Skip to main content

External monitoring integrations

Use the VeloDB Cloud Metrics API to collect warehouse and cluster metrics in an external monitoring platform.

The Metrics API returns Prometheus text exposition format over HTTPS and requires an X-API-Key header. A collector can use the API if it can scrape Prometheus or OpenMetrics endpoints and set custom HTTP headers. This is protocol-level compatibility. Configuration, credential handling, metric naming, pricing, dashboards, and alerting still depend on the monitoring platform.

The Metrics API is in Private Preview. Contact VeloDB Cloud support to enable access for your organization before configuring a collector.

The data flow is:

VeloDB Cloud Metrics API -> collector -> external monitoring platform -> dashboards and alerts

Choose an integration​

Choose the guide that matches your monitoring platform. Use the Metrics API guide directly when you are configuring Prometheus or another compatible collector.

Datadog
Use Datadog Agent's OpenMetrics check to collect selected VeloDB Cloud metrics. This is the currently validated end-to-end integration in this documentation.
Set up Datadog Agent →
Prometheus-compatible collectors
Scrape the Metrics API with Prometheus or another collector that supports Prometheus or OpenMetrics endpoints and custom HTTP headers.
View Metrics API configuration →

Grafana is a visualization layer, not a replacement for the collector. Configure a collector to pull the Metrics API first, then connect Grafana to that collector's data source. For a Prometheus-based setup, see Visualize with Grafana.

Before you start​

  • Confirm that Metrics API Private Preview is enabled for your organization.
  • Create and securely store a VeloDB Cloud Management API key with the least-privilege role that can read the target warehouse. See API Keys.
  • Confirm network access from the collector host to the Metrics API over HTTPS. If you use private connectivity, see Private endpoint access.
  • Choose a warehouse endpoint for warehouse-level metrics, or add a cluster endpoint when you need metrics from a specific cluster. Use the target warehouse's cloudProvider, region, warehouseId, and, for cluster metrics, clusterId.
  • Start with one warehouse endpoint and a small metric selection. Expand the selection after the first scrape succeeds.

Metrics collected by Datadog are billed as Datadog custom metrics. If you use Datadog, start with an explicit metric and label allowlist and review Datadog custom metric billing before expanding it. Other monitoring platforms have their own pricing and cardinality limits.

For endpoint behavior, authentication, rate limits, labels, and available metric categories, see Metrics API and Metrics.