Query Analysis Reboot

This talk will focus on using data to drive decisions; or how your users can help you better understand your own API by leveraging data pipelines and tracing tools.

Transcript

1. October 2019 Claire Knight / GraphQL Summit Query Analysis Reboot
2. Claire Knight / GraphQL Summit Query Analysis Reboot Using Data to Drive Decisions Or how your users can better help you understand your API!
3. Data? But I have the query... There are several things you need to consider from an operational point of view! Claire Knight / GraphQL Summit Query Analysis Reboot
4. Claire Knight / GraphQL Summit Query Analysis Reboot Query Performance
5. Claire Knight / GraphQL Summit Query Analysis Reboot Power of Analysis
6. Claire Knight / GraphQL Summit Query Analysis Reboot Ensuring Compliance - GDPR, PII
7. Claire Knight / GraphQL Summit Query Analysis Reboot Existing Solutions
8. Claire Knight / GraphQL Summit Query Analysis Reboot Our Data
9. Claire Knight / GraphQL Summit Query Analysis Reboot Domain Objects Being Used Parsing Problems Resolvers Unable to Complete Validation & Security Errors Things we have found valuable...
10. Claire Knight / GraphQL Summit Query Analysis Reboot Capturing It
11. Claire Knight / GraphQL Summit Query Analysis Reboot Applying Security During Parsing Within the Resolvers Before Sending A Response Where we capture the data...
12. Claire Knight / GraphQL Summit Query Analysis Reboot Statistics Capturing Systems Data Pipeline Log Files Where does it go?
13. Claire Knight / GraphQL Summit Query Analysis Reboot Insights Gained
14. Claire Knight / GraphQL Summit Query Analysis Reboot Key Learnings • Being able to learn more runtime errors and exceptions separately from validation or parsing errors to more easily identify bug fixes • Tracking expected errors (parsing for example) helps educate SRE/management that not all errors are bad • An idea of query size even without the query helps with capacity planning and timeout concerns
15. Claire Knight / GraphQL Summit Query Analysis Reboot For a (Public) API, or one of scale, it’s important to consider queries and data from the API owners perspective and not just as an integrator. This might lead to better ways to return safe and secure hints to users. But takes time, data, and delayed analysis.
16. Questions? Comments? krider2010@github.com @krider2010 @krider2010