Disaster recovery

Lekko prepares for catastrophic events with comprehensive disaster recovery plans.

AWS failures

Lekko has a primary deployment in us-east-1 across multiple availability zones (AZs) and a passive standby deployment in us-west-1.

In case of AZ failures, traffic is rerouted to healthy pods in other AZs. The system is spread across at least three AZs within a region at any given time.

In case of regional failures, you can manually initiate a failover to a standby region. The ability to change configs might be disrupted for a short period of time, but reading configs aren’t affected because edge deployments continue serving evaluations.

Internal errors and data loss

If the Lekko backend crashes, it fails over to another region or reverts to the last good version.

In case of a Planetscale data loss, Lekko restores data from the latest backup and pulls missing commits from GitHub. Backups are currently created every 12 hours and retained for two days.