Health check

As explained in Machine Learning Platform Engineering

A diagnostic test that determines whether a service or execution unit is functioning and able to handle requests.

A health check is an endpoint-based diagnostic that reports whether a deployed service or pod is functioning and able to handle requests. In Kubernetes environments, it supports observability, pod management, and troubleshooting. section 10.4.1

Distinguish the related checks by the operational decision they support:

• Liveness probes call /healthz periodically to determine whether the pod is alive and able to run its main process; failure may cause Kubernetes to restart it. section 10.4.1
• Readiness probes call /readyz periodically to determine whether the pod is operational and ready to receive traffic. section 10.4.1
• BentoML provides built-in health checks and liveness/readiness probe support, reducing the endpoint implementation required for a deployed service. section 10.4.1

Diagnosing a pod that will not initialize

Configure Kubernetes to inspect /healthz for liveness and /readyz for readiness. A failing /healthz indicates an initialization or process problem to investigate; a pod that is alive but not ready should not receive traffic. section 10.4.1

Figure 10.7 BentoML’s built-in observability endpoints for K8s environments

Inspect the BentoML observability endpoints supplied for Kubernetes environments. section 10.4.1

How Machine Learning Platform Engineering covers it

How deep it goes

Defines health, liveness, and readiness roles with endpoint behavior, but does not detail custom implementation code.

Examples

Uses /healthz and /readyz endpoints, pod initialization failures, BentoML services, and Kubernetes probe behavior.

What it assumes

Assumes familiarity with deployed services and pods while introducing health, liveness, and readiness distinctions.

What you can do after

Operate/configure deployed services with liveness and readiness monitoring in Kubernetes environments.

How it connects

Connects health checks with Kubernetes, liveness probes, readiness probes, observability, and Prometheus metrics.

liveBookmanning.com

Read Machine Learning Platform Engineering on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Learn Docker in a Month of Lunches, Second EditionA health check is an automated test that determines whether an application is functioning correctly rather than merely having a running process.
• 

  liveBookmanning.com
  Quarkus in ActionA health check is an evaluation that reports whether an application or one of its dependencies is functioning correctly, typically with a binary outcome.
• 

  liveBookmanning.com
  Amazon Web Services in Action, Third Edition: An in-depth guide to AWSA health check is an automated test that determines whether a website or backend target is able to receive traffic and provide an expected response.
• 

  liveBookmanning.com
  Build and Orchestrator in Go (From Scratch)A health check is an operation that evaluates whether a running task is functioning normally, typically by calling an endpoint and interpreting its response.
• 

  liveBookmanning.com
  Build Financial Software with Generative AI (From Scratch)A health check is an automated test or status evaluation that determines whether an application or runtime component is functioning acceptably.
• 

  liveBookmanning.com
  Bootstrapping Microservices, Second Edition: With Docker, Kubernetes, GitHub Actions, and TerraformA health check is an automated assessment of whether a running service or workload is functioning and available according to defined criteria.
• 

  liveBookmanning.com
  Kubernetes for DevelopersA health check is an operational test or signal indicating whether an application or process is functioning correctly.
• 

  liveBookmanning.com
  Building Distributed Applications that WorkA health check is an automated assessment of whether an application or dependency is ready and functioning sufficiently for interaction.

Compare how 9 books explain health check →

Related concepts

• liveness checkincludes · A liveness check is an automated health test that determines whether a running workload remains functional or should be restarted.
• readiness checkincludes · A readiness check is a health probe that determines whether a running service or workload is prepared to receive requests or traffic.
• auto-scaling groupused by · An auto-scaling group is a managed collection of compute instances whose capacity is automatically adjusted according to configured policies, conditions, or workload demand.
• uptimemeasures · Uptime is the duration or proportion of a specified period during which a system or service is operational and available to provide its intended function.
• Docker Swarm manager nodeused by · A Docker Swarm manager node coordinates the cluster, schedules services and tasks, maintains swarm state, and directs worker nodes.

This explainer was generated by Manning's knowledge graph from the text of Machine Learning Platform Engineering. The section links point to the parts of the book it draws on.