One of the goals of testing is to provide confidence for the release of a feature or product. But do you *really* know your if feature will work well in production? *Really* really? And how do you *know* you know? Too often, we do all our testing in sandboxes and then cross our fingers, hoping everything will work out fine on release. But with current monitoring tools available, there is a better way!
This talk will draw on my experiences adding tracing to production services; deciding what to track, how to display and analyze data, and using that analysis to do testing in the actual production environment. With the right monitoring in place it is possible to run realistic load tests, find issues that might have been missed in lower environments, and have greater confidence in the impact your change will have on your actual real-world users.
- What monitoring and logging tools are available
- How to decide what to track
- Ways to create meaningful reports with monitoring results
- How to use production data to create more realistic tests
- Strategies for using real-time monitoring to run tests in production