Changing the Redis queue password deletes all queued events...

Nadie ha tomado este issue todavía.

We just lost 230,000 events of customer data because we changed our Redis queue password from the settings. I can't imagine why one would ever intentionally issue a FLUSH command after a setting is updated. Or, if you knew about this why is not in all caps in that setting card.

We have to periodically cycle passwords for security. We need to be able to do that from our Matomo instance WITHOUT having our system purge our queued events.

That has to be a bug. That behavior is extremely harmful. We are trying to use this system at the scale of 10M hits/day. It's going to take days for us to try to reconstruct the lost data that the queueing system just purged unexpectedly.

No se nombran archivos ni pruebas. Empieza rastreando la ruta de actualización de la configuración de la contraseña de la cola de Redis y dónde se vacían los eventos encolados; se considera terminado cuando cambiar la contraseña desde Matomo ya no purga los eventos encolados existentes, con cobertura para garantizar que la rotación de la contraseña preserve la cola.

Escrito por el modelo de indexación a partir del texto del issue.