Skip to main content

How do I see what my tests cover with the Coverage Map?

Compute a Coverage Map from your runs, read which screens and controls your tests covered, and find the gaps.

Written by Ines

The Coverage Map shows which screens and controls of your application your Test Runs reached and used, and which ones no test used. You choose the runs, Thunders computes the map, and you read it as a tree of Apps, sections and screens, each with the share of interactions your tests covered. Use it to find the gaps before a release, and to see what a Test Set or a Test Case really exercises.

How the map is organised

The map is a tree of cards, from the whole Test Project down to single screens:

  • The Test Project card adds up all the Apps of the map.

  • An App card stands for one platform, such as Web, Android or iOS. A web Test Project usually has one App.

  • A section groups the screens that share a domain and the first part of their path, for example every screen under app.example.com/orders. A section that holds one screen shows that screen in its place.

  • A screen is one page of your application. Thunders replaces the ids in the address, so every order page is one screen, /orders/:id. On a Mobile App, a screen is a native screen, named by its title.

  • The Shared layout card holds the controls that appear on most screens of a site, such as the header and the menu. They count once, on this card, and not on each screen. A control goes there when it is on at least three measured screens and on at least half of them. A test that uses it on any screen covers it, and the card lists the screens it appears on.

Under a screen, the map can also show:

  • Next screens: the screens that a test reached from this one.

  • An unvisited screen: a link on this screen that opens a screen of the same App that no test reached. It shows where the tests stop.

  • Exits: links to another domain. The map does not follow them, so they add no screen and no percentage.

When a test follows a link to another domain, such as the sign-in page of an identity provider, the map keeps that domain apart. It hides it by default, and its screens do not count in the percentages of the App and the Test Project. Select Include other domains, next to Search, to show these domains and count them again.

A screen is measured when Thunders recorded its controls during a run, and only measured screens count in a percentage. A screen that a test reached without a record of its controls shows Visited by the number of tests, and counts in no percentage.

What the map measures

  • An interaction is one way to use a control. A button or a link has one (Activated). A text field has two (Filled, Left empty). A checkbox has two (Turned on, Turned off). A dropdown or a group of radio buttons has two (First value, Second value), which are two different values. A dropdown that a test leaves on its default value, before it goes on with the form, counts as its first value.

  • A control is covered when all its interactions are tested, and partly covered when some of them are. A control counts even when the step that used it failed.

  • The percentages are Interaction coverage: the interactions tested out of the interactions present, added up over every measured screen of the card. Hover a percentage to read how it is computed.

  • A control repeated on the rows of a list shows Repeated on the number of rows. It counts once, and a test on any row covers it.

  • A control that a test only checks with an assertion shows Checked by a test. A check is not an interaction, so it does not change the percentage.

  • A control inside a dialog shows the name of that dialog, for example In dialog "Add API Step".

  • Every percentage is an upper bound. A screen that a test visited but the map could not measure, and an unvisited screen, can hold more controls. Real coverage is equal or lower.

Select What counts as covered on any map to read the full rules.

Create a Coverage Map

  1. Open the Test Project and select Coverage Map in the left-hand navigation.

  2. Select New Coverage Map.

  3. Under Runs to read, choose one option:

    • Test set runs reads the exact runs you pick, which suits one release. Narrow the list with Test set and From, or type in Search a test set, then tick the runs. A map reads one Environment, so the first run you tick sets it, and the runs of other Environments are greyed out.

    • A period reads every finished run of one Environment between a Start date and an End date. A period can mix versions of your application, so a changed menu then counts as two menus.

  4. Select Compute map.

The map reads the runs of the Branch you are on. It opens at once and fills in as each screen finishes, and you can leave the page while it computes. If a map already exists for the same runs, Thunders opens that map instead of computing a new one. A Test Project computes up to two maps at the same time.

In Test set runs, Compute map stays disabled until the Test Project has a Test Set with a finished run. Users with the Read-only role can open every map, but they cannot create one.

Read the map

  • The Whole map pill in the header gives the interaction coverage of the whole map, for example "62% · 31 of 50". Hover it for the measured screens, the unvisited screens and the screens with no assertions.

  • The scope chips show All apps or one App at a time, each with its own percentage.

  • The map itself is a tree of cards: the Test Project, its Apps, their sections and their screens. Each card shows its percentage, a bar, the interactions tested out of the interactions present, and a No assertions tag when no assertion checks the screen. A screen card also shows how many tests reach it.

  • Expand or collapse a card with the round button on its right edge, or double-click it. Expand all and Collapse all act on the whole tree. Zoom in, Zoom out and Fit change the view.

  • The Search button in the page header opens a list of the weakest screens and the top gaps. Search screens and controls finds a screen by name or a control by its label, and moves the map to it.

The legend at the bottom of the map explains the colours:

Legend

Meaning

Under 50%

A gap: less than half of the interactions are tested.

50 to 79%

Partly covered.

80% and more

Well covered.

Unvisited screen

A screen that a link opens, but that no test reached.

Not measured

No test recorded the controls of the screen. When tests visited it, the card shows Visited by the number of tests. It counts in no percentage.

No assertions

Tests reach the screen, but no assertion checks it.

