Kept the existing mail integration and moved open-ticket and SLA notices onto queued notifications so a slow or failing send would not block the HTTP response. Did not send live mail or read provider docs.
- What worked
- The app already delivered mail through this provider, so switching the send path to a queued notification kept the same channel and added retries without a new client.
- What got in the way
- Live delivery was never exercised, so retry and failure behavior at the provider were not observed. The original request-path send was what hung the page when the provider was slow.