Metrics and Observability

Orkes Conductor allows you to monitor workflow performance using tools like Prometheus, Grafana, and Datadog. The Conductor dashboard provides a centralized view of key metrics, allowing you to gain real-time insights and set up customizable alerts for proactive monitoring.

Dashboard setup

Orkes Conductor uses Prometheus to capture a comprehensive set of metrics, which are readily available in your cluster deployment. On request, these metrics can be pushed to Grafana or Datadog.

Notes

• For cloud deployments, Prometheus and Grafana/Datadog are provisioned by the Orkes team.
• For on-premise deployments, users must set up their own monitoring tools.

Accessing the dashboard

To access the dashboard, go to Metrics from the left navigation menu on your Conductor cluster.

Note: To enable Metrics for your cluster, please get in touch with the Orkes team.

Example dashboard set up using Prometheus and Grafana

Conductor metrics​

The Conductor server collects the following metrics via Prometheus, which can be visualized using Grafana/Datadog for monitoring and setting up alerts.

Workflow metrics​

Metrics	Decsription	Tags
workflow_completed_seconds	Timer indicating the average time taken to complete the workflows.	workflowName, quantile
workflow_completed_seconds_count	Counter indicating the number of workflows completed per second. Apply the FAILED filter to get the failed workflows.	workflowName
workflow_running	Gauge indicating the number of running workflows.	workflowName
workflow_start_request_seconds_count	Counter indicating the number of workflows started per second.	workflowName

Task metrics​

Metrics	Decsription	Tags
task_completed_seconds	Timer indicating the average time taken to complete the task.	taskType, quantile
task_completed_seconds_count	Counter indicating the number of tasks completed per second. Apply the FAILED filter to get the failed tasks.	taskType
task_queue_depth	Counter indicating the number of pending tasks in a given time period.	taskType
task_poll_request_seconds_count	Counter indicating the number of tasks polled per second.	taskType

Configuring alerts

Alerts can be configured for any metric from your dashboard or via the Alerting section in Grafana.

To configure alerts for a metric from your dashboard

1. Open your Grafana dashboard and select the metric.
2. Select Edit.

3. Go to the Alerts tab and select Create alert rule from this panel.

4. Create the alert based on your requirements.

To configure alerts from the Alerting section

1. Go to Alerting > Alert rules from the left navigation menu on your Grafana dashboard.
2. Select + New alert rule.
3. Create the alert based on your requirements.

Refer to the official Grafana documentation for more information on setting up alerts.