Home Incident All Incidents INC0042891
INC0042891 Intermittent SSO Authentication Failures — Corporate Identity Provider ● P2 High In Progress ⏱ SLA: 3h 14m left
Incident Details ▲ Collapse
Number
State
In Progress
Impact
2 — Medium
Urgency
2 — Medium
Priority
2 — High
Category
Software
Subcategory
Authentication & Access
Caller
Rachel Kim · IT Operations
Opened
Apr 4, 2026 — 09:47 AM (PDT)
Assignment Group
Assigned To
Marcus Lindqvist · IAM L2 Engineer
Affected Users
~340 (Engineering, Finance, Legal)
Location
Global (multi-region)
Description

Since approximately 08:30 AM PDT on April 4, 2026, approximately 340 users across Engineering, Finance, and Legal departments have been experiencing intermittent failures when authenticating through the corporate Single Sign-On (SSO) system. Affected users report receiving "Authentication failed — please try again" errors after entering valid credentials on the corporate identity portal (sso.corp.example.com). The issue appears intermittent: roughly 30–40% of login attempts fail on the first try, with subsequent retries succeeding within 2–3 attempts.

The corporate SSO stack uses Okta as the primary Identity Provider (IdP), backed by Active Directory (AD) federation via ADFS on two on-prem servers (ADFS-PROD-01 and ADFS-PROD-02). Affected applications include Salesforce, Jira, Workday, and the internal HR portal. VPN-authenticated sessions appear unaffected; only browser-based SAML flows are impacted.

Initial triage shows elevated 500-series error rates from ADFS-PROD-01 beginning at 08:28 AM (correlation with a scheduled maintenance window for SSL certificate renewal). ADFS-PROD-02 is healthy. Okta logs show intermittent SAML assertion signing errors from ADFS-PROD-01. No changes were made to the Okta tenant configuration in the past 14 days. The scheduled maintenance completed at 08:15 AM; however the certificate trust store on ADFS-PROD-01 may not have been reloaded correctly.

{fill}

Caller has confirmed that manually forcing Okta to route all SAML requests to ADFS-PROD-02 as a temporary load-balancing override did reduce the error rate significantly, but this is not a sustainable fix and full load distribution must be restored. A permanent resolution is needed before the end-of-day SLA window.

Work Notes (3)
Additional Comments
Activity Stream
ML
Marcus Lindqvist · Apr 4, 2026 — 10:22 AM Work Note
Confirmed ADFS-PROD-01 certificate trust store not reloaded after renewal. Attempting manual IIS reset + service recycle on ADFS-PROD-01. Will monitor Okta error rate after restart. Current error rate: 34% of SAML flows from PROD-01. PROD-02: 0% errors.
JO
Julia Okonkwo (IAM Lead) · Apr 4, 2026 — 10:05 AM Work Note
Escalated to IAM L2. Okta System Log reference: event_id 0oa8xyz1234. Opened bridge call with Okta TAM (case #CS-0029184). Parallel track: checking if ADFS SSL cert renewal script left the new cert bound to incorrect service account.
RK
Rachel Kim (Caller) · Apr 4, 2026 — 09:51 AM Work Note
Ticket submitted. Confirmed ~340 affected users. Workaround applied: Okta load balanced to ADFS-PROD-02 only. Help desk fielding ~20 calls/hour. Request priority escalation due to business impact.