How gokcesurenkok/ha-flink-cluster-k8s works
Project Type Analysis
This is a Kubernetes-based infrastructure project specifically focused on deploying a highly available Apache Flink cluster. The architecture should emphasize the distributed nature of Flink and its integration with Kubernetes orchestration.
Key Components
a. Kubernetes Layer:
• Kubernetes cluster boundary
• ConfigMaps (flink-cfg, hadoop-cfg)
• Services layer
• Deployments/StatefulSets
b. Flink Components:
• JobManager StatefulSet (for HA)
• TaskManager Deployment
• Replication/scaling
c. External Dependencies:
• ZooKeeper (for HA coordination)
• HDFS (for storage)
Specific Elements
a. Kubernetes Components:
┌─ ConfigMaps ─────┐
• flink-cfg │
• hadoop-cfg │
• ┘
┌─ Services ──────┐
• flink-svc │
• ┘
┌─ Deployments ───┐
• JobManager │
• TaskManager │
• ┘
b. High Availability Setup:
• ZooKeeper connection for leader election
• Multiple JobManager instances
• TaskManager scaling capabilities