For users and administrators
Outlook not working? Start with triage
Check service health, isolate affected workflows and preserve your work before changing local settings.
Read the guide →Understand → investigate → practise
Microsoft 365 outage debugging
An outage raises two different questions: what is affected, and how should your own software recover? Start with the evidence, then practise the engineering decisions in a safe local simulation.
All reading is public. The two new labs are free with an account. Prefer to explore first? The guest lab introduces HTTP request boundaries without signing in.
Dated incident explainer · November 13, 2026
What happened, and what remains uncertain?
Our article records Microsoft’s public update published at 14:56:52 UTC, checked at 15:12 UTC on November 13. It describes improving service recovery with intermittent connectivity and mail delays still reported at that time. The root cause was not confirmed in that update. This collection is educational, not a live status monitor.
Choose the question you need answered
These evergreen guides distinguish a user’s next step from an application’s retry policy. Each links to primary technical sources and a way to verify the result.
For integration developers
Microsoft Graph 429, 503 and Retry-After
Read the retry contract, respect seconds and HTTP dates, and keep recovery traffic within a budget.
Read the guide →For developers and operators
Email queued but not delivered
Separate acceptance from delivery, investigate a backlog and avoid duplicate sends during recovery.
Read the guide →Triage → retry → recover
Build a recovery plan you can explain.
Work through three guides in order, then apply the ideas in a debugging lab. Tick each guide when you have reviewed it; you can change your checks at any time.
This is a device-only, self-reported reading checklist. It does not verify lab acceptance or earn rewards. Checks are saved in this browser when storage is available.
Triage the outage
Separate a local Outlook problem from a wider service incident before changing configuration.
Handle retries safely
Understand Microsoft Graph throttling, Retry-After and bounded retries during a disruption.
Recover queued email
Trace delayed delivery and preserve queue expiry instead of retrying indefinitely.
Next guide: Triage the outage
Continue learning pathShare the path with your team.
Send the same public reading path. Your checklist and personal information are not included in the link.
Turn the explanation into a working fix
Two free labs with an account. Both are original Buglyst simulations with deterministic clocks and fake providers. They do not use Microsoft’s source code, contact live services or establish the cause of the November 13 incident.
Medium · 24 minutes · JavaScript / Node.js
A Retry-After That Returns Too Soon
A simulated safe-read client burns through its retries before the service cooldown ends. Trace the clock and repair the boundary.
Open the retry labMedium · 25 minutes · JavaScript / Node.js
The notification backlog that never drains
A simulated notification queue keeps stale work alive after a provider interruption. Restore expiry while preserving useful retries.
Open the queue labBuild a habit beyond one incident
After the retry lab, return for the queue investigation. Compare the evidence and explain why each fix preserves a safety boundary. Share this public learning path with a teammate; they can read the guides before creating an account.
Continue with backend reliability practice →Independent educational material from Buglyst. Not affiliated with Microsoft. Reading checkmarks are self-reported on this device; accepted lab work is verified separately by the grader.