Between April 3, 2025 at 13:32 UTC and April 4, 2025 at 01:52 UTC, a subset of customers in EU region experienced degraded service when accessing certain Integration Service metadata APIs—specifically, Connections, Connectors, and Triggers. This caused intermittent instability in the UI for impacted users. All other regions and services dependent on the Integration Service remained unaffected.
The incident was triggered by a single database query from one tenant that, under increased traffic, became highly resource-intensive. This query consumed a significant number of database connections, saturating the connection pool. As a result, other customer requests were unable to establish new database sessions, leading to intermittent timeouts and degraded performance when accessing metadata APIs.
The issue was not detected by our service-specific alerting systems. It was first brought to our attention through our Site Reliability team, globally monitoring our Platform Gateway status.
Once identified, the offending process was promptly disabled, and the system was stabilized by releasing the saturated database connections. This action restored the availability and performance of the impacted services.
To prevent similar incidents in the future, we are taking the following steps: