Customer impact

Between approximately July 8, 2026 at 5:00 pm – 6:00 pm UTC and July 9, 2026 at 5:08 am UTC (about 12 hours), some customers in the U.S. and Europe regions experienced issues when deploying Automation Solutions packages that included Integration Service connectors or connections.

Affected deployments stayed in progress and could time out after one hour instead of completing normally. Deployments that did not include Integration Service connectors or connections completed successfully.

Root cause

The issue was caused by a recent change in Integration Service (Connection Service) that was intended to improve connection handling. That change interrupted the completion signal that Automation Solutions expects after Integration Service finishes processing connectors or connections in a package.

Because Automation Solutions waits for that signal before marking a deployment complete, affected deployments remained in progress until they timed out. We confirmed this by reverting to the change, after which deployments completed successfully again.

Detection

The issue was detected through automated alerts and by reproducing stuck Automation Solutions deployments. The first alerts were observed around 5:00 pm – 6:00 pm UTC on July 8, 2026. The incident was formally declared at 9:26 pm UTC, and the public status page was updated to investigate at 10:23 pm UTC.

Response

After the incident was declared at 9:26 pm UTC on July 8, 2026, our team confirmed that only Automation Solutions deployments containing Integration Service connectors or connections were affected, and identified a recent Integration Service change as the cause.

We updated the public status page as the investigation progressed (investigating at 10:23 pm UTC, identified at 11:36 pm UTC). We first attempted a targeted fix, then switched to reverting to the change that introduced the problem.

On July 9, 2026, the revert completed at 5:08 am UTC. We verified that solution deployments worked again and marked the status page resolved at 5:12 am UTC.

Follow-up

To reduce the chance of a similar issue, we are:

1. Gating any reintroduction of the reverted Integration Service change on automated validation of Automation Solutions deployments that include Integration Service connectors or connections, so this path is covered before the change ships again
2. Expanding automated tests to cover the Automation Solutions deployment path through Integration Service, so regressions of this kind are caught before production
3. Continuing to monitor the original connection-handling issue this change was intended to fix. Any future fix for that issue will be required to pass the automated regression tests described above before deployment