On 3 November 2026, between 11:42 and 12:30 AEDT (GMT+11), some emails sent through Maileroo were delayed. A small number of emails accepted during this period may not have been delivered. We are sorry for the disruption, and this post explains what happened and what we will change.

What Happened

A single customer sent an unusually large burst of email: more than 90,000 messages, over 80 GB in total in a few seconds. All of this traffic went to one of our mail transfer agents, which could not keep up, so a backlog built up in our internal queue.

The queue is held in memory, and the backlog grew until the queueing server(s) ran out of RAM. Our automatic failover then tried to move the primary role to another node, but the new node could not take on the full dataset either. The system got stuck in a loop of failovers and restarts, and the queue was unstable until we stepped in.

Timeline (AEDT, GMT+11)

  • 11:42: Issue detected

  • 11:51: Engineering team alerted and investigation started

  • 11:53: Recovery started; email sending resumed on all mail transfer agents except MTA 6

  • 12:00: Queueing database fully restored

  • 12:05: MTA 6 taken offline so its load could be spread across our other mail transfer agents

  • 12:15: MTA 6 back online with higher limits

  • 12:30: Emails stuck on MTA 6 fully drained and delivered

Impact

  • Most emails sent during this window were delivered, but with a delay.

  • Some emails accepted by our API while the queueing database was failing over may have been lost and not delivered.

What We've Changed and What's Next

Already changed

  • Sending capacity: we've raised the processing limits on our mail transfer agents so a large burst from a single customer is processed without building a backlog.

Coming next

  • New hardware: we're moving our queueing infrastructure to new hardware with more memory headroom, so a surge in queued email can't exhaust it.

  • Failing faster and louder: if an email can't be queued, our API and SMTP servers will return an error you can retry on instead of accepting it. Queue backlogs and database failovers will alert our engineers straight away.

What You Should Do

If you sent emails between 11:35 and 12:00 AEDT on 3 November and your logs show no delivery event for some of them, please resend those messages. If you need help identifying affected messages, contact our support team and we'll work through it with you.

A Note from Our Team

We are truly sorry. You trust Maileroo to deliver email that matters to your business and your customers, and on 3 November we let you down. Delays are bad enough, but losing emails is something we never want to happen, and we take full responsibility for it.

We are grateful for your patience while we fixed this. We know trust is earned through reliability, not words, and we are committed to earning yours back through the changes above. If you were affected, please contact our support team.

Thank you for sticking with us.