Microsoft 365 Outage Shows Why Records Need Independence
On Thursday 12 November, a Microsoft 365 outage cut users around the world off from their Exchange Online mailboxes and took a long list of dependent services with it, including parts of Microsoft Purview. As I write this the following day, my own Outlook and Teams are still coming and going. Every cloud provider has days like this. For compliance teams, the useful question is what else fails at the same time, and whether your records and your oversight are on that list.
Key takeaways
- A single Microsoft 365 incident disrupted email and compliance tools for more than one day, with no confirmed root cause.
- Supervision, audit search, and eDiscovery tools can go down with the platforms they oversee.
- Recordkeeping obligations don’t pause when the primary channel moves conversations to whatever channel is at hand.
- EU and UK regulators have put AWS, Google Cloud, and Microsoft under direct oversight, but accountability for resilience stays with your firm.
- Resilience for communications records means the platform, the record, and the supervision of it don’t share a single point of failure.
What happened during the Microsoft 365 outage
Microsoft’s Exchange Online incident started at 16:17 UK time on 12 November. About two hours later, Microsoft opened a wider incident for the services that depend on Exchange Online and said it may affect users in any region. Downdetector reports for Outlook peaked at around 130,000 on Thursday afternoon.
The affected list went well beyond email: Teams, SharePoint, Copilot, Defender XDR, Intune, and several Purview features. Microsoft said it wasn’t aware of any malicious activity behind the incident.
Still recovering the next day
Recovery has been slow and uneven. Overnight, Microsoft said a process on the affected servers was becoming overwhelmed, and began rolling back a network configuration change whose timing matched the start of the problem.
On Friday morning, Microsoft reported better stability after applying an authentication and routing patch and began rolling it out region by region. It also extended its email timeout window to 48 hours to protect delayed messages.
By midday UK time, it described recovery in Europe as encouraging, but it hadn’t declared the incident resolved or confirmed a root cause. Microsoft has said it will publish a full report once the incident is mitigated.
Every provider has a bad day
It would be easy to blame Microsoft, but that misses the point. AWS and Google have had major outages too. These things happen, and they’ll happen again. Spreading services across regions makes a long outage less likely, but this week showed it doesn’t rule one out. The problem began in the Americas and spread.
If your resilience plan rests on a provider never failing, it rests on a promise nobody can make. A more useful test is to ask, for each critical function, what else stops working with it. If I were back working in a bank this morning, that’s the question I’d be asking.
When records and oversight go down with the platform
Consider what many firms couldn’t access on Thursday. Email, which still carries much of client communication, became unreachable. Microsoft’s incident reports confirm that Purview Communication Compliance, Data Loss Prevention, data classification, and auto-labelling were affected, and warned that Purview portal pages might not load. They also warned that audit searches and eDiscovery tasks, including search and export, may be impacted.
For firms that rely on that stack, some of the tools they use to supervise communications, investigate issues, and retrieve records were affected too.
This has happened before. When Microsoft 365 went down for nearly 24 hours in January 2026, Purview and Defender went with it. If the platform that carries your communications also holds the record of them and runs the supervision of them, one incident can take out all three.
Conversations don’t stop when email does
Front-office staff don’t stop working because a mailbox won’t load. Some will move to a recorded line, or to whatever part of Teams is still working. Others will reach for WhatsApp, a text, or a personal email account, because the client is waiting and the trade still needs doing.
It’s understandable. But if those communications aren’t captured, the firm has created a compliance problem while working around an IT problem. That’s off-channel communication at the moment it’s least likely to be noticed. Business continuity plans usually cover how trading systems fail over. Far fewer say which channel staff should use when email is down, or whether that channel is captured. The recordkeeping requirements in the FCA’s COBS 11.8 and the SEC’s Rule 17a-4 don’t pause for a supplier’s outage.
