Intermittent Latency Affecting Serverless Robots in the US Region

Update timeline

1. resolved Jun 10, 2026, 03:36 PM UTC

   From 10 June 2026, 11:00 AM UTC to 2:20 PM UTC, Serverless Robots in the US region experienced intermittent latency due to temporary resource constraints. The issue was identified and mitigated by the engineering team, and service performance has since returned to normal.

2. postmortem Jun 18, 2026, 08:09 AM UTC

   ## Customer Impact Between June 10, 2026 at 11:00 am UTC and 2:20 pm UTC, customers using UiPath Serverless Robots in the U.S. region experienced intermittent delays and failures when starting serverless jobs. Affected jobs took longer than expected to start, and some jobs experienced delays of several minutes before execution. In some cases, jobs failed to start successfully during the impact window. ## Root Cause The incident was caused by a configuration issue that, under certain conditions, allowed an unusually high number of jobs to be submitted for concurrent processing. While UiPath Serverless Robots infrastructure automatically scales to meet demand, the surge in concurrent workload increased faster than additional capacity could be provisioned. This created temporary resource contention that degraded scheduling performance and resulted in delays and failures when starting serverless jobs. ## Detection The issue was identified through automated monitoring that detected increased latency affecting UiPath Serverless Robots. ## Response Engineering Team investigated the performance degradation and identified resource contention affecting UiPath Serverless Robots. The workload responsible for the increased resource consumption was isolated, restoring capacity for UiPath Serverless Robots and returning job start times to normal levels. Monitoring continued after recovery to confirm stable service performance. ## Follow Up We have corrected the configuration that contributed to this issue and increased available capacity for UiPath Serverless Robots. We are implementing additional safeguards to prevent similar workload spikes from affecting service performance and have strengthened monitoring to detect this class of issue earlier.