There are several tracing tools available publicly and privately for looking into execution time of GraphQL queries. These are super useful and key for integrators writing queries. But as API owners we also need to know how our API is being used. This is important for making deprecations, for choosing where we might want to optimise, understanding general usage patterns and even answering support queries. It's also key to not slow down execution gathering that data. To that end GitHub has several tools and relies heavily on data pipelines. We track parse and execution errors, which objects API queries touch, and more. All of this is beneficial when ensuring the smooth running of the API. This talk will focus on our experiences doing this, what data we've found useful, and how you might want to approach this in your APIs.
Enjoying these posts? Subscribe for moreJoin
Already have an account? Sign in