What is this article about?

This article explains how to implement graceful shutdown for Go HTTP servers deployed on Kubernetes to achieve zero-downtime rolling updates. It details the configuration of liveness and readiness probes to manage traffic during updates.

Who would benefit from reading this article?

This article is beneficial for Go developers and DevOps engineers who deploy Go HTTP services on Kubernetes and want to ensure their deployments have minimal to zero downtime during rolling updates. It's particularly useful for those struggling with proper probe configurations.

How does this solution address downtime during Kubernetes rolling updates?

Unlike a basic HTTP health check, this solution differentiates between readiness and liveness using distinct probe types. It ensures that old pods gracefully stop accepting new traffic while finishing in-flight requests before being terminated, preventing connection failures during Kubernetes rolling updates.

When should I implement graceful shutdown as described?

You should implement graceful shutdown and separate readiness/liveness probes when deploying Go HTTP applications to Kubernetes, especially in production environments where zero downtime and seamless user experience during updates are critical.

What's a common issue discussed regarding liveness probes and distroless containers?

A common issue highlighted is that distroless containers often lack basic utilities like 'cat', which is frequently used for file-based liveness checks. The article suggests writing a small Go sub-command within the application to perform the liveness check instead.

medium.com · 11 JAN '23

Graceful shutdown with Go http servers and Kubernetes rolling updates

Item: Graceful shutdown with Go http servers and Kubernetes rolling updates
Rating: 5
Author: Simon Frey

This article is a nice read on how to properly shut down Go http servers in Kubernetes. It covers configuring readiness and liveness probes to ensure zero-downtime rolling updates, even addressing issues with distroless containers.

Visit medium.com →

Questions & Answers

What is this article about?: This article explains how to implement graceful shutdown for Go HTTP servers deployed on Kubernetes to achieve zero-downtime rolling updates. It details the configuration of liveness and readiness probes to manage traffic during updates.
Who would benefit from reading this article?: This article is beneficial for Go developers and DevOps engineers who deploy Go HTTP services on Kubernetes and want to ensure their deployments have minimal to zero downtime during rolling updates. It's particularly useful for those struggling with proper probe configurations.
How does this solution address downtime during Kubernetes rolling updates?: Unlike a basic HTTP health check, this solution differentiates between readiness and liveness using distinct probe types. It ensures that old pods gracefully stop accepting new traffic while finishing in-flight requests before being terminated, preventing connection failures during Kubernetes rolling updates.
When should I implement graceful shutdown as described?: You should implement graceful shutdown and separate readiness/liveness probes when deploying Go HTTP applications to Kubernetes, especially in production environments where zero downtime and seamless user experience during updates are critical.
What's a common issue discussed regarding liveness probes and distroless containers?: A common issue highlighted is that distroless containers often lack basic utilities like 'cat', which is frequently used for file-based liveness checks. The article suggests writing a small Go sub-command within the application to perform the liveness check instead.