Workflow Runs

~11 min · runs, logs, debugging

Level 0Apprentice

0 XP0/101 lessons0/10 achievements

0/120 XP to next level120 XP to go0% complete

What you actually see in the Actions tab

Each trigger creates a run. A run has a unique ID, a list of jobs, logs per step, artifacts, and a final conclusion (success, failure, cancelled, skipped, timed_out, action_required, neutral).

Reading a run

Summary view — the workflow graph, with each job as a node and needs: edges drawn. Click a job to see its steps.
Step logs — full stdout/stderr per step. Search box at the top for grep across all steps.
Artifacts — files uploaded via upload-artifact. Download to debug locally.
Re-run — re-run all jobs, only failed jobs, or with debug logging enabled.
Cancel — stop an in-progress run.

Debug logging

Set repo or workflow secrets:

ACTIONS_RUNNER_DEBUG = true — verbose runner.
ACTIONS_STEP_DEBUG = true — verbose steps.

Or click 'Re-run with debug logging' from the run UI. The logs become 5–10× longer but show every API call the runner makes.

Code

Watch a workflow run from the CLI·bash

# Show the most recent run of every workflow
gh run list --limit 5

# Watch a specific run live (auto-tails until completion)
gh run watch <run-id>

# Show the failed step's logs only
gh run view <run-id> --log-failed

# Re-run only the failed jobs
gh run rerun <run-id> --failed

# Cancel an in-progress run
gh run cancel <run-id>

External links

GitHub CLI — gh run

Exercise

Push a commit that intentionally breaks one step (e.g., exit 1 in a test script). Use gh run watch from the terminal until it fails. Then use gh run view --log-failed to find the failure. Fix and re-run with gh run rerun --failed.

Progress

Progress is local-only — sign in to sync across devices.

← PreviousTriggering Events Deep Dive Next →Workflow File Naming & Organization

Spotted a bug or have feedback on this page?Report an Issue

Comments 0

🔔 Reply notifications (sign in)

No comments yet — be the first.