The Microsoft 365 app list on a screen

How we kept our email running through the Microsoft 365 outage

Microsoft 365 went down worldwide on the evening of Thursday 12 November. Email and calendars were unavailable for most of Friday.

Our email was back up at 08:37 on Friday morning, running on a temporary server of our own. Mail kept arriving while Microsoft was down, and once it came back we moved all of it into Outlook. Nothing was lost.

Friday morning

We confirmed the outage at 08:03. Ibernia Fraser, our Head of IT, and I had our mail switched over to the temporary server 34 minutes later, so the team could send and receive and clients could still reach us.

Microsoft's service came back in the afternoon, and we moved our mail back at 16:52. By then the temporary server had taken in 5,697 messages, and we loaded every one into Outlook. The mail stayed on systems we control throughout, so there was no data protection issue.

We also kept clients updated through the day with alerts on our client projects system and on the homepage of this website. If anything like this happens again, those are the first two places to check.

Where the gaps were

Email was the one service we had a fallback for. Teams chat and SharePoint mostly held up through the day, but calendars stayed down until Microsoft restored them on Friday afternoon, and meetings were disrupted with them. The temporary server also only carried mtc.'s own mail, so clients whose Microsoft 365 we manage were affected in the same way as everyone else. Both gaps have shaped what we've done since.

What's changed since

We're certified to ISO 27001, the international standard for information security. It sets out how a business protects the information it holds and keeps its systems running, and an independent auditor checks we're doing it every year. So the outage went through our usual incident process. It was logged, reviewed and closed, with an action to fix the gaps.

We've now built a standby system that gives us email, a chat room for each active team, and video calls in the browser. It stays switched off until we need it, updates automatically, and its user list follows our staff list. We'll test it once a quarter by switching it on, the same way we test restoring our backups. The first test is before 29 January.

We can’t stop Microsoft having a bad day, but what we can do is be ready for it. We had our emails running again on our continuity systems in 34 minutes, without compromising data security or integrity. This is exactly what ISO 27001 means day to day for us: you plan for the bad day before it arrives.

Ibernia Fraser, Head of IT, mtc.

Keeping your business running

What we built for ourselves can be built for other businesses on Microsoft 365. It's a standby email service, set up ahead of time around how your business works and tested regularly, so that if Microsoft goes down your team can keep sending and receiving.

Every business relies on different things, so we'd start by talking through what yours needs to keep going.

Practising for the next one

Outages like this will happen again, so we've started running regular "war gaming" sessions. Each one takes a realistic scenario, some of which have already happened, and walks through how we'd respond: who does what, and how we stay in touch with clients. Where we find a gap, we fix it, so the next time something goes wrong we're working from a plan we've already rehearsed.

If it had been your business

How long would it take you to get email back if your provider went down? Where would incoming mail go in the meantime, and how would your team talk to each other?

Keep your email running when Microsoft doesn't

We'll talk through how your business uses Microsoft 365 and what a standby plan would look like for you.

Talk to us