Platform status

Service status and availability

Current platform availability, 90-day monitoring history, and incident reports.

Last updated Nov 22, 18:16 UTC

Degraded Performance

Uptime over the past 90 days.

Website

Operational
90 days ago100 % uptimeToday

API

Operational
90 days ago100 % uptimeToday

Forecasting Engine

Disrupted
90 days ago92.79 % uptimeToday

Green: availability of 98% or more. Amber: availability below 98% or a documented capacity incident. Grey: no data. Select a bar for details.

Past incidents

Aug 31, 2026

High demand and compute capacity pressure

Event dates below. Timeline supplied by the platform team and published on Sep 12, 2026.

Compute access disruption — Access issues with the compute cluster became more frequent. Requests showed unusual latency patterns and elevated round-trip times. The website and API remained available.

Aug 31, 2026

Cause identified — The team identified a partial electricity outage at the data centre that supplies compute resources. Traffic was partially rerouted within the availability zone, which increased latency. These access and latency issues persisted throughout the day.

Sep 1, 2026

Inference traffic prioritised — Pre-scheduled fine-tuning runs used more bandwidth than anticipated and competed with client inference traffic for compute resources. To relieve pressure and prioritise model requests from clients, the team interrupted the fine-tuning jobs.

Sep 2, 2026

Queue issues identified — High compute pressure and interrupted fine-tuning jobs exposed issues in the job queue logic, which reduced Forecasting Engine availability. The team began work to resolve the queue issues.

Sep 3, 2026

Queue remediation continued — The job queue issues continued to affect compute workloads. The team continued work on the queue fix; the website and API remained available.

Sep 4, 2026

Queue remediation continued — Forecasting Engine availability remained reduced due to the same queue issues. The team continued work to restore regular job processing.

Sep 5, 2026

Queue fix completed — The team completed the fix for the job queue issues and restored regular availability patterns. Daily availability was 66%, which includes disruption before the fix.

Sep 6, 2026

Availability after the fix — Forecasting Engine availability reached 96% for the day after the queue fix. The website and API remained available.

Sep 7, 2026

Jul 28, 2026

Job submissions failing

Resolved — The issue has been resolved and job processing has fully recovered. Impact occurred from 10:25 UTC on Jul 28 to 06:48 UTC on Jul 29.

Jul 29, 06:48 UTC

Identified — We have identified the cause of the issue affecting job submissions and are working on a fix. We will provide an update as soon as possible.

Jul 28, 14:55 UTC

Investigating — We are investigating reports of failing job submissions. We will provide an update as soon as possible.

Jul 28, 10:40 UTC

Jul 26, 2026

Slow or failing file uploads

Resolved — The issue affecting file uploads has been resolved. Uploads are completing normally again.

Jul 26, 08:15 UTC

Identified — We have identified the cause of slow and failing file uploads and are working on a fix.

Jul 26, 06:55 UTC

Jul 24, 2026

Intermittent job failures

Resolved — An issue during the platform's initial rollout caused some submitted jobs to fail intermittently. The issue has been resolved and affected jobs completed successfully on resubmission.

Jul 24, 21:30 UTC