Skip to content

Dashboard: Notebook Validation Success Rate

Category: Developer Experience Format: OpenShift Console Contributor: @tosin2013 Requires OCP: 4.20+

Purpose

Provides a focused view of notebook validation success and failure rates across namespaces. Useful for platform teams monitoring MLOps pipeline health and for notebook authors tracking their validation outcomes over time.

Prerequisites

  • OpenShift user-workload monitoring enabled (enableUserWorkload: true in cluster-monitoring-config)
  • Operator deployed with Prometheus ServiceMonitor enabled

Installation

kubectl apply -k config/monitoring/community/

Panels

Panel Query Description
Overall Success Rate sum(rate(...{status="succeeded"}[5m])) / sum(rate(...[5m])) * 100 Percentage of validations succeeding in a 5-minute window
Validations by Status sum(rate(...[5m])) by (status) Time series of validation rates broken down by status
Success Rate by Namespace ...{status="succeeded"} by (namespace) / ... by (namespace) * 100 Per-namespace success percentage
Validation Duration (p50, p95) histogram_quantile(0.50\|0.95, ...) Validation execution time percentiles
Total Validations (24h) sum(increase(...[24h])) Count of all validations in the last 24 hours
Failed Validations (24h) sum(increase(...{status="failed"}[24h])) Count of failed validations with color thresholds

Metrics Used

All metrics are exposed by the operator out of the box:

  • notebookvalidationjob_validations_total (counter, labels: status, namespace)
  • notebookvalidationjob_validation_duration_seconds (histogram)