No components marked as affected
Resolved
The fine tuning service (across all models - gpt-3.5-turbo, babbage-002, davinci-002) is operating normally. Queue wait times are still slightly elevated but are returning to normal
Monitoring
The backlog of fine tuning jobs is steadily decreasing, we will continue to monitor the service for further disruptions. Expect an increased delay in the time it takes for a submitted job to begin running for the next few hours until the backlog settles.
Identified
The issue has been identified and jobs are now being processed again, it will take time for the queue to empty
Investigating
Approximately 5% of fine-tuning jobs have been stuck in a queued state, we are investigating