From early Thursday evening, 12 November 2026, many companies were unable to reach their email in Microsoft 365 for roughly a day. Our email kept working. Not thanks to some miracle program, but because our mailboxes are not in Microsoft 365 and our email client Veredar talks directly to our mail server. Here is what happened, what Veredar can really do, what it cannot, and how to prepare a plan B.
What happened
On Thursday 12 November, early in the evening our time, Microsoft began dealing with a problem connecting to mailboxes in Exchange Online. Exchange Online is the service that runs email in Microsoft 365. The Microsoft 365 Status account first wrote that it was investigating reports of problems connecting to mailboxes “via multiple connection methods” in the Americas. Details were available in the Microsoft 365 admin center under reference EX1550543. Later Microsoft warned that the problem could affect users in any region. It tracked the impact on other Microsoft 365 services under reference MO1550821.
According to the incident record, users could not connect to their mailbox via:
- Outlook on the web,
- MAPI, which Outlook for Windows uses to connect,
- Exchange ActiveSync, used mainly by the built-in mail apps on mobile phones,
- REST API.
Parts of Teams, SharePoint, Copilot and other services that work with email were also affected. Microsoft explicitly stated that the parts of its services that do not depend on Exchange Online kept working.
How long it lasted. According to the preliminary summary in record EX1550543, the impact lasted from 12 Nov 16:20 UTC to 13 Nov 16:10 UTC. In our time that is roughly from 17:20 on Thursday to 17:10 on Friday, i.e. about 24 hours. As early as Friday morning, Microsoft extended the delivery window for delayed messages to 48 hours so that they would not be lost. The full end came later, though: only on Saturday 14 November at 7:09 our time did Microsoft announce that it had resolved the cause and restored service to the remaining customers and regions as well. For some, the problems therefore lasted almost 38 hours.
The cause. According to Microsoft’s preliminary explanation, the systems that establish connections to the Exchange Online service failed. Traffic shifts during network maintenance exposed a latent bug. Requests piled up until the servers ran out of capacity. The bug had entered the system during an earlier expansion of network capacity. Microsoft stated that this was neither a cyberattack nor a failure of AI agents, and apologised: “We are deeply sorry for the disruption this incident has caused our customers and partners.” At the time of writing we had not found a final incident report.
The media reported on affected government offices, hospitals, schools and stock-exchange traders who went back to using the phone. Downdetector recorded more than 130,000 reports about Outlook on Thursday afternoon US time. Microsoft did not publish the total number of people affected.
It is not about who is better
Microsoft 365 is a reliable service that a large part of the business world runs on. Any provider can have an outage, even the biggest one. So the lesson is not “run away from Microsoft”, but rather: when your whole company stands on one supplier, their outage is your outage too. If email, calendar, chat, documents and sign-in are all with one supplier, a problem in a single one of its services can stop an entire day.
Why Veredar kept working
Veredar is our email client. It is in open beta for Mac, Windows and Android; iPhone and iPad are still in testing via TestFlight. One important caveat right at the start: Veredar cannot sign in to Microsoft 365 today (explanation below). So during the outage it would not have helped companies that have their mailboxes in Microsoft 365. It helped us because of how it is built and where we keep our email:
- Email goes directly between the device and the mail server you choose, via the standard IMAP and SMTP protocols. It does not pass through any Veredar server or through a Microsoft service. Only an encrypted profile with settings goes through our server, and you do not need it for reading and sending email.
- Our email does not run on Microsoft 365. So during the outage Veredar connected to the same servers as always and received and sent email without interruption. The Exchange Online outage had no way of touching it.
The other side needs saying too. A program alone will not save your email. If the mailbox had been in Exchange Online, Veredar would most likely not have received new messages during the outage either. Microsoft did not list IMAP among the affected connection methods, but the problem was in the systems that accept connections to mailboxes, not in the program on the user’s side. And as mentioned, Veredar cannot even sign in to Microsoft 365 today. Microsoft has turned off username-and-password sign-in for IMAP in Exchange Online and requires modern authentication (OAuth). So far Veredar supports that only for Google, via Modulo. So the independence did not come from a “better Outlook”. It came from where the email is stored and from the client not being tied to one supplier.
What Veredar does when a server is down
Even with ordinary outages at smaller providers, it is worth knowing what the program can handle without a connection:
- The message list is on the device. Message headers (sender, subject, date, flags) are stored in a local database. The list appears immediately, even without a connection to the server.
- On-device search. On-device AI search (Apple Intelligence or Ollama, if you have them turned on) searches the local copy across accounts, including the text of messages that have already been downloaded. However, the results only show what is already on the device.
- One failing account does not stop the others. If one account fails to connect while syncing the inbox, the other accounts load normally.
