We are at a critical inflection point. Our low-code platform is preparing for an immediate scale-up to 3,000,000 concurrent users. We currently operate on a GKE-based architecture with 78 microservices and a MongoDB Atlas backend. We need a Lead Site Reliability Engineer who can transform our current synchronous system into a high-concurrency, asynchronous engine capable of surviving massive traffic spikes without database or compute failure.
Responsibilities
● Decoupled Architecture: Transition synchronous API flows to Google Cloud Pub/Sub to act as a shock absorber for a MongoDB Atlas M60+ cluster.
● Database Guardrails: Implement and own the "Speed Limit" for our database. You will configure Subscriber-side Flow Control in and Kubernetes HPA to ensure we never exceed 10,000 IOPS or 32k connections.
● Resource Isolation: Isolate heavy Puppeteer/Chrome workloads from core platform services using Cloud Run or dedicated Spot VM node pools with taints/tolerations.
● Observability & Alerts: Build a "Nerve Center" using Cloud Monitoring. You must track Message Age, Disk Throughput, and Connection Saturation with millisecond precision.
● Platform Hardening: Work with our 78 microservices to optimize their container footprints using Vertical Pod Autoscaling (VPA) and efficient bin-packing.
Technical Requirements
● GCP Mastery: Deep experience with GKE, Pub/Sub, and Cloud Run. You should know how to request and manage high-scale CPU quotas.
● Advanced You must understand how to manage the Event Loop under heavy load and how to properly ack/nack messages in a distributed queue.
● MongoDB at Scale: Experience with Atlas M60/M80 tiers. You must know how to diagnose Index Resident Memory issues and manage connection pooling at the platform level.
● The "SRE Mindset": You believe that a "Slow" system is better than a "Broken" one. You have experience implementing Backpressure and Circuit Breakers.