A standby workspace only helps if staff can sign in without Microsoft
Microsoft says the impact ran from 4:20pm UTC on Thursday 12 November to 4:10pm UTC on Friday 13 November 2026, about 24 hours. 23 Forrester's view after the outage was that it's unrealistic to diversify application by application. Redundancy works per critical function: decide which work must keep going, and give that work a second path. 509
Sign-in decides whether any fallback works
Microsoft's backup authentication system keeps some sign-ins working when Entra ID has problems, but only for an app and device used in the last 3 days, and only for sign-ins that need no interactive step. 212
Microsoft lists Google Workspace, Slack and Zoom signed in through Entra single sign-on among the apps that backup system doesn't protect. A standby tool that signs in through Entra goes down with it. 212
Google Workspace lets an admin set a group's single sign-on profile to none, so those people sign in to Google directly. That's the setting a standby group needs. 510
Microsoft never said Entra sign-in failed in the November outage. This step is about the next one.
A standby workspace for the people who can't stop
Google Workspace Business Standard lists at A$19.80 a user a month. For 50 key staff that's about A$990 a month at list price. 511
Slack Pro lists at US$7.25 a user a month billed yearly. Zoom Workplace Pro lists at US$14.16. Zoho Workplace Standard lists at US$3. 512
A separate identity provider such as Okta lists at $6 a user a month on its Starter plan, with a $1,500 yearly minimum. 513
Set it up before you need it: accounts created, the sign-in path tested, a channel or group per critical team, and the people told where to go.
Mail should wait in a queue, and the clock is 24 hours
The email standard says senders should keep retrying for at least 4 to 5 days. 132
Exchange Online retries for 24 hours before it gives up on a message it can't deliver, so mail from other Microsoft 365 organisations can bounce after a day. 514
Continuity services hold and serve mail while the main system is down. Proofpoint Essentials Business, which includes 30 days of an emergency inbox, has a published list price of US$3.03 a user a month on a 2022 price list. 515
Keep the documents people need on the day somewhere else
OneDrive sync handles up to about 300,000 items, so a synced copy of the critical folders on key laptops is realistic. 516
Microsoft 365 Backup restores back into the same tenant, so it protects against deletion and corruption. It doesn't help while the tenant itself is down. 517
Name an owner for the list: contact sheets, the incident plan, customer commitments due that week, payroll files near pay day.
Decide now who calls the switch, and at what hour
NIST's contingency planning guide puts the call with one senior official, such as the CIO, with a named successor, and says to activate when the expected outage will run longer than the recovery time objective. 518
Write the hour down. If the plan says switch at 2 hours, a 24 hour outage costs 2 hours. If nobody has the authority, it can cost the day.
Regulated firms already have to plan for this
In Australia, APRA's CPS 230 requires a tolerance level for each critical operation: the longest disruption the entity would tolerate, the most data it would lose and the minimum service under alternative arrangements. The plan must include triggers that start it. 519
CPS 230 asks entities to tell APRA no later than 24 hours after a disruption to a critical operation goes outside tolerance. The standard has applied since 1 July 2025, and the amended version from 1 July 2026. 519
In the EU, DORA has applied since 17 January 2025. Article 28 requires exit strategies for ICT services that support critical or important functions, and Article 29 asks firms to assess concentration risk when a provider isn't easily substituted. 520
In the UK, the PRA and FCA rules required firms to stay within impact tolerance for each important business service in a severe but plausible disruption by 31 March 2025. 521
The UK named Microsoft a critical third party to its financial sector in July 2026, and EU supervisors designated it under DORA in November 2025. 522
The FCA said it was monitoring the outage's impact on the City. APRA hadn't commented publicly by 16 November. 523
Some organisations had already split their critical work off Microsoft
London Ambulance Service and London Fire Brigade reported no impact. The ambulance service reportedly doesn't run its critical operations on Microsoft. 523
The Bank of England moved email traffic to Teams chats, which kept staff talking but stayed inside Microsoft. 523
The lessons page names 51 organisations in 13 countries and what kept each one working.
This is one step in the free outage playbook.