ark trigger

Trigger a team run in a live Kubernetes cluster. Resets all pipeline steps and starts a new run.

ark trigger <team-name> [flags]

---

Flags

Flag	Default	Description
-n, --namespace	default	Namespace of the ArkTeam
--input	—	JSON object of input values: --input '{"key":"val"}'. Patches spec.input before the run.
--sync	false	Wait for the run to complete and print the output
--watch	false	Stream step status as the run progresses
--timeout	60s	Maximum wait time when --sync or --watch is set
--kubeconfig	~/.kube/config	Path to kubeconfig

---

Examples

# Trigger a run with input
ark trigger research-pipeline -n my-org --input '{"topic": "distributed systems"}'

# Wait for completion
ark trigger research-pipeline -n my-org \
  --input '{"topic": "distributed systems"}' \
  --sync --timeout 120s

# Stream step output as it runs
ark trigger research-pipeline -n my-org \
  --input '{"topic": "distributed systems"}' \
  --watch

# Trigger in default namespace
ark trigger quickstart

---

What trigger does

ark trigger performs a full re-run:

1. Patches spec.input with the provided --input values (if any)
2. Resets all pipeline steps to Pending
3. Clears status.output, status.startTime, status.completionTime
4. Sets status.phase to Running

The operator picks up the change on next reconcile and begins submitting steps.

To retry only failed steps from the last run (preserving succeeded steps), use ark retry instead:

ark retry research-pipeline -n my-org
ark retry research-pipeline -n my-org --step researcher   # retry one specific step

---

See also

• Triggering Pipelines guide — all trigger patterns
• ArkEvent concept — webhook and cron triggers