Outage Phishing: A Fake Service Notice That Cites the Real Incident
Microsoft Threat Intelligence, 14 Nov 2026: phishing emails cite the real incident number of the 12 Nov Microsoft 365 outage and threaten to delete held mail.

AegisAI Threat Watch · Phishing
Outage Phishing: A Fake Service Notice That Cites the Real Incident
Microsoft Threat Intelligence, 14 November 2026: emails that impersonate Microsoft 365 service health notices cite the real outage incident EX1550543 and say held mail will be deleted within 24 hours unless the recipient clicks.
- What
- Emails that impersonate Microsoft 365 service health notices cite the real outage incident EX1550543 and say held mail will be deleted within 24 hours.
- Who it hits
- Organizations across multiple US sectors, per Microsoft Threat Intelligence. It gives no count and no actor.
- Why filters miss it
- The incident number is real and the links start at a sending service's tracking address. Microsoft does not say how many emails were delivered.
- Do today
- Hunt for subjects that start with the word Outage, check who clicked, and tell staff where real incident status lives.
Reported by Microsoft Threat Intelligence, 14 November 2026.
Attackers sent fake Microsoft 365 service health emails that cited the real outage incident EX1550543 and claimed held mail would be deleted within 24 hours unless the recipient clicked.
Credential phishing is a message that tries to get a person to type a sign-in into a page the attacker controls. Click tracking is a feature of bulk email services that rewrites each link so the service can count clicks, then sends the reader on to the real destination. EX1550543 is Microsoft's incident number for the 12 November 2026 Exchange Online outage.
Analysis based on Microsoft Threat Intelligence reporting, a post on X dated 14 November 2026. Microsoft does not name a threat actor, and it says credential collection and malware delivery have not been confirmed. The outage facts come from Microsoft's own service health entries. A related Threat Watch article covers a help desk impersonation that starts with a flood of email.
Source: Microsoft Threat Intelligence post on X, 14 November 2026; Microsoft 365 service health text for EX1550543, read through a University of Pennsylvania mirror, entry of 13 November 2026, 09:55 UTC.
Key judgments
- The lure borrowed a real incident number. Microsoft says the messages reference the legitimate incident EX1550543, so a check that stops at the number finds a real incident. What is false is what the email says about the incident.
- The timing made the claim believable. Microsoft detected some of the emails on 13 November, while its own entries for the incident still described delayed mail queues and a backlog. Staff had a real reason to expect delayed messages, which is the claim the lure makes.
- The first hop is a tracking link. Each message carries a recipient-specific SendGrid click-tracking link that redirects to reallifeaccounting[.]com. The first address in the link belongs to the sending service, not to the page the reader lands on.
- The payload is unconfirmed. Microsoft calls the activity "consistent with credential phishing" and says credential collection and malware delivery have not been confirmed.
- Scale and actor are not stated. Microsoft says similar messages targeted organizations across multiple US sectors. It gives no count, no victim and no actor.
Evaluating email security tools? The 2026 Buyer's Guide to Email Security has nine evaluation areas and the questions to ask vendors. Get the guide.
What happened in the Microsoft 365 outage?
Microsoft's first public post about the incident came from its Microsoft 365 Status account on 12 November 2026. It said Microsoft was investigating reports of users having trouble reaching Exchange Online mailboxes through multiple connection methods in the Americas, and pointed to EX1550543 in the Microsoft 365 admin center.
Microsoft's service health text lists Outlook on the web, MAPI, Exchange ActiveSync and the REST API as the connection methods affected. It adds that mail flow between on-premises systems and Exchange Online may have been affected. A second record, MO1550821, covers other Microsoft 365 services, among them Microsoft Graph, Intune, SharePoint Online, Microsoft 365 Copilot and the Microsoft Defender portal. SecurityWeek, reporting on 12 and 13 November, names Teams and Defender XDR in the same spread.
