Skip to main content

Load Balancer Metrics

info

Load Balancer Metrics is currently in Limited Availability. Metrics are automatically available for all customers with access to Load Balancers.

Load Balancer Metrics provide comprehensive insights into the throughput of your load balancers. These metrics help you monitor traffic patterns, analyze connection behavior, and understand data flow through your load balancers. The following key metrics are collected in 30-second intervals for a period of 30 days for each load balancer.

MetricsDefinitionSuggested Query
Outbound PacketsThe number of packets transmitted by the load balancer.rate(crusoe_elb_out_packets[5m])
Inbound PacketsThe number of packets received by the load balancer.rate(crusoe_elb_in_packets[5m])
Bytes OutThe total bytes transmitted by the load balancer.rate(crusoe_elb_bytes_out[5m])
Bytes InThe total bytes received by the load balancer.rate(crusoe_elb_bytes_in[5m])
Active FlowsThe number of currently active connections through the load balancer.crusoe_elb_active_flows
New FlowsThe number of new connections established through the load balancer.rate(crusoe_elb_new_flows[5m])

Load Balancer Metrics are available by default for all load balancers—no agent installation is required. They're accessible via a PromQL API, Telemetry Conduit, and Crusoe MCP. Console UI support will be available in a future release.

By default, queries return metrics for all load balancers within a project. To query metrics for a specific load balancer, add a label selector to your query:

crusoe_elb_active_flows{elb_name="my-load-balancer"}

For token generation and instructions on querying via API, Grafana, or Telemetry Conduit, see Get started.