I think Windows Azure SQL Database , which used to called SQL Azure, and internally we called CloudDB is an amazing service that it was the first multi tenant cloud based SQL database. As you can image, it is super complex under the hook, and we sometime thought the architect is not suitable for modern service. However, running such complex service is sometime fun and frustrated (when you are on called for one week, and been called up 3 or 4 time per night). it always has big challenge and you can learn lot from it. In this blog series, I will write a couple of ideas focusing on how we testing SQL Azure. Note, most of the ideas are not mine, I just have pleasure to work with many people who invent and implemented them. Hope this can help your team to better testing your service.
Here is topics I plan to write and I am pretty sure that I will not follow the order to write my blogs since some topics are much hard to write than others.
- Introduction to Windows Azure SQL Database
- Our test environment Onebox, Multibox, Test Cluster and Dogfood
- How to do functional testing?
- FailOver Stress: you love it when it opens bug, you hate it when it always failed?
- Redline Stress and Failure Injector: our quality gate to ship into production
- Upgrade Perf Testing: My first metric driven improvement
- SQL Azure Dashboard: Testing is monitoring, monitoring is testing
- External Monitoring: the good, the bad
- Dogfood: the best test environment I have ever seen
- Thought on data collection and data pipeline
- Lessons learned from running a data pipeline
- Thought on alert strategies
- Lesson learned from all feature switch off
- Case Study “Issue Mitigation and Exposition Control: tens of dumps in production is not so scary as you thought”
- Case Study “NewT: network emulator: it is much easy to repro hard problem”
- Case Study “Ship faster does matter”
- Livesite: your only source to learn, but you may hate it.