On June 24, 2026, between 08:26 PM and 08:55 PM PDT (June 25, 03:26 UTC to 03:55 UTC), E2B experienced a full service outage affecting the API, sandbox provisioning, and the dashboard. Service was unavailable for approximately 29 minutes.

The outage was caused by configuration drift between our production environment and our infrastructure-as-code (IaC).
On June 4, production database credentials were rotated during routine operations, but the change was applied directly to production without a corresponding update to IaC. This left the committed configuration out of sync with the live environment.
On June 25, a routine infrastructure update reapplied the stale IaC configuration. This published a new version of the database secret containing an outdated connection string. As API instances picked up the new secret, they lost the ability to connect to the database.
Because a secret update triggers a simultaneous restart across all API instances, every instance reloaded the bad configuration at once, leaving no healthy instances and taking the service down.

• API: Fully unavailable

• Sandbox provisioning: Fully unavailable

• Dashboard: Fully unavailable

This incident was caused by an invalid infrastructure configuration. There is no evidence of data loss, data corruption, or unauthorized access.

Our on-call team identified the invalid database secret, restored the correct production database connection string, and verified database connectivity before returning the service to production. Service was fully restored at 08:55 PM PDT (03:55 UTC).

We take this incident seriously and are implementing the following improvements:

• Health-gated deployments: Configuration and secret changes will roll out gradually with health checks between deployment steps, preventing simultaneous failures across all API instances.

• Infrastructure-as-code hygiene: Production configuration changes will be immediately reflected in infrastructure-as-code to eliminate configuration drift.

• Database connectivity health checks: API health checks will validate live database connectivity, allowing deployment issues to be detected before traffic is routed to unhealthy instances.

• Improved secret handling: Secret updates will be decoupled from live service restarts, reducing the blast radius of configuration changes.

We apologize for the disruption. Reliability is a core priority for E2B, and we are committed to strengthening our deployment processes to prevent incidents like this from recurring.
If you have any questions or were affected in a way not covered here, please reach out to us directly.

The E2B Team