- Send Later does not give up right away. If a scheduled message cannot be sent (the server rejects it or does not respond), Veredar retries every five minutes, at most six times in total, while it is running on the device. The error stays visible in the Scheduled folder.
- Signatures and rules are in the Veredar profile, not with the email provider. Veredar keeps contacts as a local copy (from Modulo, Google or CardDAV), so you can see them even without a connection.
- A second route via Modulo. Anyone who uses Modulo also sees the mailboxes from Modulo’s email module in Veredar. From those, they can reply and write a new message via Modulo even without their own IMAP account.
We will not hide one limitation: Veredar loads the full message body from the server when you open a message. So without a connection you can see the list and search, but you cannot open the message.
Plan B for business email: six steps
- Know where your email is and where to check its status. Write down your provider’s status page and, for Microsoft 365, also the path to service health in the admin center. During an outage, the incident number (for example EX1550543) is the first thing the IT department needs.
- Have a backup address with another provider. One mailbox for support or sales on a different domain or with a different hosting provider is enough. Agree in advance who monitors it and when you announce it to customers.
- Use a program that is not tied to one supplier. A client built on open protocols (IMAP, SMTP) can handle mailboxes with several providers at once. If one is down, the others keep working.
- Keep the contacts and phone numbers of key people outside email too. When neither email nor chat works, you need to know whom to call.
- Accept customer requests through more than one channel. For example by email, a form on your website and by phone. If one channel fails, the others keep going.
- After the outage, check what arrived late. This time Microsoft extended delivery to 48 hours. So messages could keep arriving long after service was restored. Go through bounce messages and delayed orders.
Where Modulo fits in
In Modulo we designed email so that it does not depend on a single route. The email module can receive email via Cloudflare Email Routing and send it via Resend or via the company’s SMTP mailbox. It fetches external IMAP mailboxes every five minutes and keeps its own copy of the messages, including attachments. Helpdesk creates tickets from received emails. But even Modulo cannot deliver messages that the source server is not handing out at that moment. Here too, independence starts with where you keep your mailbox.
Summary
The outage on 12 and 13 November did not show that Microsoft 365 is a bad choice. It showed how costly it is to have everything in one place without a plan B. Our email kept working because it is not with a single supplier and Veredar talks directly to a server that we choose. If you like, we will help you look at where your email is and what your plan B looks like.
Sources
Primary (Microsoft):
- Microsoft 365 Status on X, 12 Nov 17:54 UTC, first report on EX1550543 (Americas): https://x.com/MSFT365Status/status/2120938282204688468
- Microsoft 365 Status on X, 12 Nov 18:07 UTC, impact in any region: https://x.com/MSFT365Status/status/2120941584220827979
- Microsoft 365 Status on X, 12 Nov 22:09 UTC, overloaded process on the servers: https://x.com/MSFT365Status/status/2121002505911857577
- Microsoft 365 Status on X, 13 Nov 09:24 UTC, delivery extended to 48 hours: https://x.com/MSFT365Status/status/2121172248241348867
- Microsoft 365 Status on X, 13 Nov 18:05 UTC, apology: https://x.com/MSFT365Status/status/2121303408313303423
- Microsoft 365 Status on X, 14 Nov 06:09 UTC, service restored to the remaining customers and regions: https://x.com/MSFT365Status/status/2121485746334777593
- Incident record EX1550543 from Microsoft 365 service health (public copy; the Microsoft 365 admin center requires sign-in): https://www.cloudadminhub.com/microsoft-365-service-health/ex1550543/
- Microsoft Learn, Deprecation of Basic authentication in Exchange Online: https://learn.microsoft.com/en-us/exchange/clients-and-mobile-in-exchange-online/deprecation-of-basic-authentication-exchange-online
Secondary (media):
- CNBC, 12 Nov 2026, updated 14 Nov 2026: https://www.cnbc.com/2026/11/12/microsoft-outage-outlook-aws-google.html
- SecurityWeek, 12 Nov 2026, updated 13 Nov 2026: https://www.securityweek.com/microsoft-365-outage-disrupts-outlook-exchange-online-and-teams/
- TechRepublic, Microsoft Outlook Down: Over 130,000 Outage Reports: https://www.techrepublic.com/article/news-microsoft-outlook-outage-130000-reports-2026/
- Wikipedia, 2026 Microsoft outage (overview with links to the BBC, The Guardian, Bloomberg and others): https://en.wikipedia.org/wiki/2026_Microsoft_outage
Veredar and Modulo:
- Veredar source code, version 0.20.4 (private repository): core
crates/veredar-core, applicationapps/desktop - Modulo changelog:
apps/admin/config/changelog.ts(versions 2.52.0, 2.110.0, 2.125.0, 2.130.0, 3.4.0)
Verified as of 15 Nov 2026.






