01Production Observability Stack
Designed and deployed a fully self-hosted observability platform powering three production environments. Centralises metrics, logs, and traces for a microservice fleet processing 80M+ events/day.
- Sub-60s alert latency across 200+ metrics rules
- Long-term metric storage via Thanos with 13-month retention
- Automated runbook dashboards linked to PagerDuty escalation
PrometheusGrafanaLokiTempoAlertmanagerKubernetesHelm
02Multi-Cluster Kubernetes Platform
Built a fleet of Kubernetes clusters across three cloud regions serving 500+ microservices. Implemented GitOps-driven deployments with full multi-tenancy, RBAC, and automated certificate management.
- CKA-certified cluster architecture with zero-downtime upgrades
- Karpenter-driven autoscaling reduced compute costs by 34%
- GitOps enforcement via ArgoCD ApplicationSets across 12 teams
KubernetesArgoCDFluxcert-managerKarpenterTerraformGolang
03High-Throughput Golang API Gateway
Architected a low-latency API gateway in Golang that proxies and rate-limits traffic for a B2B SaaS platform. Replaced a Node.js layer and delivered a 4x throughput improvement with a smaller footprint.
- Sustained 2.4M RPS at p99 latency under 12ms in load tests
- Distributed token-bucket rate limiting via Redis Cluster
- Full distributed tracing with OpenTelemetry and Jaeger
GolanggRPCRedisPostgreSQLOpenTelemetryDockerKubernetes