Send Twilio notification when a workflow fails

Get a Twilio notification the instant any workflow in your namespace fails. This system flow watches every execution under a namespace prefix and fires a Twilio message on FAILED or WARNING status, so on-call engineers learn about broken pipelines from their phone instead of from a downstream stakeholder. It solves the classic monitoring gap: failures that happen overnight, on weekends, or in pipelines nobody is actively watching go unnoticed until the damage spreads.

How it works

• A io.kestra.plugin.core.trigger.Flow trigger listens to executions across all flows in the instance.
• Two conditions filter what fires the alert: io.kestra.plugin.core.condition.ExecutionStatus matches FAILED and WARNING, and io.kestra.plugin.core.condition.ExecutionNamespace matches the company prefix so every nested child namespace is covered.
• When a matching execution occurs, the send_twilio_notification task of type io.kestra.plugin.twilio.notify.TwilioExecution sends the alert.
• The message body includes the failed {{ trigger.executionId }} and the {{ taskrun.startDate }}, and authentication is read from secrets via accountSID, authToken, and the notification url.

What you get

• Real-time mobile alerts for every failed or degraded run, no manual polling.
• One central system flow instead of per-flow notification logic duplicated everywhere.
• Coverage of an entire namespace tree from a single prefix condition.
• Direct links back to the failing executionId for fast triage.

Who it's for

• On-call and platform engineers who need to react to incidents fast.
• Data and DevOps teams running critical scheduled pipelines.
• Anyone responsible for SLAs on overnight or unattended workflows.

Why orchestrate this with Kestra

A tool like Twilio sends messages but has no idea what your pipelines are doing. Kestra's event-driven Flow trigger reacts to execution state across your whole platform, applies declarative status and namespace conditions in plain YAML, and adds retries, full execution lineage, and a searchable audit trail. A standalone scheduler or cron job cannot observe other workflows' outcomes and fan out alerts on failure: that cross-flow, status-aware reactivity is exactly the gap Kestra fills.

Prerequisites

• A running Kestra instance with the Twilio plugin available.
• A Twilio account with a configured notification endpoint.

Secrets

• TWILIO_ALERT_URL: the Twilio notification URL.
• TWILIO_ACCOUNT_SID: your Twilio account SID.
• TWILIO_AUTH_TOKEN: your Twilio auth token.

Quick start

1. Add the three Twilio secrets to your Kestra instance.
2. Create this flow as a system flow.
3. Adjust the ExecutionNamespace prefix to match your own namespace.
4. Run any flow under that namespace that fails (for example a io.kestra.plugin.core.execution.Fail task) and confirm the alert arrives.

How to extend

• Add more ExecutionStatus values or a label condition to narrow what triggers alerts.
• Enrich the message body with the flow id, namespace, or a direct execution link.
• Add parallel tasks to also post to Slack, PagerDuty, or email for layered escalation.
• Branch on namespace so production failures page on-call while dev failures stay low priority.

Links

• Twilio plugin documentation