API Gateway @ Kong Summit

At Kong Summit 2022, I had the privilege of sharing how to uplevel its reliability engineering through strategic use of API gateways. The talk, “How Uber Uplevels Reliability Engineering by Leveraging API Gateways,” highlighted how centralized API infrastructure can play a pivotal role in building resilient, large-scale distributed systems.

I walked through Uber’s three-pronged reliability strategy: mitigation, detection, and prevention—all enabled or significantly enhanced by capabilities embedded at the API gateway layer.

For mitigation, traffic failover mechanisms were introduced, allowing the gateway to intelligently route requests based on city-level awareness while preserving session stickiness. On the prevention front, global rollback beacons were implemented to track a core set of edge metrics, enabling rapid rollback decisions when issues arise. Another preventive measure is tenancy injection, which allows test and staging scenarios to safely run within the production environment by tagging requests with contextual tenant information.

Detection is strengthened through advanced observability. Custom headers are used to enrich every request with metadata like API tier, method, user agent, device version, and more—providing deeper visibility across the system. Additionally, real-time funnel analysis and user impact tracking make it easier to identify regressions and quantify how specific issues are affecting users, enabling more informed and faster remediation.