FYND Backend Microservices 🛠️

Expert debugging for FYND's Kubernetes/GCP/Kafka/Node.js backend.

🎯 Use When

"Pods crashing" | "Kafka lag" | "Redis memory high" | "API latency"
"Database migration failed" | "LLM costs spiking" | "Memory leak" | "Service failures"

🛠️ 8 Core Skills

1. K8s/GCP Deployment (40% issues) - pods, scaling, graceful shutdown
2. Kafka Resilience (25%) - consumer lag, DLQ, rebalancing
3. Redis Optimization (15%) - memory, TTL, pub/sub
4. Distributed Tracing (10%) - correlation IDs, Langfuse
5. Database Patterns (8%) - Sequelize, pgvector, MongoDB
6. LangGraph Orchestration (5%) - multi-LLM, token counting
7. Performance Analysis (4%) - heap profiling, slow queries
8. Resilience Patterns (3%) - circuit breaker, backoff

🔍 Diagnostic Flow

1. Gather metrics (kubectl, kafka-consumer-groups, redis-cli)
2. Form hypotheses (OOM? poison pill? slow query?)
3. Test systematically
4. Provide fix + monitoring

📦 Bundled Resources (load as needed)

• scripts/diagnose.js - quick pod snapshot (kubectl get pods)
• references/patterns.md - fast CLI patterns for pod crash, Kafka lag, Redis memory
• references/fynd-backend-skills.md - full architecture + skills matrix + flowcharts + checklists
• references/fynd-agent-integration.md - LangGraph/LangChain tool integration guide
• references/fynd-backend-skill-template.md - template for creating new skills or extensions

Search tips: use rg -n "Use Cases|Agent Actions|Checklist|Flowchart" in the reference files to jump to relevant sections quickly.

📊 Success

94% accuracy | <5s response | 30% MTTR reduction | $740k/year savings