffer size in bytes.
  • nostr.websocket.max-binary-message-buffer-size (default: 1048576) — WebSocket binary message buffer size in bytes.

Example (application.properties):

nostr.websocket.await-timeout-ms=30000
nostr.websocket.max-idle-timeout-ms=7200000
nostr.websocket.max-text-message-buffer-size=2097152

Retry behavior

WebSocket send and subscribe operations are annotated with @NostrRetryable:

  • Included exception: IOException
  • Max attempts: 3
  • Backoff: initial 500ms, multiplier 2.0

To customize globally, create a custom annotation or replace @NostrRetryable with your configuration.

Virtual Thread dispatch

Relay subscription callbacks and async operations (connectAsync, sendAsync, subscribeAsync) are dispatched on Virtual Threads using named thread factories:

  • nostr-relay-io-* — relay I/O operations
  • nostr-listener-* — listener/callback dispatch
  • nostr-http-* — HTTP client operations (NIP-05 validation)

Notes

  • Timeouts apply per send; long-running subscriptions are managed separately.
  • If a relay response is not received before the timeout elapses, NostrRelayClient throws RelayTimeoutException.
  • The maximum number of events accumulated per blocking request defaults to 10,000 to prevent unbounded memory growth.
  • Ensure your relay endpoints' SLAs align with chosen timeouts and backoff.