Skip to main content

Introduction to CoreWeave Telecaster™

Use CoreWeave Telecaster™ to forward logs and metrics

CoreWeave Telecaster™ is a fully-managed solution for forwarding logs and metrics to destinations outside of CoreWeave's observability platform with the following features and benefits:

  • Comprehensive data aggregation: Ingests telemetry data from multiple sources.
  • Flexible data filtering and routing: Allows fine-grained data filtering before sending to configured endpoints.
  • Guaranteed delivery: High SLOs for data completeness and latency.
  • Fully managed with easy configuration: Zero-operational overhead for configuring data filters and endpoint routing.

By supporting multiple data streams to a variety of destination endpoints, CoreWeave Telecaster™ provides flexibility for managing and analyzing your data.

Supported logs and metrics

The table below lists the supported logs and metrics for forwarding:

StreamData
Audit LogsCKS audit logs
CoreWeave API audit logs
Console audit logs
Support Access Management logs
Application LogsCustomer CKS Pod logs
Platform LogsCKS Control Plane logs (other than audit logs, for example etcd logs, controller manager logs)
CKS system component logs (for example, CSI, CNI)
Node logs (for example, kernel logs)
Network fabric logs
Platform MetricsCKS Control Plane metrics
Redfish metrics
Ethernet routing metrics
Ethernet interface metrics
BMC (Node and DPU) metrics
NVswitch metrics
DCGM metrics

Supported endpoints

You can route logs and metrics to the following endpoints:

  • HTTP
  • Prometheus Remote Write v1.0+
  • Fluentbit v2.0.x+ (TCP Socket w/ JSON codec)

Configure CoreWeave Telecaster™

To set up and configure CoreWeave Telecaster™, contact Support to define the following:

  • Data streams
  • Forwarding and filtering requirements
  • Destination endpoints

Limitations

Note that Prometheus metrics from your application workloads are not included in CoreWeave Telecaster™. In general, CoreWeave Observe™ does not ingest application workload metrics. If you're interested routing application workload metrics, see Application Metrics Forwarding.