Call Check my tenant

Microsoft 365 Outage, November 2026: What Microsoft’s Incident Report Says Went Wrong

Outlook stopped connecting for most Australian businesses at 3:20 am on Friday 13 November and stayed down until early Saturday. Microsoft has now given admins its preliminary post-incident report. We read the whole thing. This is what it says broke and why recovery took most of a day, followed by the things worth doing before the next outage.

Quick answer

Microsoft 365 went down globally from 16:20 UTC on 12 November 2026 to 16:10 UTC on 13 November, which in Sydney was 3:20 am Friday to 3:10 am Saturday. Microsoft’s preliminary report blames a latent bug in the network gateway servers in front of Exchange Online, set off by extra load and missing capacity. It was not a cyber attack. Delayed email was held and delivered late.

Key takeaways

  • A whole Australian business day: Exchange Online was unreachable for most users from 3:20 am Friday 13 November until 3:10 am Saturday (AEDT). Microsoft tracked it as EX1550543 and MO1550821.
  • The cause was inside Microsoft: a bug introduced in July, in Windows code written only for Microsoft 365’s gateway servers. It surfaced when Teams traffic moved onto the same gateway while about 14% of its servers were already out of service.
  • It spread worldwide in under two hours: more than 80% of gateway servers stopped responding, and the boundaries meant to keep a failure inside one region didn’t hold.
  • You may be owed a credit: Microsoft’s Exchange Online SLA credits 25% to 100% of the Exchange fee when monthly uptime falls below 99.9%. Credits aren’t automatic, and Microsoft must receive claims for this outage by 31 December 2026.
  • Get ready for the next one: send Service health alerts to an address outside Microsoft 365, write down what staff do when Outlook is down, and treat “outage compensation” emails as phishing.
On this page
  1. What happened
  2. The timeline in Sydney time
  3. What caused it
  4. Why it took almost a day
  5. What Microsoft has changed
  6. What to do now
  7. Claiming an SLA credit
  8. Do you need a backup email system?
  9. Frequently asked questions

What Happened in the November 2026 Microsoft 365 Outage?

For just under 24 hours, most Microsoft 365 users around the world couldn’t open their mailbox. Outlook desktop, Outlook on the web, phones using Exchange ActiveSync and apps connecting through Microsoft’s REST APIs all failed to connect, and the email that did get through arrived late. Microsoft tracked it in Service health as EX1550543 for Exchange Online and MO1550821 for the other services that depend on it.

The damage went well past email, because a lot of Microsoft 365 leans on Exchange in the background. Microsoft’s incident notice lists these among the affected features:

  • Teams: scheduling and editing meetings, calendar views, creating live and virtual events, and creating or deleting private and shared channels. Chats, calls and meetings mostly kept working.
  • Copilot: Copilot Chat failing to load, and Copilot unable to reach email, calendar and files.
  • Defender and Purview: portal pages not loading, and delays or failures in data loss prevention, sensitivity labels, eDiscovery and audit searches.
  • Intune: some device enrolments, check-ins and compliance checks.
  • SharePoint: sharing files with people outside your organisation. Opening files stored in SharePoint kept working.
  • Power Apps and Power Automate: sign-in failures in the mobile apps, and flows that use Microsoft Graph data failing on and off.
  • Everything else on the list: Planner, To Do, Universal Print, parts of the Microsoft 365 admin center, and Windows 365 Cloud PCs that use Purview Customer Key.

Microsoft’s explanation is a preliminary post-incident report (PIR) dated 15 November 2026. It is published to tenant admins inside the Microsoft 365 admin center rather than on a public web page, although Microsoft executive Charles Lamanna posted an abridged version on X. Everything in this article comes from that report and Microsoft’s own Service health updates for the incident, unless we say otherwise.

24h Active impact, less 10 minutes, per Microsoft’s report
80%+ Gateway servers unresponsive at the worst point
14% Gateway capacity already out of service before it started
July When the bug shipped. It ran for over three months without trouble

When Was Microsoft 365 Down in Australia?

