Health checks in containers in Microservices - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Purpose of Liveness and Readiness Probes

In container orchestration systems like Kubernetes, what is the main difference between a liveness probe and a readiness probe?

A Liveness probes check if the container is alive and should be restarted if failed; readiness probes check if the container is ready to accept traffic.

B Liveness probes check network connectivity; readiness probes check disk space availability.

C Liveness probes verify CPU usage; readiness probes verify memory usage.

D Liveness probes are used only during startup; readiness probes are used only during shutdown.

💡 Hint

Think about what happens when a container is unhealthy versus when it is not ready to serve requests.

Liveness probes detect if a container is stuck or dead and need restarting. Readiness probes determine if a container can receive traffic. They serve different purposes in container lifecycle management.

❓ Architecture

intermediate

Designing Health Checks for a Microservice

You have a microservice that depends on a database and a cache. Which health check design best ensures the service is only marked ready when both dependencies are available?

A Use no health checks and rely on external monitoring tools.

B Implement a readiness probe that checks connectivity to both the database and cache before marking the service ready.

C Use only a liveness probe that checks if the service process is running.

D Implement a readiness probe that checks only the database connectivity.

💡 Hint

Consider what readiness means for a service that depends on multiple components.

Readiness probes should verify all critical dependencies are available before the service receives traffic. Checking only the process or partial dependencies risks sending requests to a non-functional service.

❓ scaling

advanced

Impact of Health Checks on Auto-Scaling

How can improperly configured health checks affect the auto-scaling behavior of a containerized microservice?

A If health checks fail frequently, the orchestrator may kill and restart containers unnecessarily, causing scaling instability.

B Health checks have no effect on auto-scaling since scaling depends only on CPU usage metrics.

C Health checks speed up scaling by automatically increasing container memory limits.

D Improper health checks cause the orchestrator to ignore scaling policies and keep the number of containers fixed.

💡 Hint

Think about what happens when containers are repeatedly marked unhealthy.

Failing health checks cause container restarts, which can interfere with scaling decisions and cause instability. Proper health checks help maintain stable scaling behavior.

❓ tradeoff

advanced

Tradeoffs in Health Check Frequency

What is a key tradeoff when choosing how often to run health checks on containers?

A Health check frequency does not affect system performance or reliability.

B Less frequent health checks reduce resource usage but cause containers to restart unnecessarily.

C Frequent health checks improve container startup time but reduce security.

D More frequent health checks detect failures faster but increase resource usage and network traffic.

💡 Hint

Consider the balance between responsiveness and overhead.

Running health checks too often uses more CPU and network resources, but helps detect problems quickly. Running them too rarely saves resources but delays failure detection.

❓ estimation

expert

Estimating Health Check Load in a Large Cluster

You manage a cluster with 1000 containers. Each container runs a health check every 10 seconds, and each check sends a 1 KB request and receives a 1 KB response. Estimate the total network bandwidth used by health checks per minute.

A Approximately 2000 KB per minute

B Approximately 200 KB per minute

C Approximately 20000 KB per minute

D Approximately 200000 KB per minute

💡 Hint

Calculate requests per minute and multiply by request and response size.

Each container sends 6 checks per minute (60/10). Each check uses 2 KB (1 KB request + 1 KB response). Total = 1000 containers * 6 checks * 2 KB = 12000 KB = 12 MB per minute. Closest option is 20000 KB (20 MB) which accounts for overhead and rounding.