Think Twice Code Once with Automations
Automation tests were given low priority by developers unless there is a strict process. It should be a habit to complete automation with implementation. Engineers are keen to provide the solution; They bang their heads for an efficient solution. Then they are ready to cheer upon their achievement. Until the Good Automation Test generates a report, the solution is still pending.
The word Good Automation represents many other hidden attributes. Automation must have Fast enough, Consistent, Decent Coverage, Alert and Reporting Metrics. This code piece checks the heart of the business code; Test is as important as implementation. Write enough tests on different Test groups; Unit test, Sanity test and Regression test.
Think about corner cases before coding. Data validation is a wall of functional code. In exceptional scenarios, Data is the primary citizen protected first.
These three steps are simple to safeguard from data corruption.
- Identification
- Notification
- Auto Correction
Identify exceptional scenarios early and code to notify the failures. Fix data using self-healing on the subsequent requests or involve a background automation system.
Microservice is the talk of the town in recent years; Data out of Sync is another big problem among multi-system architecture. Microservice is designed to avoid single-point failures. But Simple improper exception handling may lead to data Out-of-Sync issues. This Nightmare is called the data issue.
Data Auto-correction and True-up are other areas that need more focus. Application nodes are brought back on other node failures. Database replicas and Container orchestration are giving the ability for systems to high availability. These designs focus on Infrastructure and Application, not on Functional. Functionality failure because of data is different from Application failure. The issue is addressed only on proper Data correction.
Treating customer data is like doing surgery; A simple miss calculation could disturb customer business. It is not a good practice to fix the data at the back-end. This bad process involves high maintenance costs. Automation system for data correction and testing must be established to keep the better business continuation.