Microsoft reports in UTC. Converted to Sydney time (AEDT, UTC+11), the outage covered the whole of Friday 13 November. For Brisbane take an hour off the times below, for Adelaide half an hour, and for Perth three hours.

Microsoft 365 outage timeline in Sydney time, from Microsoft’s preliminary post-incident report
Sydney time (AEDT) What happened
Thu 12 Nov, 12:07 pm Microsoft’s automated failover moves Teams traffic onto the gateway servers shared with Exchange Online. Load climbs through the afternoon and evening, and automation handles isolated server failures.
Fri 13 Nov, 3:03 am Microsoft’s connectivity monitoring picks up a rise in connection failures.
3:20 am A high-priority alert for Exchange Online fires and EX1550543 goes up on Service health. Microsoft counts this as the start of impact.
5:27 am MO1550821 is posted for Teams, Copilot, Defender and the other services affected.
6:00 am Microsoft pauses all deployments and configuration changes.
6:20 am Engineers work out the load isn’t coming from more traffic. Servers are dropping out and pushing their share onto the rest.
10:55 am Servers returned to service are found to be healthy until they receive traffic, then fail again.
11:30 am to 12:00 pm The process that sets up connections on each server is identified as the part being overwhelmed.
6:20 pm An update that reverts the faulty change is applied to part of the fleet.
7:00 pm The update is confirmed stable and rolled out further, starting with regions in or entering business hours.
10:00 pm Recovery is completed across parts of the Asia Pacific infrastructure. Mailbox access and mail processing improve.
Sat 14 Nov, 12:00 am Recovery holds across all regions and queues of delayed email start to shrink.
1:17 am Outlook on the web availability is back to normal worldwide.
3:10 am Microsoft confirms customers are no longer seeing active impact and moves to monitoring and clearing the backlog.
3:15 pm Final customer updates are posted to both incidents.

Microsoft brought regions back in order of their business hours, and Asia Pacific infrastructure was the first it reported as recovered, at about 10 pm on Friday. By then most Australian offices had closed for the weekend, so the working day was gone. Mail sent to you during the outage was queued and delivered later. Microsoft extended its email timeout window to 48 hours while it recovered, so delayed messages wouldn’t bounce back to the people who sent them.

What Caused the Microsoft 365 Outage?

Most Microsoft 365 traffic first passes through a networking gateway: a large fleet of servers that accept your connection and hand it to the right service. Microsoft runs two separate gateway systems so services can move from one to the other when there’s trouble. Exchange Online is too big to move, so it only uses one of them. Microsoft’s report says three things lined up on that gateway.

1. A bug that had been waiting since July

The gateway servers ran a fix Microsoft had written specifically for them. Under certain load it caused lock contention inside an internal Windows process. Work piled up waiting on the same resource until the server ran out of CPU and memory and stopped responding. The change started rolling out in July, went through Microsoft’s staged deployment process, and ran worldwide for more than three months without a problem.

Microsoft’s load testing missed it, and the report gives two reasons. The lab traffic mix was different from production, and the test hardware was less powerful, with fewer processor cores. The contention gets worse on newer machines with more cores, which is what the production servers run.

2. Teams moved onto the same gateway

About 15 hours before the outage, at 12:07 pm Sydney time on Thursday, Microsoft’s automated failover moved Teams off the alternative gateway and onto the primary one that also carries Exchange Online. Microsoft describes this kind of failover as routine. It added load and changed the mix of traffic the servers were handling.

3. Capacity was already down

Nearly 14% of the gateway’s servers were out of service because of an unrelated problem the day before, and were due back through normal maintenance. On its own that didn’t matter, because the gateway had enough spare capacity.

Put together, more traffic on fewer servers was enough to set off the July bug. It hit during peak business hours in the United States.

