> Source: https://www.gitdash.info/docs/metrics-workflow · GitDash v4.7.1

# Workflow Overview Tab

The Overview tab of the Workflow Detail page shows four charts that give a quick health pulse for a specific GitHub Actions workflow, plus an Optimization Tips banner.

## Overview Charts

| Chart | What it shows | How to read it |
| --- | --- | --- |
| Rolling Success Rate | Moving average of CI pass rate over every 7 consecutive runs. | A dip below the red 80% reference line that persists across multiple windows indicates a systemic problem, not just a fluke failure. |
| Action Duration Trend | Two independent (non-stacked) area series: purple = execution time only (run_started → completed), amber = queue wait only (triggered → run_started). Total elapsed = purple + amber. | Rising purple = workflow getting slower (test suite growth, cache misses). Rising amber = runner capacity bottleneck. The two series share the Y-axis but are NOT added together. |
| Outcome Breakdown | Donut chart of run conclusions over the last 60 runs: success, failure, cancelled, skipped, timed_out. | A large failure or timed_out slice needs immediate attention. Cancelled runs often indicate force-pushes interrupting in-flight runs. |
| Run Frequency | Bar chart of runs triggered per calendar day over the last 14 days. | Gaps are expected on holidays. Unusual spikes may indicate retry storms, misconfigured cron schedules, or a flood of PRs. |

## Optimization Tips

GitDash automatically analyses workflow patterns and surfaces actionable suggestions in a dismissible banner at the top of the Overview tab. Tips are generated from the last 60 runs.

| Tip type | Trigger condition | Suggested action |
| --- | --- | --- |
| Weekend runs | >50% of runs triggered on Sat/Sun | Move scheduled/cron workflows to weekday-only schedules to save CI minutes. |
| High cancel rate | >20% of runs cancelled | Consider using concurrency groups to cancel superseded runs instead of letting them start. |
| Long queue wait | Queue wait P95 > 5 minutes | Add more runners or switch to larger GitHub-hosted runner tiers. |
| High re-run rate | >10% of runs re-triggered manually | Investigate flaky tests or infrastructure instability. |
| Duration regression | P95 duration grown >25% in last 14 days vs prior 14 days | Profile slow jobs — look for cache misses, dependency bloat, or uncapped test parallelism. |