On cause, Microsoft's final status says the incident "was a technical issue not related to cyber or AI / agents". It says the incident was introduced by an issue during an earlier network capacity upgrade that surfaced after a recent network change. Earlier entries said Microsoft had no evidence of malicious activity associated with the incident. Microsoft promised a full post-incident report.
The records do not give one clean timeline. The header fields show a start of 10:05 UTC on 12 November, the description gives 16:20 UTC, and the end time is 16:10 UTC on 13 November, yet entries kept arriving until 04:15 UTC on 14 November, when Microsoft said it was still restoring service to remaining customers and regions. This article states no single duration for that reason.
Sources: Microsoft 365 Status post, 12 November 2026; Microsoft 365 service health text for EX1550543 and MO1550821 (University of Pennsylvania mirror, read 15 November 2026); SecurityWeek, 12 and 13 November 2026.
How did attackers turn the outage into a phishing lure?
Microsoft Threat Intelligence observed an active campaign that impersonates Microsoft 365 service health notifications and cites the legitimate incident EX1550543. Some of the emails Microsoft detected on 13 November falsely claim that messages were held during the outage and will be deleted within 24 hours unless recipients click to restore them. The details Microsoft reports are below.
- Subject: the pattern "Outage - [six-character code]".
- Sender name: an organization-specific prefix followed by "Outage Notification" or "Service Health".
- Sender domains: three, listed in the indicators section.
- Links: recipient-specific SendGrid click-tracking links that redirect to reallifeaccounting[.]com, including the observed path /themelist0.html.
- Targets: organizations across multiple US sectors.
Source: Microsoft Threat Intelligence post on X, 14 November 2026. iTnews, 14 November 2026, reports the same details.
What the recipient sees and what is happening
The table pairs each moment with the action behind it. It is AegisAI's arrangement of Microsoft's description.
| Moment | What the recipient sees | What is happening |
|---|---|---|
| Inbox | A service health notice from a display name that carries the organization's own prefix | The prefix varies by organization; the three sender domains are not Microsoft's |
| Subject | The word Outage, a hyphen and a short code, which reads like an incident reference | The code is six characters long |
| Body | The real incident number and a warning that held messages will be deleted in 24 hours | The number is real; the deletion claim is false |
| Click | A link to restore held messages | A recipient-specific SendGrid tracking link counts the click and redirects |
| After | The destination page | reallifeaccounting[.]com/themelist0.html. Microsoft has not confirmed what the page collects |
Source: Microsoft Threat Intelligence post on X, 14 November 2026. The "What the recipient sees" column is AegisAI's reading of the lure text Microsoft describes.
Why a tracking link hides the destination
Twilio's SendGrid documentation says click tracking rewrites all the links in an email and that the links then point either to SendGrid's servers or to the sender's branded domain. SendGrid can then log each click. This section is AegisAI's analysis: a check that reads only the first address in such a link sees the sending service, not the page behind it, and the real destination appears only when the link is followed.
Microsoft's own advice fits that gap. It points to Safe Links, which provides time-of-click protection "when properly configured". Microsoft does not say SendGrid was compromised, and it does not say whether the sending accounts were created by the attackers or taken over. That question stays open.
Is outage phishing new?
No. After a widespread IT outage in July 2024, the US Cybersecurity and Infrastructure Security Agency said it had observed threat actors taking advantage of the incident for phishing and other malicious activity. It urged organizations to follow instructions only from legitimate sources, and on 26 July it advised blocking malicious domains. The pattern is the same: a real, widely visible incident gives a false message a reason to arrive.
Source: CISA alert on the July 2024 outage, issued 19 July 2024 and updated through 6 August 2024.
How can an organization check whether staff were hit?
