Microsoft's outage was a warning. Plan for the next one
Last week, many Microsoft 365 customers went about a day without reliable email. In a separate incident the same evening, contact center agents on Dynamics 365 Contact Center could not sign in to take calls or messages for close to eight hours. The outage is over, but the question it raised is still open: if a cloud service you depend on stops working tomorrow, what does your business actually do?
What happened
According to Microsoft's preliminary post-incident report, as reported by The Register, a change made in July to Microsoft 365 network gateway servers carried a defect that stayed dormant for months. On November 12, a routine Teams failover pushed extra load onto the gateway that also serves Exchange Online, while roughly 14% of that gateway's capacity was already out of service. The defect spread as traffic moved to healthy servers, and at the low point more than 80% of those servers stopped responding. Teams, OneDrive and SharePoint were able to move off the faulty gateway. Exchange Online could not, because of its scale.
Microsoft's monitoring flagged the problem at 16:03 UTC on November 12. A fix that returned the code to a known good state went out at 07:20 UTC the next morning, and recovery was confirmed that afternoon. Microsoft said in its report that its "normally automated recovery processes were not designed to restore service from this broad unresponsive state." It has promised a final review on November 20.
The contact center incident was separate. CX Today reports that Dynamics 365 Contact Center agents could not sign in from 17:00 GMT on November 12 until 00:57 GMT on November 13, after part of the database infrastructure behind the service lost connectivity. CX Today also reports that Microsoft Threat Intelligence saw phishing emails posing as Microsoft 365 service health notices and quoting the real incident number.
Every cloud provider has outages
This is not a reason to leave Microsoft, or any provider. Every large cloud platform has bad days, and a well-run cloud service is usually more reliable than what a mid-size business could build alone. The lesson is about dependency. When email, calling, the contact center, file sharing and your AI tools all sit on one platform, a single failure can reach all of them at once.
Microsoft says plainly that continuity is shared. Its guidance on customer and partner continuity responsibilities states that Microsoft provides the services and the customer is "responsible for connecting your client endpoints, managing identity and access, and determining how those services are used." Your continuity plan is yours to write.
Know what your SLA actually pays
Many buyers treat an uptime percentage as protection. It is a commitment with a defined remedy, and that remedy is usually a credit against the fees for the affected service, not compensation for lost revenue or overtime. Read the current version of the Microsoft Online Services SLA, or your own provider's equivalent, and confirm three things: what counts as downtime, how credits are calculated, and who has to file the claim and by when. If you buy through a partner, ask whether the partner files for you.
Microsoft's service health documentation says that for incidents with broad impact, a preliminary post-incident review arrives in the admin center within 48 hours of resolution, with a final review within five business days. Make sure someone on your side actually reads those reports and decides whether anything should change.
Questions to settle before the next outage
- Which business processes stop if email is down for a full day? Who decides when to switch to a fallback, and how do you reach staff and customers?
- If contact center agents cannot sign in, where do calls go? Can the carrier or platform reroute to an overflow number, a voicemail with a clear message, or mobile phones?
- Do your phone service, contact center and collaboration tools share the same identity, network gateway or provider, and does that create a single point of failure?
- How will you communicate with staff during an outage? Microsoft suggests an internal status dashboard to reduce helpdesk traffic.
- Do employees know not to trust service health emails, and to check the admin center or your IT team instead?
- When your contracts come up for renewal, what does each provider commit to on uptime, failover, notification and post-incident reporting?
Put it into your next renewal
The best time to fix a dependency is before you sign the next term. When a phone, contact center or collaboration renewal comes up, ask each provider to explain its failover design in plain terms, show its incident history, and describe how customers are notified. Compare options that keep a second path for your most critical customer contact, such as separate voice routing for the contact center. Then run a short tabletop exercise with your own team. One hour walking through a day without email will show you more than any SLA.
How StrategIQ IT can help
Before founding StrategIQ IT, Brian spent years as the enterprise customer accountable for technology, then built and led a CX and collaboration business on the partner side. We help clients map where their communications, contact center and collaboration tools depend on each other, set practical fallbacks, and build resilience requirements into provider selection and renewals. We are vendor-agnostic by design, with partnerships across more than 200 providers, so we can put the full range of options in front of you. In most cases StrategIQ is paid by whichever provider you choose, so we have no reason to steer you to one platform, and working through us does not raise your price. Our disaster recovery checklist is a practical place to start. If you want someone to own provider governance and renewals on an ongoing basis, our fractional CIO / CTO service provides part-time senior technology leadership. Scope, fees and how StrategIQ is paid are explained before you proceed. The first 20-minute conversation is free.