Few customers using Computer vision had spike of 500 errors in US region

Update timeline

1. resolved May 05, 2026, 07:20 PM UTC

   Between 16:17 and 16:22 UTC on 5/5/2026, a few subset of customers using Computer Vision in US region observed a spike of 500 status code errors due to MLS not being reachable. This issue is resolved and we are actively working on investigating the root cause.

2. postmortem May 19, 2026, 07:18 PM UTC

   ## Customer Impact Between 16:17 and 16:22 UTC on May 5, 2026, a subset of customers using Computer Vision in the US East region experienced a spike in errors when accessing the platform. During this 5-minute window, requests were returning 500 status code errors due to Location Service, a core component responsible for routing platform requests, being temporarily unreachable. ## Root Cause The incident was triggered by a planned infrastructure maintenance activity, a routine upgrade to an internal networking mesh component used across all UiPath Platform services. This upgrade caused all services to restart simultaneously, placing unexpected demands on shared compute resources. As a result, the Location Service was unable to spin up replacement instances quickly enough to maintain full capacity. The service ran on limited capacity for approximately 5 minutes, after which the service recovered automatically. ## Detection and Response The issue was identified through internal monitoring and later corroborated through customer reports. A formal investigation was initiated to confirm the root cause. ## Follow-up We take this incident seriously and have already begun implementing improvements to prevent a recurrence. Actions include: * Improving how our infrastructure scales up fast enough and ensuring we maintain minimum required capacity during high-demand events