Microsoft's advice gives a hunt and a verification step. The table lists them with AegisAI's reading of what each can tell a team.
| Control | What it sees in this campaign | Result |
|---|---|---|
| Defender for Office 365 Threat Explorer | Messages with the Outage subject pattern and the three sender domains | Finds delivered messages |
| Defender XDR advanced hunting (EmailEvents, EmailUrlInfo, UrlClickEvents) | The embedded links and which users clicked them | Shows who reached the redirect |
| Safe Links | The destination at the moment of the click, if the policy is configured | Microsoft: protection "when properly configured" |
| Microsoft 365 admin center | The real incident record under EX1550543 | Confirms the real outage and what Microsoft said about it |
Controls and wording from Microsoft Threat Intelligence, 14 November 2026. The "Result" column is AegisAI's reading, except where it quotes Microsoft.
The incident number, the outage and the delay were all real. The deletion deadline was not.
AegisAI Threat IntelligenceWhat this means for a security team
This section is AegisAI's analysis. The lure needs no new technique. It needs a moment when staff expect an email from the platform, and an outage that sends status updates, delays and backlogs gives it that moment. The risk can outlast the incident, because delayed mail is still expected after service returns.
Three questions settle the exposure: whether staff know where real incident status lives, whether anyone can show which users clicked a link in these messages, and whether a message that cites a real incident number is read for what it asks the recipient to do.
Why standard controls miss it
This section is AegisAI's analysis. Microsoft detected some of these emails and does not say how many reached an inbox. What follows is why a simpler check could pass one. The incident number checks out, the topic matches real events, and the first link points to a well-known sending service. A rule that trusts a real reference, or a link check that stops at the first address, has little to object to. The request is the signal: click to restore held mail, inside a deadline, from a sender that is not Microsoft.
How AegisAI protects against it
AegisAI has not observed this campaign. Microsoft reports it as email, so the approach below applies to messages that arrive in a Microsoft 365 or Google Workspace mailbox. It does not describe how a different channel would be handled.
AegisAI's agents evaluate the message and the destination it points to together, and pull a malicious message out of mailboxes before users see or click it. More on how that works: inbound email security.
When a lure or a compromised session is caught, the response compounds: the message is withheld at classification, copies already delivered are retracted, rewritten links refuse a late visit, the account's outbound and internal mail gets rescanned, and lateral phishing attempts to colleagues are blocked at the same bar rather than treated as a fresh, unrelated message.
Recommendations for defenders
- Run Microsoft's hunt now: search Threat Explorer and advanced hunting for subjects that start with the word Outage, the three sender domains and clicks on the redirect destination.
- Treat each user who clicked as a possible credential exposure until shown otherwise: review the account's sign-ins, revoke its sessions and reset its password. Microsoft has not confirmed credential collection, so this is a precaution.
- Check that the Safe Links policies are configured, as Microsoft advises, so the destination is checked when the link is clicked.
- Tell staff in advance where real incident status lives, the Microsoft 365 admin center and an internal channel that does not depend on email links, and repeat it at the start of the next outage.
- Match on the three sender domains and the redirect destination, not on the sending service as a whole. Microsoft does not say the service itself was compromised.
FAQ
How do attackers use a real Microsoft 365 outage for phishing?
Microsoft Threat Intelligence reported emails that impersonate Microsoft 365 service health notifications and cite the real incident EX1550543. Some say messages held during the outage will be deleted within 24 hours unless the recipient clicks. The links are SendGrid click-tracking links that redirect to a site Microsoft named in its 14 November 2026 post.
Does a real incident number mean the email is real?
No. The incident number in these messages is real, and Microsoft calls the incident legitimate. The attackers copied it, so the number cannot be used to judge the email. The place to confirm an incident is the Microsoft 365 admin center, which is what Microsoft advises.
How can I tell a real Microsoft 365 service health notice from a fake one?
Check the incident in the Microsoft 365 admin center instead of following a link in the email. Treat any message that asks you to click to restore held mail, or sets a short deadline, as suspect. This is AegisAI's general guidance; Microsoft's advice is to confirm service disruptions directly through the admin center.
What should I do if I clicked the link?