Why one bug became a global outage

When a server stops responding, Microsoft’s traffic management sends its users to healthy servers. That normally keeps everyone online. This time the extra traffic pushed the next servers into the same bug, so the failure moved from server to server. In under two hours it had reached the whole fleet, and at the worst point more than 80% of gateway servers were unresponsive.

The gateway is split into thousands of “capacity units” grouped into regions, with boundaries between them that are meant to stop one region’s problem reaching another. Microsoft says a separate issue meant those boundaries didn’t contain the failure this time. The preliminary report doesn’t say what that issue was.

Does this bug affect your own Windows servers or PCs?

No. Microsoft says neither the defect nor the fix that introduced it applies to Windows outside its Microsoft 365 gateway servers. There’s nothing to patch on your own computers or servers because of this outage.

Why Did It Take Almost a Day to Fix?

Microsoft’s report gives several reasons recovery was slow.

  • Servers fell over again as soon as they came back. Each server returned to service was hit straight away by the queue of waiting connections and stopped responding again.
  • The automated recovery tools weren’t built for it. They are designed to deal with servers failing here and there, and this time most of the fleet was down at once.
  • The usual ways of undoing a change didn’t work. Microsoft normally reverses a bad change by rolling it back, pausing a staged rollout or switching off a feature flag. Those work on recent changes. This one had been in place since July.
  • Finding the cause took time. Early on, the evidence pointed at failing health checks and overloaded servers. Microsoft paused all deployments at 6 am Friday, Sydney time, but it was about midday before engineers traced the problem to the process that sets up connections.

The fix was an update that returned the connection-handling code to its last known good state. Engineers took large groups of servers out of service, often half a region or more at a time, applied the update, restarted them and brought them back gradually so they weren’t swamped again. The first servers got the update at 6:20 pm on Friday, Sydney time.

Why Teams kept working while Outlook didn’t

Teams, OneDrive, SharePoint and the Office apps can move to Microsoft’s other gateway, and during the outage they did. That’s why Teams chats and meetings worked for most of the day and files stayed reachable. Exchange Online can’t use the other gateway because of its size, so it had to wait for the main gateway to be repaired. Teams features that rely on Exchange, such as calendars, failed with it.

What Has Microsoft Changed Since the Outage?

The preliminary report lists five repair items as completed:

  • The update that put the connection-handling code back to its previous state.
  • A longer email delivery window, so delayed messages weren’t returned as undeliverable and everything queued during the outage was processed.
  • Extra gateway capacity, added to help recovery.
  • New monitoring for traffic redirection patterns and for the health of connection handling.
  • Guardrails on how traffic is redistributed, so one server’s failure can’t cascade to the next.

Microsoft says the longer-term fixes from its full root cause analysis will be in the final post-incident report, and its repair item table dates the comprehensive list 20 November 2026. The preliminary report leaves some questions open, such as why the regional boundaries failed and how load testing will change. We’ll update this article when the final report is out.

Want someone to look over your Microsoft 365 setup?

An engineer goes through your Microsoft 365 tenant, licences and security settings, then tells you what they found and what they would fix first. There’s no charge.

Get your tenant checked

What Should Australian Businesses Do After the Outage?

Nothing in your own tenant caused this, and there’s nothing to repair on your side. The useful work is getting ready for the next outage, whether it comes from Microsoft or somewhere else.

