How bitdrift scaled to 121 million concurrent gRPC connections on Amazon CloudFront for live telemetry sporting events
Architecture Blog
This article describes how bitdrift resolved a critical scaling issue handling 121 million concurrent gRPC connections for live T20 World Cup cricket telemetry by fixing their Route 53 DNS routing configuration.
- Weighted routing returned single IP per DNS query, causing all CloudFront edge nodes to resolve to one NLB per TTL window (thundering herd effect)
- Long-lived gRPC connections amplified DNS routing problems compared to stateless HTTP traffic
- Scaling from 2 to 6 NLB IPs had no effect because weighted routing still returned only one IP per response
- Switched to multi-value answer routing, which returns up to 8 IPs per DNS query with built-in health checks
- After DNS configuration change, handled 121 million devices and 110K+ peak requests/sec with zero server-side errors
- Peak 5xx error rate improved from 79.8% to 0.033% (99.96% reduction)
DNS routing policy selection is critical at extreme scale; multi-value answer routing eliminates single-origin bottlenecks for CloudFront origins with persistent connections.
The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.
Related articles
2026
2026
2026
2026
The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.