fix(nats): revert to consumer.messages() — fetch() defaults no_wait:true which skips empty queues
Some checks failed
lint / lint (push) Has been cancelled
test / unit (push) Has been cancelled
test / integration (push) Has been cancelled
lint / lint (pull_request) Failing after 9m21s
test / unit (pull_request) Failing after 10m57s
test / integration (pull_request) Failing after 16m44s

This commit is contained in:
2026-05-14 23:34:51 +02:00
parent a4377fe209
commit 4d2d56c8ae

View File

@@ -139,19 +139,14 @@ impl MessageSource for NatsMessageSource {
tracing::info!("NATS pull consumer ready");
loop {
// fetch().expires() keeps the request open server-side until messages arrive
// (up to 30 s), then the batch ends and we loop. Without expires, the fetch
// returns immediately with zero messages when the queue is empty.
let mut messages = match consumer
.fetch()
.max_messages(100)
.expires(std::time::Duration::from_secs(30))
.messages()
.await
{
// consumer.messages() uses long-poll (no no_wait flag) — NATS holds the
// request open and delivers messages as they arrive.
// fetch() in async-nats 0.48 defaults to no_wait:true which returns
// immediately when the queue is empty, so we avoid it here.
let mut messages = match consumer.messages().await {
Ok(m) => m,
Err(e) => {
tracing::error!("NATS fetch failed: {e}");
tracing::error!("NATS messages() failed: {e}");
let _ = tx.send(Err(DomainError::Internal(e.to_string()))).await;
return;
}