All Products
Search
Document Center

ApsaraMQ for Kafka:Inspection

Last Updated:Mar 11, 2026

ApsaraMQ for Kafka uses HouseKeeping to regularly monitor cluster health. HouseKeeping checks messaging, data integrity, and process stability. It alerts the ApsaraMQ for Kafka technical team when issues arise.

How it works

HouseKeeping performs three types of automated checks:

Messaging verification

HouseKeeping sends and receives messages on internal topics to validate messaging within seconds. When issues occur, the system immediately alerts the ApsaraMQ for Kafka technical team. Detected issues include:

  • Messaging failures

  • Increased latency

  • Abnormal throughput

  • Connection exceptions in dependent components

Strict topic offset and metadata verification

HouseKeeping verifies information such as topic offsets and metadata to make sure that in-memory data remains correct and consistent. This prevents issues caused by data inconsistency.

Process monitoring

HouseKeeping monitors the runtime health of ApsaraMQ for Kafka processes. When a critical failure is detected, it automatically restarts ApsaraMQ for Kafka processes. Detected issues include:

  • I/O hangs

  • Thread deadlocks

  • JVM crashes

  • Connection flooding attacks

  • Memory leaks

Usage notes

Internal topics

HouseKeeping uses two internal topics:

  • __alikafka_housekeeping_local_topic

  • __alikafka_housekeeping_cloud_topic

Important
  • These topics are automatically created when you deploy an ApsaraMQ for Kafka instance.

  • These topics can only be deleted after you delete the corresponding instance.

Resource consumption

HouseKeeping consumes a small amount of cluster resources. The following estimates apply to the smallest specification:

ResourceConsumption per topic
Traffic (production and consumption)Approximately 100 bits/s
StorageApproximately 10 MB

Bandwidth consumption scales with cluster specification. Larger clusters consume more inspection bandwidth.