Medius incident

Capture - OCR degradation

Minor Resolved View vendor source →

Medius experienced a minor incident on December 9, 2025 affecting Capture (US), lasting 17h 32m. The incident has been resolved; the full update timeline is below.

Started
Dec 09, 2025, 08:12 PM UTC
Resolved
Dec 10, 2025, 01:45 PM UTC
Duration
17h 32m
Detected by Pingoru
Dec 09, 2025, 08:12 PM UTC

Affected components

Capture (US)

Update timeline

  1. investigating Dec 09, 2025, 08:12 PM UTC

    We are experiencing significant performance degradation in Capture OCR. We are investigating the issue.

  2. investigating Dec 09, 2025, 08:23 PM UTC

    We are currently experiencing a high backlog of OCR jobs. The worker pool reached its maximum size and was unable to process incoming traffic fast enough. We have switched to manual scaling and doubled the number of workers. Processing is now catching up and the queue is shrinking.

  3. identified Dec 09, 2025, 09:27 PM UTC

    The root cause has been identified. Ongoing infrastructure patching impacted OCR service performance. After temporarily overprovisioning the infrastructure, we were able to process the entire queue, and the system has returned to normal operation.

  4. monitoring Dec 09, 2025, 09:27 PM UTC

    A fix has been implemented and we are monitoring the results.

  5. resolved Dec 10, 2025, 01:45 PM UTC

    This incident has been resolved.