Microsoft 365 goes down: the plan B nobody tests (and that NIS2 and DORA require)
This morning, like many people, I lost my Microsoft 365 email. Microsoft acknowledged it on its official status page: a service degradation that left many users unable to connect to several of its services. Email, Teams and documents, all at once.
I have spent more than 30 years in IT in regulated sectors, and in these cases the useful question is never “why did it go down?”. It is a different one: “what did we have ready?”. Your provider will fail one day, and regulation does not ask you to prevent it. It asks you to have planned for it.
The lesson: everything depended on the same provider
What strikes me most about these outages is not the outage itself. It is seeing whole teams left with no way to talk to each other, because email, messaging, video calls and even sign-in depend on the same provider. The continuity plan exists, but it sits in a shared document… with that same provider.
In highly regulated banking environments I learned a simple rule: plan B cannot depend on plan A. If your crisis channel, your contacts or your procedure live in the service that has failed, you have no plan B.
What regulation requires when a provider fails
NIS2 requires business continuity and supply-chain security (Article 21). And if the outage causes severe disruption to your services, you must report it: early warning within 24 hours, notification within 72 hours and a final report within one month. I explain it in detail in how to report an incident within 72 hours.
DORA goes further for financial entities: a register of information covering all ICT providers, minimum contractual clauses and tested exit strategies for those supporting critical functions. And the GDPR reminds us of something often forgotten: losing access to personal data is also a breach, an availability one. It usually does not need to be reported, but it must always be recorded.
If you work with several regulations at once, it is worth seeing how DORA, NIS2, the ENS and the AI Act overlap: almost everything they ask of you after an outage is the same thing under a different name.
My plan B in five pieces
- A crisis channel outside your main provider. A messaging group and a phone list that do not depend on the same service as your email.
- The continuity plan, outside too. An up-to-date copy, printed or in another storage, with who decides and what happens in the first hour.
- An emergency account. A protected and tested emergency admin access, in case the usual sign-in fails.
- Backups outside the provider. If your documents only exist in one place, a long outage becomes a business problem.
- A clock and a decision. Someone records the detection time and decides, with judgement and in writing, whether to report or not.
What I would do on Monday
- Check which critical services depend on a single provider.
- Make sure the continuity plan can be read without that provider.
- Document today’s outage: timeline, impact and the reporting decision.
- Schedule a simple exercise: “no email or Teams today, how do we work?”.
If you want to review your dependence on providers and prepare a continuity plan that withstands an outage like today’s, I can help. Let’s talk.
Frequently asked questions about a cloud provider outage
Only if the outage causes, or can cause, severe operational disruption to your services and you are in scope of NIS2, DORA or the ENS. Even if you do not report it, document the timeline, the impact and the decision.
It can be: it is an availability breach. You only notify the authority if it is likely to result in a risk to individuals, but the GDPR requires you to record it every time.
A crisis channel and a copy of the continuity plan that do not depend on your main provider. Without them, the rest of the plan cannot be carried out.
Do you have to apply this under DORA, NIS2 or ENS? Tell me about it.
Book 20 minutes