Uptime, Outages and Vendor Reliability

#168 - Serverless as a Game Changer - Joseph Emison

Topic 9 of 14

Published Mar 25, 2024 · Clip 40:47–45:26 (4:39)

Joe shares Branch's reliability record: running in us-east-1 since 2008, only one significant regional outage in five years (Route 53) and one January 2022 vendor outage; core services like AppSync, DynamoDB, and Cognito have essentially never gone down. He discusses evaluating vendors by historic uptime rather than SLAs (which are non-compensatory), notes Amazon treats outages seriously and publishes postmortems, and describes adopting Neon serverless Postgres early in private beta with backup caching, while warning against Auth0/Okta after reliability issues.