We don't have QA which means quality is dependent only on us. Use the following list to verify that the end result that is being pushed to production is in high quality
✔ Are we going to break backward compatibility? If so then raise a flag
✔ What are the cost estimations of the feature?
✔ Error handling (exceptions, timeouts, retries) when using streams as lambda inputs
✔ DynamoDB Resources
- How many indexes are required?
- How will an index affect the cost?
✔ Lambda roles
- Who is allowed to access the Lambda
✔ Code + scenarios are covered by testing
✔ Integration tests pass
✔ If code change depends on tracer input then check all tracers to work with the new code
✔ Your feature works on your AWS environment
✔ Code review
✔ The deployment was successful on CircleCI
✔ No errors/warnings appearing in the changed module on logz.io
✔ Verify it works in production by login into our customers
✔ DDB Metrics
- Did the Write Units increase?
- Did the Read Units increase?
- How does it affect the cost?