Identity Platform Degradation

The MFA Provider has begun returning elevated latency and intermittent failures, while Session Store write/lookup pressure is independently rising at the same time — two concurrent, separately-caused problems, not one. Token Verification remains fully healthy throughout, which must not be read as 'everything is fine': an already-valid token verifying successfully says nothing about whether a new login or a fresh MFA challenge can currently succeed. Five business invariants govern every decision here: (1) existing valid tokens may be verified without forcing unnecessary MFA; (2) high-risk authentication may never bypass mandatory MFA, no matter how much it would improve availability metrics to do so; (3) fail-open is never permitted for privileged or high-risk actions — if MFA cannot reliably protect them right now, they are blocked or explicitly taken offline, never silently waved through; (4) low-risk users may enter a bounded degraded mode, but only when policy explicitly allows it and risk tiers can still be reliably told apart; (5) session-store relief must not invalidate active sessions unnecessarily. A broad MFA bypass is this incident's central trap: it makes login latency and failure-rate dashboards look dramatically better while directly violating invariants (2) and (3), and it creates a permanent, non-resetting historical count of high-risk actions authenticated without MFA that restoring policy afterward does not erase. Decide how to keep low-risk traffic moving without weakening protection for high-risk and privileged actions, how to relieve Session Store pressure without unnecessarily invalidating active sessions, and how much elapsed recovery time the MFA Provider and the Session Store each genuinely need before handing the incident back.

Difficulty

Principal

Expected Time

~22 min

Category

Security & Identity