Gorilla Experiment Builder had 4 outages in the last 2 years totaling 18h 45m of downtime — averaging 0.2 incidents per month.
There were 4 Gorilla Experiment Builder outages since July 30, 2024 totaling 18h 45m of downtime. Each is summarised below — incident details, duration, and resolution information.
A sudden increase in traffic to the platform resulted in heavily impaired performance for a short period. The issue was diagnosed, rectified, and measures have been taken to prevent future occurrences.
We are aware of an issue with Microsoft Azure (our cloud service provider) which is currently impacting the Gorilla platform. We are continuing to monitor the issue and our technical team will take whatever mitigating options become available to them. We will provide an update as soon as we are able to.
We are continuing to investigate this issue.
Azure have successfully rolled back to their last know good configuration, and the service's performance should continually improve. They have forecast recovery by 23:20 UTC on 29 October 2025, though performance may still be degraded until then as they continue to restore nodes.
A fix has been implemented and we are monitoring the results.
Microsoft Azure are no longer reporting any issues and the platform has returned to normal stability, though we are continuing to monitor and respond to the aftereffects of the outage
A service outage on Azure resulted in task assets distributed via an Azure CDN (Content Delivery network) being unavailable. For Task Builder 1 and Questionnaire Builder 1 this will have resulted in some stimuli failing to load. For Task Builder 2, Questionnaire Builder 2, Shop Builder and Game Builder, this will have resulted in participants receiving an error message "No-Mani-1" on initial task load. While the service issue appears to have been resolved, we'll continue to investigate and identify any routes for mitigating similar issues in the future (especially as this is the second time in as many months that the Azure CDN has experienced difficulties!)
Azure (our cloud service provider) are experiencing issues with their CDN (Content Delivery Network) and associated services. As we use Azure's CDN to stream task/questionnaire assets such as image/videos etc. as well as some core definition/manifest files within the tooling, some tasks/questionnaires and experiments are now failing to load. Stimuli failing to load or tasks/questionnaires failing to load with the No-Mani-1 error are symptomatic of this issue. We are closely monitoring the situation on Azure and can only ask for the continued patience and understanding of our users while Azure works to resolve the core issues!
Now that sufficient access to Azure has returned, we've deployed a code change to temporarily remove the affected services from our workflow. Task/Questionnaires and Experiments should now function correctly (though may take slightly longer to load than normal!) We're continuing to monitor the issue, to make sure no other problems arise!
While the wider issues with Azure persist, our fixes to circumvent the affected resources (Azure CDN's) seems to have worked as desired! Once the main issue on Azure is resolved, we'll reinstate the CDN's, which should bring stimuli load times back down to normal levels!