Customer Impact

Between May 20, 2026 at 3:46 pm UTC and May 20, 2026 at 3:59 pm UTC, the Location Service in UiPath Commercial Cloud was unavailable for a subset of customers. Affected customers in the US region may have been unable to use functionality that depends on the Location Service during this 13 minutes incident.

The incident was declared customer-impacting at 3:47 pm UTC on May 20, 2026, and our public status page was updated to reflect the disruption.

Root Cause

The incident was triggered by a rollout of a new version of Location Service. The application version was healthy, but the rollout configuration incorrectly under-provisioned the canary deployment. A bug in the canary configuration made the canary start with only a replica count of 1 instead of having parity with the number of replicas in the primary deployment. As the canary runs in steps and as each step created a new replica set, the issue started to manifest once the canary started to serve a higher and higher percentage of the traffic but with not enough capacity for handling all allocated traffic. Location Service became partially unavailable; the incident was treated as customer-impacting; mitigation was reported at 3:58 pm UTC, once the switch to the new primary replica set, started with a correct number of pods, and the service was confirmed fully operational at 4:13 pm UTC on May 20, 2026.

Detection

Monitoring alerted immediately, and the incident was classified customer-impacting within one minute.

Response

At 3:46 pm UTC, the automated alert fired and our engineering team began coordinating a response. By 3:47 pm UTC, the incident was classified as customer-impacting and a public status page entry was created to keep customers informed.

At 3:58 pm UTC, the public status page was updated to "Monitoring," confirming that the issue had been mitigated and that services were operating normally. At 3:59 pm UTC, the service was confirmed as available.

At 4:13 pm UTC, the public status page was updated to "Resolved," confirming that the outage had been fully resolved and the Location Service was operational. The incident was formally closed at 4:14 pm UTC.

Follow-Up

1. The bug in the canary mechanism has been fixed and released: the number of minReplicas for the canary is now aligned with the primary deployment.
2. The canary mechanism has also been changed so a new replica set is not created for each step, reducing any chance of disruption due to pods restart problems