  1. Make sure the right people can see Service health. In the Microsoft 365 admin center, Health then Service health shows incidents affecting your tenant, with updates as Microsoft works through them. Global admins can see it, and so can anyone with the Service Support admin or Helpdesk admin role. If you can’t sign in at all, Microsoft’s public service status page and the @MSFT365Status account on X carry updates for large events (Microsoft’s guide to Service health).
  2. Send outage alerts somewhere that works when email doesn’t. Service health can email each admin about new incidents, at up to two addresses. If both are Microsoft 365 mailboxes, the alert lands in the system that’s down. Use one address outside your tenant, or install the Microsoft 365 Admin app on a phone for push notifications.
  3. Write a one-page plan for a day without email. Name who checks Service health and tells everyone else. Decide how staff hear about it: Teams worked for most of this outage, and a group text never hurts. Keep a short message ready for your website or phone greeting, and keep the numbers for key clients and suppliers somewhere other than Outlook.
  4. List what quietly depends on Exchange. Shared mailboxes for orders and enquiries, scan-to-email on printers, invoices or alerts that software sends through Microsoft 365, meeting room bookings and approvals in Power Automate all stopped on Friday. Knowing the list in advance tells you what needs a manual workaround.
  5. Don’t troubleshoot a Microsoft outage on your own machines. Rebuilding Outlook profiles, reinstalling Office or resetting passwords won’t help when the fault is in Microsoft’s data centres, and it leaves a mess to clean up afterwards. Check Service health first. During this outage, Microsoft’s advice to affected Power Apps users was to stay signed in and not to clear app data or reinstall.
  6. Expect outage-themed phishing. Large outages are followed by fake emails offering compensation, or asking staff to sign in again to “restore” their mailbox. Real SLA credits are claimed by your administrator or licensing provider through Microsoft support, and staff never need to click anything to get one. Our article on device code phishing in Microsoft 365 covers one sign-in trick worth blocking while you’re at it.

Can You Claim an SLA Credit for the Outage?

Possibly. Microsoft’s Service Level Agreement for Online Services sets a 99.9% monthly uptime target for Exchange Online. For that SLA, downtime means users can’t send or receive email with Outlook on the web, and it’s counted in user-minutes: each affected user’s minutes of downtime, added up across the month.

Exchange Online service credits by monthly uptime, from Microsoft’s Service Level Agreement for Online Services
Exchange Online monthly uptime Service credit
Below 99.9% 25%
Below 99% 50%
Below 95% 100%

November has 43,200 minutes. A user who couldn’t use Outlook on the web for the full 23 hours and 50 minutes had about 96.7% uptime for the month, which sits in the 50% band. Users affected for part of the day count only for the time they were affected, so your tenant’s figure depends on how many of your people were hit and for how long.

Before you claim, know how it works:

  • Credits aren’t automatic. Someone has to lodge a claim with Microsoft support that describes the incident, when and for how long you were affected, how many users were affected and where, and the errors they saw.
  • There’s a deadline. Microsoft must receive the claim by the end of the month after the incident. For this outage, that’s 31 December 2026.
  • It’s a percentage of the affected service’s fee. If you buy a suite such as Microsoft 365 Business Premium, Microsoft credits only the affected service within it.
  • Providers lodge claims for their customers. If you buy licences through a Cloud Solution Provider (CSP), Microsoft only accepts the claim from that provider. They need your tenant ID, the incident ID (EX1550543), and an email from someone at your business, sent from your own domain, asking for the credit. Providers work to a deadline of one month after the end of the billing month, so ask early.
  • A credit is the only remedy. Microsoft’s SLA says credits are the sole remedy for an outage, and they don’t cover lost sales, overtime or any other cost to your business.

Does a Small Business Need a Backup Email System?

For most small businesses, a second email platform running all year costs more than an outage like this one. What helps more is deciding in advance how you’ll work through a day without email, and knowing what the services you already pay for do in that situation.

Mail filtering services

If your mail passes through a filtering service before it reaches Microsoft, ask your provider what happens when Microsoft can’t accept mail. Some services hold incoming mail and give users a web page to read it until the mailbox is back. On Friday, Microsoft held delayed mail itself and delivered it once the service recovered, so the bigger loss for most businesses was a day of not being able to read or send.

Microsoft 365 backup

A Microsoft 365 backup copies your mailboxes, OneDrive, SharePoint and Teams data to storage outside Microsoft. Every AyeTech managed plan includes one, on Acronis, stored in Sydney. It’s there to get data back after something is deleted or encrypted. It doesn’t keep Outlook connected when Microsoft’s gateway is down, so it wasn’t the answer to Friday. Our backup and recovery page explains what it does cover.

