junkerderprovinz 654f0e8e20 fix(postprocessing): retry publishing events instead of killing the server ...

A single transient nats: timeout on event publish was fatal to the whole
single-binary server, with no retry despite the service already having
backoff machinery for step failures. Now retries with the same exponential
backoff, capped at half the ack wait so the source event isn't redelivered
to another worker mid-retry. Also stops acking the source event when
publish ultimately fails, so the upload resumes after restart instead of
staying orphaned, and no longer exits during graceful shutdown.

2026-08-19 11:23:46 +02:00

..

service_suite_test.go

fix(postprocessing): retry publishing events instead of killing the server

2026-08-19 11:23:46 +02:00

service_test.go

fix(postprocessing): retry publishing events instead of killing the server

2026-08-19 11:23:46 +02:00

service.go

fix(postprocessing): retry publishing events instead of killing the server

2026-08-19 11:23:46 +02:00