Candidate workspaces inaccessible

Update timeline

1. investigating May 08, 2026, 12:50 AM UTC

   We're currently investigating an issue that is causing errors when launching candidate workspaces.

2. identified May 08, 2026, 01:17 AM UTC

   We've identified the root cause as an AWS outage in US East 1, Availability Zone 4. We're actively working on provisioning infrastructure in non-impacted availability zones to unblock candidates as fast as possible. You can also follow AWS's status page for the latest updates on service in AZ4: https://health.aws.amazon.com/health/status

3. identified May 08, 2026, 02:46 AM UTC

   We are continuing to work on a fix by deploying changes to route traffic to newly provisioned infrastructure in a different AZ.

4. resolved May 08, 2026, 06:17 AM UTC

   All traffic is now being successfully routed to an unaffected availability zone. Candidates should no longer encounter errors when loading their workspaces. Despite intermittent signs of recovery in AZ 4, we continued to observe degraded performance. Therefore, we moved forward with deploying new infrastructure in a separate AZ and routing traffic there.

5. postmortem May 08, 2026, 03:20 PM UTC

   ## Summary On May 7, 2026 at 8:42 PM ET, candidates began experiencing failures in the Ropes assessment platform. The root cause was a power and cooling failure caused by a thermal event at an AWS data center in the US-East-1 region \(availability zone \`use1-az4\`\), which hosted a core component of our candidate workspace infrastructure. Other parts of the Ropes platform — the customer portal, ATS integrations, results engine, problem creation, API, and webhooks — remained operational. Candidate data, including names, emails, and pre-existing results, were never impacted and remains secure throughout. ## Customer Impact Candidates attempting to start new assessments were unable to launch their workspaces. Candidates already in an assessment began seeing a blank screen as their workspace lost connection to our infrastructure. Once we confirmed the scope of the issue, we proactively blocked new candidates from starting assessments to prevent them from entering a broken experience. ## Timeline \(all times UTC\) * May 8, 00:25 — AWS reports impairment in US-East-1 availability zone \`use1-az4\` due to a thermal event. * May 8, 00:30 — Ropes begins detecting elevated errors on candidate workspace launches and opens an internal incident. * May 8, 00:50 — Incident posted to our public status page, and a banner is deployed across customer subdomains blocking new candidates from starting assessments. * May 8, 01:17 — Root cause identified as the AWS AZ-4 outage. Engineering begins provisioning replacement infrastructure in an unaffected availability zone. * May 8, 02:46 — Code and infrastructure changes deployed to route candidate traffic to the newly provisioned infrastructure. * May 8, 02:46 – 06:17 — Engineers work through residual issues caused by the cutover: database references to the old infrastructure are updated, DNS record caches are flushed, and SSL certificates are provisioned. * May 8, 06:17 — All candidate traffic successfully serving from healthy infrastructure. Incident marked resolved. * May 8, post-resolution — A subset of candidates continued to observe issues due to database configurations that had not been fully updated to reference the new control plane infrastructure. These were patched as they were identified, and assessments were re-enabled at 12:51 PM UTC. * May 8, 11:00 — As of 11:00 AM ET on May 8, AWS has not yet repaired the impacted hardware in \`use1-az4\`, which is complicating recovery of a small set of in-progress candidate assessment files \(the working files candidates were editing during their assessment\). We will continue working with AWS on recovery and follow up with affected customers directly. To be clear: candidate data, including names, emails, and pre-existing results, were not stored on this hardware and were not affected. ## Root Cause Ropes runs candidate assessment workspaces on dedicated infrastructure provisioned, configured, and monitored by a central control plane. That control plane was running in the same zone affected by the AWS thermal event. When the AZ went down, the control plane became unreachable, which prevented candidate workspaces from launching or reconnecting. Rather than wait for AWS to fully restore the zone, the Ropes engineering team stood up new control plane infrastructure in a different availability zone, updated DNS, and shipped configuration changes to point our system at the new location, at which point candidate workspaces became accessible again. ## What We're Doing Events of this nature at AWS are extremely rare, but we are prioritizing better failover mechanisms in our candidate control plane so that we are more resilient to similar events in the future. Throughout the incident, we posted timely updates on ropes.ai/status, and we will continue to operate that page transparently for any future events. ## Closing We know our customers and their candidates count on Ropes to be available when trying to take an assessment, and we fell short of that. We're sorry for the disruption this AWS outage caused to your hiring workflows and to candidates who had to re-work their plans. If you have questions about specific candidates or assessments affected, please reach out to your Ropes contact directly. — The Ropes Team