Skip to main content

Quickstart

This example creates a two-step pipeline:

  1. a scheduled ingest job
  2. a downstream report job that uses captured output

1. Create husky.yaml​

version: "1"
defaults:
  timeout: "10m"
  retries: 1
  retry_delay: exponential
  default_run_time: "0900"
  timezone: "UTC"

jobs:
  ingest_events:
    description: "Download fresh events and print the output file path"
    frequency: every:30s
    command: "./scripts/ingest.sh"
    output:
      file_path: last_line

  generate_report:
    description: "Generate a report from the latest ingest output"
    frequency: after:ingest_events
    command: "./scripts/report.sh {{ outputs.ingest_events.file_path }}"
    on_failure: stop

2. Add example scripts​

mkdir -p scripts
cat > scripts/ingest.sh <<'EOF'
#!/usr/bin/env sh
echo "[ingest] downloading events"
echo "/tmp/husky/events.json"
EOF
chmod +x scripts/ingest.sh

cat > scripts/report.sh <<'EOF'
#!/usr/bin/env sh
echo "[report] using input: $1"
EOF
chmod +x scripts/report.sh

3. Validate config​

husky validate

4. Start the daemon​

husky start

5. Inspect status​

husky status
husky dag

6. Trigger the root job manually​

husky run ingest_events --reason "quickstart smoke test"

7. Watch output​

husky logs ingest_events
husky logs generate_report
husky history ingest_events
husky audit

8. Open the dashboard​

husky dash

Or visit the address stored in .husky/api.addr.

9. Stop Husky​

husky stop

Next steps​