Harness experienced a critical incident on September 15, 2026 affecting Platform and Platform and 1 more component, lasting 20m. The incident has been resolved; the full update timeline is below.
Affected components
Update timeline
- investigating Sep 15, 2026, 05:40 PM UTC
We are currently investigating this issue.
- identified Sep 15, 2026, 05:40 PM UTC
The issue has been identified and a fix is being implemented.
- monitoring Sep 15, 2026, 05:43 PM UTC
A fix has been implemented and we are monitoring the results.
- resolved Sep 15, 2026, 05:43 PM UTC
This incident has been resolved.
- postmortem Sep 29, 2026, 12:11 AM UTC
# Root Cause Analysis Report ## Brief API Login Failures Due to a Security Policy Misconfiguration ## 1. Executive Summary On September 11, 2026, a change to one of our network security policies was deployed with a configuration error. For approximately 6 minutes, this caused login requests to be incorrectly blocked, preventing some users from logging in. The issue was detected automatically by our monitoring systems. The problematic change was fully reverted within 5 minutes of detection, restoring normal login functionality. # 2. Impact * For approximately 6 minutes \(12:30–12:36 PM CT on September 11, 2026\), login requests matching certain URL paths were incorrectly blocked. * Any user attempting to log in during this window could have been affected * No other functionality was affected. Users already logged in prior to the window were not impacted. # 3. Root Cause As part of a planned security enhancement, our team implemented a new network security policy intended to block a small number of specific, exact URL paths. During implementation, the rule was configured with an erroneous matching condition instead of the intended exact-match condition. As a result, the policy blocked any request whose path began with those values, rather than only the exact paths intended , which included legitimate login requests that happened to share that prefix. This caused those login attempts to be rejected during the affected window. # 4. Resolution * The problematic policy change was identified and fully reverted within 5 minutes of the first alert. * Normal login functionality was confirmed restored immediately following the revert. # 5. Risk of Recurrence & Preventive Actions | **Corrective / Preventive Action** | | --- | | Enhanced our QA validation step in the process for this class of security policy change, so the deployed configuration is checked against its intended specification before go-live. | | Enhanced automated test coverage that specifically validates exact-match versus prefix-match behavior for path-based rules, as an additional safeguard on top of the review process. |