This incident has been resolved. We switched our EU Central worker servers to an alternative DNS resolver at around 16:45 UTC, after which import timeouts stopped and queue times returned to normal. All regions and conversion types are operating normally. Jobs that failed during the incident can be retried.

Since approximately 15:25 UTC, conversions processed in our EU Central region are experiencing elevated queue times and an increased failure rate. Import tasks that download files from URLs or cloud storage are particularly affected and may fail with a download timeout. The cause is a DNS fault at our infrastructure provider in the affected data centers, which also prevents additional capacity from coming online. The API, the web interface and the US East region are operating normally. We are monitoring the provider's incident and working on mitigation. Failed jobs can be retried.