Liveness Probe
System Analysis
Normal Behavior
An automated mechanism to regularly verify if a system, service, or specific endpoint is operational.
Failure Behavior
Returns HTTP 200 OK while the underlying database has been fully deleted.
Business Consequence
Monitoring dashboards look perfectly green while actual users cannot log in or perform any actions.
Visual Manifestation
"A brightly glowing green checkmark next to a totally blank application screen."
Satirical Behavior
"A simple ping endpoint that successfully returns '200 OK' while the actual database it relies on has been on fire for the last 45 minutes."
Known Aliases
Technical Terminology
Failure Indicators
System Architecture (Graph)
Used By (Characters)
FAQ
How does it normally behave?
An automated mechanism to regularly verify if a system, service, or specific endpoint is operational.
How does it fail?
Returns HTTP 200 OK while the underlying database has been fully deleted.
What is the business consequence?
Monitoring dashboards look perfectly green while actual users cannot log in or perform any actions.
Why does performing deep synchronous database queries in Kubernetes liveness probes cause catastrophic cascading cluster restarts?
A liveness probe exists solely to detect unrecoverable process deadlocks requiring a container restart (SIGKILL). If a liveness probe synchronously queries an external database, a transient database query spike or network blip causes every replica's liveness check to fail simultaneously. Kubernetes immediately kills and restarts all service pods at once, overwhelming the database with bootup connection surges and turning a minor 5-second database slowdown into an unrecoverable cluster-wide outage.
How should modern microservices decouple readiness checks from liveness checks to achieve graceful zero-downtime deployments?
Liveness probes should only verify internal process state (e.g., checking if event loops are cycling or memory is below fatal limits) without evaluating external dependencies. Readiness probes should evaluate critical dependencies (e.g., cache connectivity and schema migration state) and flip to unhealthy to temporarily remove the pod from the load balancer rotation without triggering a process restart, allowing in-flight requests to drain cleanly.
Explore the system
AI Summary
Health Check Service is a OBSERVABILITY system in TinyCTO.tv. Dispatches lightweight periodic synthetic probes to microservice endpoints, verifying internal state flags and non-blocking I/O loops within tens of milliseconds to validate cluster readiness.