Answer.

Manual testing during the maintenance phase is testing an already existing and functioning system when making enhancements, fixing bugs, or integrating new components.

Background

Previously, maintenance was performed on a residual basis, often testing only the new features. However, over time it became clear that any interference could affect already functioning scenarios.

Problem

A typical situation is:

Local changes are made, but their impact on existing functionality is often underestimated.
Regression occurs in seemingly unrelated modules.
Lack of a systematic approach increases the risk of unexpected "breakdowns" in production.

Solution

Effective organization of maintenance testing requires:

Identifying and constantly updating a "set of key scenarios" that are checked with each enhancement.
Using checklists and regression maps.
Including exploratory testing to find unexpected effects from changes.

Key features:

Quick response to small changes with minimal rollback.
Focus on real user scenarios that may be indirectly affected.
Flexibility in choosing methods: from checklists to creative exploratory testing.

Trick Questions.

Is it necessary to test only those modules that have been modified?

No, it is essential to test related parts of the system to avoid missing side effects from changes.

Is full regression testing sufficient after each fix?

No, often it's enough to check key (critical) paths, while complete regression is only performed before a release or with significant changes.

Can we completely abandon exploratory testing during the maintenance phase?

No, exploratory testing allows finding non-trivial bugs outside the script coverage and should accompany the maintenance phase.

Common mistakes and anti-patterns

Neglecting related modules: testing only the "patched" area.
Lack of an up-to-date regression test set.
Ignoring understanding of architecture slows down risk zone identification.

Real-life examples

Negative case

After fixing a bug in the user profile, only the profile is tested, but authorization and profile display on other pages are not checked. As a result, a bug emerges: the profile does not update on the main page.

Pros:

Quick completion of testing a specific task.

Cons:

Missing bugs in related sections.
Decreased trust in QA and the product.

Positive case

The fixed bug in the profile is tested both separately and comprehensively: the profile is checked everywhere it's used. A checklist of key scenarios is utilized.