01Understand the current system
Before changing an application, review its architecture, repository, deployment process, dependencies, data flows, known issues and existing documentation. Clarify ownership of hosting, credentials, third-party services and production access. This reduces the risk of making changes without understanding their effects.
02Separate urgent fixes from planned improvements
Classify issues by impact and urgency, reproduce problems where possible and agree on acceptance criteria for fixes. Keep feature requests and refactoring visible alongside defects so that short-term repairs do not silently create a growing maintenance burden.
03Protect release quality
Use a repeatable workflow for code review, automated checks, testing, environment configuration and release notes appropriate to the application. Backups, rollback plans and monitoring should be considered according to the system’s data and operational importance.
04Make support expectations explicit
Document communication channels, coverage, response expectations, escalation paths, included maintenance and out-of-scope work. Do not assume 24/7 availability or a guaranteed response time unless it is explicitly agreed in writing.