Troubleshooting
Find the message or the number that surprises you. Each entry says why it happens, what to check and how to fix it.
In short
On this page you can
- Fix a tab that says it could not be loaded.
- Find out why a period shows no runs, or a pipeline is missing.
- Explain numbers that differ from Report or from the Tasks page.
- Understand why coverage stays low while everything passes.
- Solve saved report, download and clipboard errors.
How to use it
- Click Refresh first: InSight does not update by itself.
- Find your message or symptom in the headings below.
- Read Why, then do the Check list in order.
- Apply the Fix; if it persists, write to info@smartestqa.ai with the project name.
Loading and empty tabs#
“Execution could not be loaded” / “Test Health could not be loaded” — “The server did not answer. Try again in a moment.”Loading
- Why
- The calculation did not come back: a network problem, an expired session, or a very large project taking too long.
- Check
-
- Other pages load; if not, log in again.
- A project is selected in the top bar.
- For Execution, a shorter Period loads faster.
- Fix
- Click Try again. If it keeps failing, write to info@smartestqa.ai with the project name and the time.
“No runs in this period”Execution
- Why
- No pipeline run, and no test case run on its own, started in the period. Dry runs are never counted.
- Fix
- Click Show the last 90 days, or Open pipelines and run one. See Quick Start, step 6.
“Nothing to measure yet”Test Health
- Why
- The project has no sub-requirements, no scenarios and no pipelines yet.
- Fix
- Add requirements in Analyze, or click Copy MCP prompt and let your assistant walk you through setup. The Setup Roadmap in Discover shows the order; see Progress meters.
I ran a pipeline but InSight shows the old numbersBoth tabs
- Why
- InSight calculates when the tab opens and when you click Refresh. The live chip says how old the numbers are.
- Fix
- Click Refresh after the run has finished.
InSight opens on Test Health instead of ExecutionBoth tabs
- Why
- Projects with AI agents open on Test Health until you pick a tab; after that, the last tab you used in this browser opens.
- Fix
- Click Execution once, or bookmark
/dashboard/default?tab=execution.
Execution numbers#
InSight and Report show different numbersExecution
- Why
- They count different things.
- Report’s Result chip counts every attempt, retries included; the InSight strip uses each test case’s latest attempt.
- The success rate is per scenario run over the whole period; Report shows one run.
- Report’s list filters by the start of a run’s first test case, InSight by each scenario run’s start.
- The classic view counts per step.
- Fix
- Compare like with like using the formulas in InSight numbers and Report numbers.
“{n} runs stuck in RUNNING”Execution
- Why
- Scenario runs from the last 30 days are still Running more than 2 hours after they started: the runner died and the result never came back.
- Check
- The agents in the same row are online; the runs in Report.
- Fix
- Stop them in Report (Stop a run), bring the agent back and rerun. They also turn Disconnected by themselves; see the 2-hour rule.
“Agents 0 / {n} online” in redExecution
- Why
- None of your company’s agents is connected. Pipelines that target an agent cannot start.
- Fix
- Start the agent app on the machine and wait for green. See Install an agent and agent offline. Runs on the server do not need an agent.
A pipeline is missing from the Pipeline run stripExecution
- Why
- It has not run in the last 56 days (or the longer period you chose) and has no active schedule.
- Fix
- Run it once, or give it a schedule. Pipeline levels are in Pipeline maturity.
A flaky scenario shows “uncertain · {n} runs”, or the card number is lower than the listExecution
- Why
- It passed and failed in the period, but ran fewer than 5 times, so no score is given. The card counts only scored scenarios, so it can be lower than the number of rows.
- Fix
- Pick a longer period, or let it run more. See Flaky score.
A failure cluster is not marked Fixed although the test passes nowExecution
- Why
- Fixed needs every test case of the cluster to pass in a later, different run. A pass on an automatic retry inside the same pipeline run does not count, and one test case still failing keeps it open.
- Fix
- Run the pipeline again, then Refresh. See Clusters and Fixed.
Time saved shows “—” or is very lowExecution
- Why
- No scenario passed in the period (“—”), the passing runs wrote few step rows because Skip Successful Logs was on, or the runs took longer than 30 seconds per step.
- Fix
- Nothing is wrong with the tests. See how it is counted: Time saved.
Open tasks differs from the Tasks pageExecution
- Why
- InSight counts every task whose status is not closed, resolved or done, at any date, including your own custom statuses.
- Fix
- Close finished tasks with a closing status. “new in {n} days” follows the period; “open” does not.
“The file could not be created.”Execution
- Fix
- Wait for the tab to finish loading and try Download again. Use the PDF option as a fallback.
Test Health numbers#
Scenarios pass but requirement coverage stays lowTest Health
- Why
- A sub-requirement only counts when a linked scenario checks a result and passed its last run. Passing scenarios that only click and type prove nothing; scenarios not linked to a sub-requirement prove nothing for it.
- Check
- Trust issues › “scenarios check no result”, and Gaps › Unlinked sub-requirement.
- Fix
- Add an assertion or a visibility check with text to the end of each scenario, and link scenarios to their sub-requirements in Analyze. See Pillars.
A pillar says “not measured yet”, or maturity shows “—”Test Health
- Why
- The pillar has nothing to divide by: no sub-requirements, no risky sub-requirements (risk pillar), or no actions. Maturity is left out until at least one pillar is measured.
- Fix
- Add requirements in Analyze and set severities on the risky ones. Missing pillars are left out of maturity, not counted as 0. See Maturity.
A module is missing from Modules × coverageTest Health
- Why
- Modules are the top-level folders of the Requirements tree. A folder without sub-requirements is listed only under “+ {n} modules not started yet”. Folders deeper down count toward their top-level folder.
- Fix
- Add sub-requirements under the folder’s main requirements, or move the folder to the top level in Analyze.
Scenarios in a called pipeline are counted as “in no pipeline”Test Health
- Why
- Test Health only follows a pipeline’s own rows (test cases, requirement rows, suites). Test cases of a pipeline that is only called from another pipeline are not followed.
- Fix
- Give the called pipeline its own run or schedule, or add the test cases directly.
“No history yet.” / “Only one point in this period; pick a longer period to see the trend.”Test Health
- Fix
- Pick a longer period in Progress (default Last 90 days). Points before the project’s first record are not drawn.
“The report could not be saved.” / “…could not be opened.” / “…could not be deleted.”Test Health
- Why
- The server did not accept the request, or the saved report belongs to another project.
- Fix
- Switch to the right project, wait for the tab to load, and try again. Keep the name under 200 characters. See Saved reports.
“Could not copy to the clipboard.”Both tabs
- Why
- The browser blocked clipboard access, which happens on pages not served over HTTPS or with strict browser settings.
- Fix
- Allow clipboard access for SmartestQA in the browser’s site settings and click Copy MCP prompt again.
Didn’t find your answer?Write to support with your project name and what you see on screen.
info@smartestqa.ai