15 reviewed items across 2 content types.
Netflix operated 1,200 microservices across 3 AWS regions handling 250 million subscribers. Their legacy Eureka-based service discovery lacked mutual TLS and service-to-service authorization, leaving east-west traffic unencrypted and unauthenticated.
Stripe processed 500 million API requests per day across 4 data centers. Their hardcoded service endpoints in configuration files caused 45-minute outages during datacenter failovers because DNS entries had to be manually updated across 800+ services.
Airbnb needed to achieve PCI-DSS Level 1 compliance for their payment processing subsystem spanning 45 microservices. Auditors required documented, enforceable network segmentation between cardholder data environment (CDE) services and non-CDE services.
Uber needed to manage feature flags and dynamic configuration across 4,500 microservices serving 130 million monthly active users. Configuration changes deployed through their CI/CD pipeline took 25 minutes to propagate, too slow for incident response.
Goldman Sachs ran 60% of their trading platform on VMs and was migrating 40% to Kubernetes. They needed a unified service mesh that spanned both environments during a 2-year migration window, with no disruption to $1.2 trillion in daily trading volume.
Lyft needed to expand from 2 to 5 AWS regions to reduce rider-to-driver matching latency below 200ms globally. Their existing service discovery was region-local, requiring manual intervention for cross-region failover that averaged 12 minutes.
Pinterest ran 800 microservices serving 450 million monthly users. Service health degradations took an average of 8 minutes to detect through their Datadog monitoring stack, and another 15 minutes for manual remediation by on-call engineers.