A Coverage Map. The tree runs from the Test Project through an App and a section to two screen cards, each with a percentage, a bar and a No assertions tag. The selected screen is outlined, and the side panel on the right shows its interaction coverage, its Not covered, Tests and Unvisited screens figures, then its covered and partly covered controls. The legend runs along the bottom.

Find the gaps

Select a card to open its details in the side panel. The panel has three tabs: Properties shows the details of the selected node, Gaps shows its gaps, and Tests lists the Test Cases that reach it. The Gaps and Tests tabs show a count.

  • The figures give Interaction coverage, Not covered and Unvisited screens. A screen also gives Tests, and a Test Project, an App or a section gives Measured, the screens the map could measure.

  • Top gaps, in the Gaps tab, lists the controls where a new test adds the most, each with its reason, for example Key action, Form field or Opens a screen no test reached.

  • On a screen, the controls are grouped into Covered, Partly covered and Not covered. A partly covered control shows the interaction it is missing, for example Missing: Left empty.

  • Select a control to see each of its interactions, and the Test Cases and runs that covered it.

  • On a Test Project, an App or a section, select the note about the screens with no assertions to list those screens. Select a screen in the list to show it on the map.

  • The Tests tab lists the Test Cases that reach the selected node, with a link to each Test Case and its latest run.

  • Cover with Thunders AI, at the bottom of the panel, opens Thunders AI in a new tab with a prompt that asks it to cover the gaps of the selected section, screen or control. For an App or the Test Project, the prompt covers the gaps of the whole map. Thunders AI adds the missing steps to an existing Test Case that reaches the screen, and creates a new Test Case only when none fits. Read the prompt, then send it. Users with the Read-only role do not see this button.

Mark a control, a screen or an exit as not worth testing

Some parts of your application do not need a test. To record that:

  1. Select the control, the screen or the unvisited screen. For a link to another domain, select the screen, then the domain under Exits.

  2. Select Not worth testing.

  3. Choose a Reason: Out of scope, Tested elsewhere, Not a real control or Other. Add a Note (optional) if you want.

  4. Select Save.

The item shows Dismissed for everyone who opens the map, and it stays on the map. A dismissed screen leaves the percentages at once. A dismissed control or unvisited screen leaves the percentages of every map that you compute after the dismissal. Select Undo to remove the dismissal. Users with the Read-only role see dismissals but cannot add or undo them.

See coverage from the pages you already use

  • Test Runs list. On a finished Test Set run, open the row menu and select Coverage map. Select Open coverage map when a map holds that run, or Compute coverage map to open the creation dialog with the run already ticked.

  • Test Set page. When a map holds runs of the Test Set, the Coverage map section of the properties panel shows the interaction percentage, for example "Up to 62% of interactions", the measured screens, the map name and date, and Open map. When newer runs exist, it says From an older map and offers Compute a new map.

  • Test Run report. See in coverage map opens the map with the screens of that run highlighted under Screens reached by this run. The figures stay those of the whole map. Select Clear to see the whole map again.

  • Test Case page. The Coverage section shows how many screens the Test Case reaches and how many interactions it covers in the latest map that holds it, and Show in map opens that map with its screens highlighted. It shows counts only, because a single Test Case has no percentage of its own.

  • Thunders AI. When a map exists, the chat of a Test Run suggests What did this run not cover?, and the chat of a Test Set suggests Which screens does this set never reach?. You can also ask Thunders AI about your maps, their gaps and the tests behind a screen.

  • Thunders MCP server. The tools list_coverage_maps, get_coverage_map, get_screen_gaps and get_test_case_coverage read your maps, and compute_coverage_map computes one from Test Set runs.

Frequently asked questions

How do I see coverage for my latest runs?

Compute a new map. A map never changes after it is ready, so it always describes the runs it read. On the Test Set page, Compute a new map opens the creation dialog with the latest finished run of the Test Set already ticked.

How do I get a screen measured?

Run a Test Case that reaches that screen, then compute a new map from that run. Opening the screen is enough, and the test does not need to use its controls. A screen shows Visited by the number of tests when the runs that reached it gave no list of its controls. It stays on the map, so you can see it was visited, but it counts in no percentage.

How do I count only the screens of my own application?

The map hides the other domains by default, such as the sign-in pages of an identity provider, and leaves them out of the percentages of the App and the Test Project. Select Include other domains to show them and count them again.

How do I take a screen out of the percentages?

Select the screen and mark it Not worth testing. It leaves the percentages at once, and it stays on the map with Dismissed, so you can select Undo later.

How do I close a gap?

Select the gap in Top gaps to see the control and the interaction it misses. Add a step that uses that control in an existing Test Case, or write a new Test Case for it, run it, then compute a new map to see the change.

How do I compare the coverage of two Test Sets?

Compare the gaps and the screens rather than the percentages alone. Each percentage only counts the screens its own map measured, so two maps that measured different screens give percentages that do not compare directly. The measured count next to each percentage tells you how much of the map it rests on.

How do I delete a Coverage Map?

On the Coverage Map page, open the row menu of the map and select Delete. The map leaves the list, and the Test Runs it read are not changed.

Related articles on this topic that you might find helpful

Did this answer your question?