What regulators now expect
Over the past year, regulators in the EU and the UK have brought the largest cloud providers inside the regulatory perimeter. In November 2025, the European Supervisory Authorities designated 19 critical ICT third-party providers under DORA, including AWS, Google Cloud and Microsoft. In July 2026, HM Treasury designated AWS, Google Cloud, Microsoft, and Oracle as critical third parties to the UK financial sector, overseen by the Bank of England, the PRA, and the FCA.
Oversight of providers doesn’t shift responsibility away from firms. Microsoft’s own DORA guidance says financial entities remain accountable for their obligations. DORA asks firms to manage ICT concentration risk, and Article 12 requires that when you restore backup data on your own systems, those systems be physically and logically segregated from the source.
In the UK, firms have had to be able to stay within impact tolerances for their important business services since 31 March 2025. Starting in 2027, they’ll also report more detail about the third parties they rely on.
None of these rules prescribes an architecture. Taken together, they make it hard to defend a setup where your records become unreachable whenever the source platform does.
What independent recordkeeping looks like
Resilience for communications records rests on independence. The system that carries a conversation and the system that keeps and supervises the record of it should be able to fail separately.
That works in both directions. In an independent setup, if archive provider has an incident, conversations carry on and capture catches up once the archive is back. If the communications platform has an incident, the record of everything before the outage stays searchable, supervision continues on the channels still running, and you can still respond to a regulator’s request or a legal hold that lands that day.
Hold your archive provider to the same standard: ask how it contains its own failures and how it would prove that to you.
Signs it’s time to reassess your resilience plan
These signs suggest your records may share a single point of failure with your communications platform:
- Your communications, archive, and surveillance all sit with the same provider.
- Your business continuity plan doesn’t name an approved fallback channel for when email or chat is unavailable.
- Your compliance team has never tried to search records or run surveillance while the primary platform is down.
- Your DORA register of information, or your UK third-party mapping, lists the platform but not the compliance functions that depend on it.
How to reduce concentration risk in your records
Start by mapping failure domains — the points where one outage takes several functions down at once. For each communications channel, write down where the record is kept and where it’s supervised. If the answer is the same provider every time, document that concentration and the reasoning behind it, so you’re ready to explain it to a regulator.
Then write the outage fallback into policy: name the approved channel, confirm it’s captured, and tell staff before they need it.
Finally, test the scenario rather than only documenting it. Confirm your compliance team search last month’s records and run surveillance while the primary platform is unavailable. Running the test is the only way to know.
Frequently asked questions
It hasn’t said whether any mail was lost. Microsoft said email delivery might be delayed, and on Friday it extended its email timeout window to 48 hours to protect delayed messages. Check message trace for your tenant, and read Microsoft’s post-incident report when it’s published.
Not in those words. DORA requires firms to manage ICT concentration risk and to restore backups on systems segregated from the source. A separate archive is one practical way to support both.
It makes a long outage less likely, but it doesn’t remove the risk. This incident began in the Americas and spread to other regions.
Your firm. Designation gives regulators oversight of the provider. It doesn’t transfer your obligations to it.
Share this post!
Smarsh Blog
Our internal subject matter experts and our network of external industry experts are featured with insights into the technology and industry trends that affect your electronic communications compliance initiatives. Sign up to benefit from their deep understanding, tips and best practices regarding how your company can manage compliance risk while unlocking the business value of your communications data.
Ready to enable compliant productivity?
Join the 6,500+ customers using Smarsh to drive their business forward.




Subscribe to the Smarsh Blog Digest
Subscribe to receive a monthly digest of articles exploring regulatory updates, news, trends and best practices in electronic communications capture and archiving.
Smarsh handles information you submit to Smarsh in accordance with its Privacy Policy. By clicking "submit", you consent to Smarsh processing your information and storing it in accordance with the Privacy Policy and agree to receive communications from Smarsh and its third-party partners regarding products and services that may be of interest to you. You may withdraw your consent at any time by emailing [email protected].
FOLLOW US