Having everything with one provider

Friday showed both sides of running everything on Microsoft 365. Email, calendars and Copilot went down together. Teams and SharePoint recovered early, though, so staff still had chat and files for most of the day. Moving platforms because of one outage rarely makes sense, since every cloud email service has outages. If you’re weighing it up for other reasons, our Microsoft 365 vs Google Workspace comparison covers the trade-offs.

Get your Microsoft 365 tenant checked

An engineer goes through your Microsoft 365 tenant, licences and security settings, then tells you what they found and what they would fix first. There’s no charge.

Get your tenant checked Microsoft 365 support and security

Or call 02 9188 8000, Monday to Friday, 8am to 6pm. Forms get a reply within one business hour.

Frequently Asked Questions

When was Microsoft 365 down in November 2026?

Microsoft reports active impact from 16:20 UTC on Thursday 12 November 2026 to 16:10 UTC on Friday 13 November. In Sydney and Melbourne (AEDT) that is 3:20 am Friday 13 November to 3:10 am Saturday 14 November, so the outage covered the whole Australian business day on Friday. Microsoft posted its final customer updates at 3:15 pm AEDT on Saturday, after a period of monitoring.

What caused the Microsoft 365 outage on 12 and 13 November 2026?

Microsoft’s preliminary report names three factors. A bug introduced in July, in Windows code written only for Microsoft 365’s networking gateway servers, caused lock contention that exhausted server CPU and memory under certain load. About 15 hours earlier, Teams traffic had failed over onto the gateway that also serves Exchange Online. And nearly 14% of gateway servers were already out of service from an unrelated issue. Together they set off the bug, and the failure cascaded across the fleet.

Was the Microsoft 365 outage a cyber attack?

No. During the incident Microsoft said it had no evidence of malicious activity and described it as a technical issue unrelated to cyber attacks or AI agents. Its preliminary post-incident report puts the cause down to a latent software defect in its own gateway servers. There is nothing to clean up or reset in your own Microsoft 365 tenant because of this outage.

Was any email lost during the Microsoft 365 outage?

Microsoft says delayed email was protected. During recovery it extended the email timeout window to 48 hours so queued messages wouldn’t be returned to senders as undeliverable, and its report lists processing everything queued during the event as a completed repair item. Mail arrived late, sometimes many hours late. If a customer says a message they sent on 13 November bounced, ask them to send it again.

Which Microsoft 365 services were affected by the outage?

Exchange Online was hit hardest: Outlook desktop, Outlook on the web, ActiveSync on phones and REST API connections all failed, and mail was delayed. Features of other services that rely on Exchange also failed, including Teams calendars and meeting scheduling, Copilot, the Defender and Purview portals, Intune enrolment and compliance checks, Planner, To Do, Universal Print, Power Apps sign-in and external sharing in SharePoint. Teams chats and meetings largely kept working.

Why did Teams keep working when Outlook didn’t?

Microsoft runs two separate networking gateways. Teams, OneDrive, SharePoint and the Office apps can move between them, and during the outage they moved off the failing one, which is why Teams chats and meetings worked for most of the day. Exchange Online is too large to move to the other gateway, so it had to wait for the main gateway to be repaired. Teams features that depend on Exchange, such as calendars, still failed.

Why did the outage last almost 24 hours?

Servers returned to service were immediately swamped by waiting connections and failed again. Microsoft’s automated recovery wasn’t designed for most of the fleet being down at once. And because the faulty change dated from July, the usual rollback, staged rollout and feature flag tools couldn’t reverse it quickly. Engineers built an update that restored the old code, applied it to large groups of servers, restarted them and brought them back gradually.

Does the Windows bug affect our own Windows servers or PCs?

No. Microsoft says the defect came from a fix prepared specifically for its Microsoft 365 networking gateway servers, and that neither the defect nor the fix applies to broader Windows deployments. There is no patch to install on your own computers or servers because of this outage.

Can we get a service credit from Microsoft for the outage?

