This is a discussion topic for the original post at How to prevent your automations from breaking in RPA - BotCity | Python RPA | Blog
In RPA, we know that automations will stop working. It will happen; it is a fact. Learn about the leading causes, how to avoid them, and make your bots more resilient 
In companies, we see that essential and complex processes are 100% automated. If they stop working, they can significantly impact the business areas, which are no longer prepared to meet these demands manually.
📌 Learn about the leading causes of breakage in robots:
1. Application Updates: RPA bots interact with systems that, if they undergo changes in their interfaces, may fail;
2. Environment Problems: Bots can be sensitive to server and network configurations, resolution, fonts, positioning, limits on systems capacity, and so on;
3. Exceptions and Errors: When encountering exceptions or errors, bots can fail if not handled correctly;
4. Data Variability: If the input or output data format changes, the bot logic or processing may fail;
5. Process Changes: If processes undergo significant changes, the bot logic may become incompatible;
6. Licensing and Expiration: RPA software often has licenses or subscriptions, which, if they become invalid, the bot will stop working;
7. Human Errors and Attacks: Errors in configuration, logic or development, or security attacks can cause failures;
✅ To avoid problems and make your automations more resilient, follow these practices:
Error and exception handling: Implement comprehensive mechanisms, including potential scenarios and actions, such as retries or alternate paths;
2. Automate by IDs / XPath and use APIs (if any): interfaces change frequently, and computer vision is sensitive to changes in resolution, fonts, and so on;
3. Monitoring: use a tool that provides real-time indicators, complete logs, and preventive notifications to anticipate problems, interface failures, bottlenecks, and resource escalation;
4. Scalability: Make sure the infrastructure can handle the increased workload and can scale quickly to accommodate more extensive processes or transaction volumes;
5. Maintenance, Updates, and Security: Perform regular maintenance and updates to your RPAs, ensuring compatibility, license and certificate renewals, more secure versions, and application of best security practices;
6. Testing: validate changes before they go into production, do automated, unit, and integration testing, and release incrementally;
7. Training and documentation: Train your teams regularly and ensure complete and accessible documentation;
8. Stability of the environment: Avoid unnecessary changes and identify dependencies that may impact it. Coordinate with teams to minimize unavailability;
9. Version control and Backup: Implement a backup and recovery strategy and make sure it works. Use versioning to control changes and rollback;
10. Sustainment and Improvement: have people responsible for monitoring, preventing, acting on problems with the bots, and making revisions;