Tell your IT or security team at once. Microsoft recommends investigating user clicks, and it has not confirmed that credentials were collected. As a precaution, which is AegisAI's general guidance, sign out of active sessions and change your password from a different device.
For researchers
What the reporting does not tell us 6 gaps
Payload. Microsoft says the activity is consistent with credential phishing and that credential collection and malware delivery have not been confirmed. What the destination page collects is not described.
Scale and victims. Microsoft says similar messages targeted organizations across multiple US sectors. It gives no message count, no number of recipients and no number of clicks.
Actor. No threat actor or group is named, and Microsoft does not say whether the three sender domains share an owner.
The sending service. The links are described as recipient-specific SendGrid click-tracking links. Microsoft does not say whether the sending accounts were created by the attackers, compromised, or abused another way.
Microsoft's response. Some secondary reports say Microsoft blocks the domains and quarantines the messages and attribute it to a 15 November follow-up. AegisAI found no such post from Microsoft Threat Intelligence on 15 November 2026, so this article makes no claim about blocking.
Duration. The outage records give two start times and an end time that precedes the last update. The campaign's own end date is not reported.
Reported indicators 8 items
Indicators from Microsoft Threat Intelligence, written defanged exactly as Microsoft wrote them. AegisAI did not open or test any of these addresses. Treat them as historical indicators, not a live blocklist.
| Type | Indicator | Role |
|---|---|---|
| Sender domain | moments-notice[.]com | Sender domain |
| Sender domain | planetmaison[.]com | Sender domain |
| Sender domain | newbooker[.]com | Sender domain |
| Redirect destination | reallifeaccounting[.]com | Where the SendGrid links redirect |
| Path | /themelist0.html | Observed path on the destination |
| Subject pattern | Outage - [six-character code] | Observed subject lines |
| Display name | An organization-specific prefix, then "Outage Notification" or "Service Health" | Observed sender names |
| Reference | EX1550543 | Real incident number the messages cite. Not an indicator of malice: legitimate Microsoft notices carry it too |
MITRE ATT&CK mapping 2 techniques
These techniques are AegisAI's mapping of the Microsoft description. IDs and names were checked against attack.mitre.org on 15 November 2026. Microsoft names no technique, and no further technique is supported by its post.
| ID | Name | Use in this campaign |
|---|---|---|
T1566.002 | Phishing: Spearphishing Link | Emails carry a link to restore held messages. |
T1204.001 | User Execution: Malicious Link | The attack depends on the recipient clicking the link. |
References 8 sources
- Microsoft Threat Intelligence (14 November 2026). Post on X about a phishing campaign exploiting the 12 November Microsoft 365 service disruption. x.com
- Microsoft 365 Status (12 November 2026). First public post about Exchange Online incident EX1550543. x.com
- Microsoft 365 service health entries for EX1550543 and MO1550821, read on 15 November 2026 through a University of Pennsylvania mirror of the text. Microsoft's own admin center needs a tenant sign-in and was not opened. mailservices.isc.upenn.edu
- SecurityWeek (12 November 2026, updated 13 November 2026). Microsoft 365 Outage Disrupts Outlook, Exchange Online and Teams. Used for the spread to other services. securityweek.com
- iTnews (14 November 2026). Phishers quick to exploit extended M365 outage. Used to confirm the campaign details against a second outlet. itnews.com.au
- CISA (19 July 2024, updated through 6 August 2024). Alert on the widespread IT outage of July 2024. Used for the earlier outage-themed phishing. cisa.gov
- Twilio (current documentation). SendGrid Tracking settings: how click tracking rewrites links. twilio.com
- MITRE ATT&CK. Techniques T1566.002 and T1204.001. attack.mitre.org
Every email, investigated.
When a message cites a real incident and asks for a click, see how AegisAI reads the request and where the link leads.
Phishing that reaches the inbox, stopped before anyone clicks.
Book a demoRelated posts
More from the AegisAI blog.

