🗓️ 03092024 1153
📎 #resilience #microservices

health_live_readiness_checks

Health Check

General term that can encompass both liveness and readiness checks

Determine whether a service or application is functioning as expected
Used to monitor various aspects of an application
- Example
  - Ability to respond to requests
  - Connection to a database
  - Availability of necessary resources.
Monitor the overall health status of the application or service.

Used to determine if an application is still running and should not be restarted

Answers the question: "Is the application alive or dead?"

Determine if the application is running correctly and not stuck in a deadlock or unrecoverable state

NOTE

If a liveness check fails, the container orchestrator (like Kubernetes) will kill the container and attempt to restart it

Useful for detecting if
- An application has crashed
- Stuck in a bad state (e.g., deadlock or infinite loop)

Determines if an application is ready to start accepting traffic

Crucial for ensuring:
- Application is fully initialized
- Has its dependencies ready before starting to serve requests

NOTE

If this check fails, the container orchestrator will not route traffic to that instance, and it will also not restart the container

Useful for scenarios where an application needs time to:
- Initialize
- Wait for dependencies to be ready

Aspect	Health Check	Liveness Check	Readiness Check
Purpose	Monitor overall health status.	Check if the application is alive and should not be restarted.	Check if the application is ready to accept traffic.
Action on Failure	Could trigger alerts or corrective actions depending on configuration.	Container is killed and restarted by orchestrator (e.g., Kubernetes).	Traffic is not routed to the instance until it passes readiness checks.
Example Condition	Service is up, connected to DB, and dependencies are healthy.	Application is not deadlocked or crashed.	Application has fully started and all dependencies are ready.
Typical Use Cases	Monitoring dashboards, alerts, CI/CD pipelines.	Detecting crashed containers, unresponsive services.	Rolling updates, scaling operations, ensuring smooth deployments.