Avoiding Disaster: Practical Strategies for Error Handling

Transcript

1. AVOIDING DISASTER: PRACTICAL STRATEGIES FOR ERROR HANDLING Huy Du | exponentdev.com RubyConfTH 7th October, 2023
2. Huy Du Software Engineer @ @dugiahuy https://exponentdev.com
3. PROGRAMMING f(x) = y • f() is solution • x is input • y is expected output Programming, in simple term, involves solving problem by providing instructions to a computer.
4. WHAT ARE ERRORS? Software errors are defined as results in unexpected outcomes from a computer program or cause it to act in ways that were not intended.
5. EXPECTED ERROR These are errors that developers anticipate and expect as part of normal program execution • Part of business logic ⚬ Ticket has been sold out. ⚬ Purchase ticket cannot process after midnight. • External dependencies ⚬ Wrong library usage ⚬ Connection error
6. UNEXPECTED ERROR Errors are not expected to occur. They are interrupted our program or also known as bugs. • Database corrupted • Memory exceeded • Developer mistake
7. EXCEPTION IN RUBY Exception is a unique data structure representing an error or unexpected condition in the Ruby program. Exception objects carry information about: • Type • Message • Backtrace
8. EXCEPTION IN RUBY Exception is a unique data structure representing an error or unexpected condition in the Ruby program. Exception objects carry information about: • Type • Message • Backtrace
9. EXCEPTION IN RUBY Exception is a unique data structure representing an error or unexpected condition in the Ruby program. Exception objects carry information about: • Type • Message • Backtrace
10. As a User, I want to buy a Ticket for attending Ruby Conference. EXAMPLE PROBLEM
11. User Frontend API Server Database Payment Gateway Infrastructure HIGH LEVEL SYSTEM DESIGN Buy a ticket Send API request Interact with data Submit payment
12. Model Controller Database Payment Gateway APPLICATION COMPONENTS Application
13. NAIVE IMPLEMENTATION Ticket not found? User not found? Order created? User buy ticket NO NO
14. NAIVE IMPLEMENTATION Ticket not found? User not found? Order created? User buy ticket Insufficient balance? Payment success? NO NO YES YES Render success YES
15. NAIVE IMPLEMENTATION Ticket not found? User not found? Order created? User buy ticket Insufficient balance? Render Error Raise Exception Raise Exception Render Error Payment success? YES YES NO NO NO YES YES Render success Render Error Raise Exception YES Render Error NO NO Raise Exception Render Error
16. Can we make it better? NAIVE IMPLEMENTATION • Chaotic • Error Prone • Poor extendibility • Poor readability • Not reusable
17. Model Controller Database Payment Gateway BUSINESS LOGIC LAYER Application Business Logic
18. CREATE BUSINESS LOGIC SERVICES
19. APPLY TO CONTROLLER
20. CONTROLLER REFACTORING User buy ticket Create Order failed? Render Error Checkout Order failed? NO Insufficient balance? Render Error YES NO YES YES NO Payment failed? YES Render Error NO Render success
21. REFACTORING ISSUES User buy ticket Create Order failed? Render Error Checkout Order failed? NO Insufficient balance? Render Error YES NO YES YES NO Payment failed? YES Render Error NO Render success Leaky error details • Return inconsistent value • No standardize
22. How to resolve it? REFACTORING • Better code flow • Better extendibility • Better readability • Not reusable • Poor encapsulation
23. RETURN MONADS AS RESULT
24. MONADS Monads is structure that wrap return value of function in monadic way
25. MONAD RESULT Result Success, Value Failure, Error
26. MONAD RESULT IN RUBY
27. SERVICE OBJECT Service Object are Ruby objects, that are designed to execute one single action in the business logic, and do it well.
28. SERVICE OBJECT
29. MONADIC HANDLING Monadic handling is the combine between Service Object and Monads Result. Monadic handling involves using Monad Result to manage the flow of execution within the Service Object
30. MONADIC HANDLING
31. APPLY MONADIC HANDLING Monadic Handling Refactoring
32. APPLY MONADIC HANDLING Monadic Handling Refactoring
33. APPLY MONADIC HANDLING Naive Implementation Monadic Handling
34. APPLY MONADIC HANDLING User buy ticket Create Order failed? Render Error Checkout Order failed? NO Render success YES NO YES Render Error
35. MONADIC HANDLING • Provide standardize and consistency way error handling • Explicit and reduce complexity of conditional cases • Making code more robust and maintainable • Completely encapsulate logic and reusable everywhere
36. ERROR HANDLING f(x) = y Input value x must meet valid criteria to resolve problem. • f() is solution • x is input • y is expected output Solution f() is expected to produce final result without any issues.
37. VALIDITY VS EXECUTION f(x) = y Validity errors pertain to issues related to the input x • f() is solution • x is input • y is expected output Execution errors refer to issues that arise during execution of f() with the valid input x
38. BEFORE RAISING EXCEPTION • Is this action truly unexpected? • Am I going to disrupt the execution flow? • Will I have a heart attack when this exception thrown?
39. Huy Du ”The correct way of error handling will stop your lifespan for a couple of seconds."
40. WRAP UP • Understand different type of errors in software application • Understand how exception works in Ruby • Create boundaries for Business Logic layer and encapsulate with Service Object • Use monadic error handling by combining Service Object and Monad Result • You have the practical strategies for error handling
42. Huy Du "Pray is not prevented your system from crash. You need monadic handling."
43. SALUTE! Huy Du https://exponentdev.com/talks https://github.com/dugiahuy/talks