Possibly. Microsoft’s SLA for Exchange Online promises 99.9% monthly uptime, counted in user-minutes, with credits of 25% of the Exchange Online fee below 99.9%, 50% below 99% and 100% below 95%. A user who couldn’t use Outlook on the web for the whole outage had about 96.7% uptime for November. Credits aren’t automatic, apply only to the affected service, and are Microsoft’s only remedy, so they don’t cover lost revenue. See how to claim.

What is the deadline to claim an SLA credit for this outage?

Microsoft’s SLA says claims for Microsoft 365 services must reach Microsoft by the end of the month after the month the incident happened. For the 12 to 13 November 2026 outage, that is 31 December 2026. If you buy through a Cloud Solution Provider, the provider has to lodge the claim within one month of the end of the billing month, so give them what they need well before Christmas.

We buy Microsoft 365 through a provider. Who lodges the SLA claim?

If you buy through a Cloud Solution Provider (CSP), only that provider can lodge the claim. Microsoft asks for your tenant ID, the incident ID (EX1550543 for the Exchange Online outage), details of when and how your users were affected, and an email from your business, sent from your own domain, asking for the credit. If you buy direct from Microsoft, your administrator lodges the claim with Microsoft support.

Where do I find Microsoft’s post-incident report?

Sign in to the Microsoft 365 admin center, go to Health, then Service health, then Issue history, and open MO1550821 or EX1550543. An incident with a report shows the status Post-incident report published and a Download post-incident report link. You need an admin role that can view Service health. Issue history only goes back 30 days, so download a copy before mid-December if you want to keep it.

When will Microsoft publish the final post-incident report?

Microsoft’s preliminary report says a final report will follow within five business days of full resolution, and its repair item table gives 20 November 2026 for the comprehensive list of actions. The final report replaces the preliminary one. We’ll update this article when it is published.

How do I check whether Microsoft 365 is down?

Admins should check Service health in the Microsoft 365 admin center first, because it shows incidents affecting your own tenant. If you can’t sign in, Microsoft’s public status page at status.cloud.microsoft and the @MSFT365Status account on X post updates for large incidents. The Microsoft 365 Admin app on a phone can push Service health notifications, which helps when email itself is down.

Would a Microsoft 365 backup have helped during the outage?

Not with the outage itself. A third-party backup copies your mailboxes, OneDrive, SharePoint and Teams data to storage outside Microsoft, so you can restore it after something is deleted or encrypted. It doesn’t keep Outlook connected to Exchange Online when Microsoft’s gateway is down. Backup is still worth having for the problems it does solve, because Microsoft’s own retention of deleted items is limited.

Should a small business have a second email system for outages?

Usually not. A second platform running all year costs more than an occasional outage. A written plan for a day without email, Service health alerts sent somewhere outside Microsoft 365, and knowing what your mail filtering service does when Microsoft can’t accept mail will cover most of the risk. If a day without email has a real dollar cost for you, for example because orders arrive by email, talk to your IT provider about continuity options.

What should staff do when Outlook stops connecting?

First find out whether it’s just them or everyone. If it’s widespread, check Service health or status.cloud.microsoft before changing anything. During a Microsoft outage, rebuilding Outlook profiles, reinstalling Office or resetting passwords won’t help and creates clean-up work later. Use Teams or phones for urgent internal messages, and let customers know by phone or on your website if email will be slow.

Should we move off Microsoft 365 after this outage?

One outage is rarely a good reason to move platforms. Every cloud email service has outages, and a migration costs time and money and disrupts staff. The better question is whether your business can keep working for a day without email, and that comes down to planning more than platform. If you’re comparing Microsoft 365 with Google Workspace for other reasons, our comparison covers the trade-offs.

Are there scams linked to the Microsoft outage?

Expect them. Large outages are a common hook for phishing emails offering compensation or asking users to sign in again to restore their mailbox. SLA credits are claimed by an administrator or licensing provider through Microsoft support, and staff never need to click a link to get one, so treat any email like that as a scam and report it to your IT provider.