Kafka Topology Governance

Topology & Copartitioning Validation

Mismatched stream topologies lead to silent join failures in production. AegisMesh intercepts deployment requests and performs strict validation on your stream topologies.

Before a microservice is deployed, the mesh verifies that all input topics and repartition topics have identical partition counts. If it detects a Copartitioning Mismatch, it actively blocks the deployment to prevent data corruption.

High Availability Guardrails & Admission Control

Developers often deploy topics with default, unsafe settings. The AegisMesh Admission Webhook enforces enterprise reliability rules automatically via a mutating webhook:

• Replica Enforcement: Automatically injects JSON Patches to ensure min.insync.replicas is high enough for cluster availability.
• Disk Leak Prevention: Validates that internal Streams repartition topics strictly utilize cleanup.policy=delete with aggressive retention limits to prevent broker storage exhaustion.
• Delete Protection: Blocks malicious or accidental deletion of critical production topics without explicit overriding annotations (e.g., kafka.health.io/allow-delete: true).

Zombie Consumer Detection

High volume traffic and orphaned microservices can leave behind empty consumer groups. AegisMesh monitors partition lag in real time. It automatically identifies Zombie Consumer Groups—groups with massive accumulated lag but zero active members—allowing operators to clean up resources before thread starvation occurs.