Microsoft 365 Outage: The First Hour Is a Leadership Test

At 11:17 Eastern on Thursday, Microsoft confirmed problems with Exchange Online. By early afternoon it had spread. Microsoft’s service health notices listed Outlook, Teams, SharePoint, Copilot, the Microsoft 365 admin center and the Defender XDR portal among the affected services, across multiple regions. Recovery dragged through the night. On Friday morning Microsoft was still reporting “intermittent connectivity issues or delayed email delivery” for some users.
Microsoft says it wasn’t an attack. Its updates said a large number of servers became unavailable, the remaining ones couldn’t carry the load, and “overall traffic had not increased.” Charles Lamanna, the Microsoft executive who spoke for the company, said “there is no indication this is the result of malicious or cyber activity.” As of Friday, Microsoft hadn’t confirmed a root cause and had promised a full post-incident report once the problem was fixed.
So it was a plain outage, the kind every cloud provider has eventually. The technical side was Microsoft’s problem. What happened inside your company in the first hour was yours.
The first hour is a leadership test
When email and Teams go dark, most organizations don’t have a technology problem they can solve. They have a coordination problem. Nobody can reach anybody through the usual channels, rumors start in the hallway, and people start improvising, some of it badly.
That first hour shows whether a leader has thought about this before. It comes down to three things: what you say, what you decide and what you hand off.
What you say
Say something early, even if all you know is that it’s not you.
The first question in every office on Thursday was whether this was a local problem or a big one. You can answer that in a few minutes. Microsoft posts updates on its public status page and on its Microsoft 365 Status account on X, and news sites were covering it by early afternoon. On Thursday the admin center itself was one of the affected services, so don’t count on getting into it.
Then tell your people three things, through whatever channel still works:
- What’s happening. Microsoft 365 is down across the country. It’s not our network and, as far as Microsoft can tell, it’s not an attack.
- What to do. Which work continues, which waits, and how to reach each other until it’s back.
- When you’ll update them. Pick a time and keep it, even if the update is “no change.”
Customers need a version of the same message. If you can’t email them, a short phone call to the ones with deadlines this week goes a long way. One firm affected by the outage posted on its website that urgent inquiries should go to the phone, which is exactly the right instinct.
What you don’t say matters too. Don’t speculate about causes, and don’t promise a recovery time Microsoft hasn’t given.
What you decide
A leader’s job during an outage is to make the calls nobody else is authorized to make. A few that came up Thursday:
- Which deadlines move. If a proposal, a deliverable or an invoice is due today and depends on mail, decide now whether you ask for an extension, deliver another way or wait. Don’t let it drift until 4:59.
- Whether to treat it as a security event. Microsoft said no evidence of malicious activity. That’s Microsoft’s tenant-wide view. It’s still worth five minutes to have someone check your own sign-in logs and alerts once the tools come back, because outages are a good cover for phishing. Expect “your mailbox is full, log in here” messages in the days after.
- What people may not do. This is the one that bites small defense contractors. When work email is down, people forward files to personal Gmail or text documents to get the job done. If those files contain CUI, you’ve just turned an outage into a spill. Say plainly, early, that controlled information stays where it is, even if work waits. If you’re not sure what counts, my post on CUI vs. FCI in your contracts covers it.
- Whether to change anything. Rerouting mail, changing DNS or standing up a new system in the middle of someone else’s outage usually causes more trouble than it solves. Unless you planned it in advance, the right decision is usually to wait.
What you hand off
You can’t do all of this yourself, and you shouldn’t. Give each job to one named person:
- Vendor watch. Someone who checks Microsoft’s status every half hour and tells you when something changes.
- Customer contact. Someone who calls the customers with deadlines at risk.
- Internal updates. Someone who sends the scheduled update to staff.
Your job is to make decisions and stay reachable. If you’re the one refreshing the status page, nobody is leading.
When the first hour turns into the first day
Thursday’s outage didn’t end in an hour. Microsoft was still restarting and rebuilding servers late that night, and some users had problems into Friday. So plan for the second decision point too.
Around the end of the business day, decide what tomorrow looks like. Do people come in and work offline? Do you move a customer meeting to a phone call? Does anyone need to work late once mail comes back, to clear a backlog of delayed messages? Make the call before people go home, and send it through the alternate channel so nobody wakes up wondering.
Also expect a flood when service returns. Delayed mail arrives all at once, sometimes out of order. Tell people to check for anything time-sensitive that came in during the outage, and to confirm by phone before acting on a payment or banking change that arrived in that window.
A cheap backup plan for mail and calls
The Army plans communications with a PACE plan: primary, alternate, contingency and emergency. Every mission has one, because every primary system eventually fails. A small business can use the same idea for almost nothing.
- Primary. Microsoft 365: Outlook and Teams.
- Alternate. A group text or a messaging app on company phones, set up in advance with every employee in it. This is how you send the “here’s what’s happening” message.
- Contingency. A call tree: a short list of who calls whom, plus phone numbers for your key customers and partners. Keep it somewhere that isn’t in Microsoft 365, on paper or in a password manager.
- Emergency. A conference bridge or phone number you can publish on your website, and a small banner you can post that tells customers how to reach you.
For mail specifically, some small firms keep a basic account with a different provider, for outbound notices only, not for controlled information. It costs a few dollars a month. I resell Google Workspace, so weigh my opinion with that in mind, but the point isn’t which provider. The point is not to have only one.
Write the plan on one page, tell everyone where it is and test it once. A plan nobody has tried is just a document.
The single point of failure nobody mentions
Look at that list of affected services again. Defender XDR, Microsoft’s security console, was on it. So was the admin center.
A lot of small organizations run their mail, their files, their chat, their AI assistant and their security monitoring on the same platform from the same vendor. That’s often a sensible choice. It’s simpler, cheaper and easier to secure than a patchwork. But it means one bad day takes out the tools you’d use to tell whether the bad day is also a security problem.
You don’t have to fix that by buying more tools. You fix it by knowing which things depend on the same platform, and having a plan for the hour when they all go down together. I wrote about the bigger version of this in your security stack is not your security strategy.
After it’s over
When the post-incident report comes out, read it. Then do three short things:
- Check your contract. Microsoft’s service level agreement offers credits for downtime, but you generally have to claim them. Ask your reseller or your licensing admin.
- Write down what happened inside your company. Who couldn’t do what, which deadlines slipped, what people improvised. That list is your real impact report.
- Run a 30-minute tabletop. Walk your leadership team through “email and Teams are gone for six hours, starting now.” You’ll find the gaps fast.
For defense contractors there’s one more check. DFARS 252.204-7012 requires you to report a cyber incident affecting covered defense information within 72 hours. That report goes through DoD’s DIBNet portal, not your email, but the people who need to coordinate it are probably on Teams. Make sure your incident response plan says how they’d reach each other if Teams is down too.
Leadership is mostly preparation
Thursday wasn’t a test of anyone’s technology. Nobody outside Microsoft could fix it. It was a test of whether leaders had thought about the hour when the tools disappear, and whether their people knew what to do without being told.
The good news is that this is one of the cheapest risks to reduce. One page, one call tree, one group text and one practice run. Do it this week while everyone still remembers Thursday.
Where to go next
- If you handle DoD information, your incident response plan is part of your assessment. The CMMC readiness quiz takes about five minutes.
- Our CMMC guide for small contractors walks through the levels from the beginning.
- Need an incident response plan and policies you can actually use? Start with our CMMC packets.
Discover more from CyberCloudAI.tech
Subscribe to get the latest posts sent to your email.