Reliability

The Reliability pillar focuses on making your supergraph resilient to change. That means ensuring it's appropriately resourced, observable, and generally production-ready. These pillar's best practices build confidence in your supergraph's availability for graph consumers. They also ensure your graph is a trusted source of data in your organization.

Principles

Understand graph usage

This principle's practices promote an understanding of how your system is being used. Knowing which clients use which fields and how their operations use upstream resources is crucial for ensuring system reliability. This understanding lets you make informed decisions about optimizing and improving reliability.

Enable effective testing in the supergraph

Graph contributors should test their changes in the supergraph. To do so, they need clear guidance on testing best practices for both subgraphs and entity resolvers.

End-to-end testing of the supergraph is also crucial for ensuring reliability. Providing clear instructions on how to test changes across supergraph components helps catch potential issues before they impact users.

Use compute resources efficiently

When using the GraphOS Router and subgraph services, balancing performance with efficiency is key. By monitoring pod metrics and graph execution metrics, you can identify optimization opportunities and maintain efficient system performance.

Catch problems before they go into production

One of the best ways to avoid production outages is to catch potential problems early in the development lifecycle. This approach lets you address issues before they impact production, ensuring system stability. These proactive measures can save time and resources and help avoid costly downtime and unhappy customers.

Identify issues in production

Lowering mean-time-to-recover (MTTR) in your supergraph is crucial for maintaining the system reliability. By proactively monitoring and addressing issues, you can minimize the impact of downtime on your users. Additionally, tracking and analyzing metrics such as error rates and response times can identify potential issues before they become critical problems.

Minimize risks in production

Making the right infrastructure and process choices is key to preventing production outages. It's important to consider any potential issues and take steps to prevent them. Best practices include:

• Performing risk assessments
• Conducting regular audits
• Staying up-to-date with the latest industry trends and best practices

Teams can minimize the risk of production outages by being proactive about infrastructure and process choices.