Current load and volume
- Ingress and egress PPS
- Ingress and egress Mbps
- Packet and byte counters
- Active flows
Monitoring / OpenTelemetry native
Load balancer and backend health, built in.
XLB exposes traffic, connection, backend load, capacity, and discovery metrics. Use ready-to-import Grafana and Datadog packages for fleet-wide monitoring, plus a local status interface for immediate health.
One-second local statusOTLP exportGrafana + DatadogKubernetes autoscaling
What XLB measures
The signals you need when traffic is high.
A process-up check is not enough at millions of requests per second. XLB shows whether traffic is balanced, backends are responding, connection health is changing, and the load-balancer tier is approaching a limit.
Distribution and churn
- New connections per backend
- Active backend connections
- Backend selection share
- Connection lifetime
Exact TCP behavior
- Handshake p50, p95, and p99
- Success, reset, and timeout rate
- SYN retransmissions
- No-backend-response connections
How connections end
- FIN and RST outcomes
- Orphaned connections
- Premature closure patterns
- Per-backend failure rates
Pressure before failure
- Flow-map occupancy
- Flow insertion failures
- NAT port allocation failures
- Maintenance scan duration
Backend pool freshness
- Eligible backend count
- Provider last-success age
- Readiness and synchronization
- Instance and deployment identity
Ready to use
Monitor XLB where your team already works.
The same metric contract powers the local status interface, Grafana dashboards, and Datadog dashboards. Your team gets a useful default without giving up its existing monitoring stack.
Built-in status
Immediate per-instance traffic, backend, capacity, and health information with no external monitoring dependency.
- One-second refresh
- Per-backend status
- Readiness and warning states
Ready-to-import dashboard
Premade Collector and Prometheus configuration, cluster-wide dashboards, backend health views, and alert rules.
- Cluster-wide traffic and capacity
- Backend comparison
- Prometheus alert rules
Prebuilt Datadog view
Premade dashboard, OpenTelemetry Collector configuration, and monitors for XLB traffic, capacity, and backend health.
- Cluster and fleet grouping
- Backend and transport health
- Datadog monitors
Built-in status interface
See what each XLB instance is doing right now.
Get one-second visibility into traffic, connections, backends, and capacity on every XLB instance without setting up an external monitoring platform.
Use the premade Grafana and Datadog packages for cluster-wide history, dashboards, and alerts.
01Instance overview
Traffic, connection load, backend readiness, and capacity headroom in one local view.
Open full screenshot ↗
02Backend distribution
Compare load, traffic, state, and idle removals across every backend serving the instance.
Open full screenshot ↗
03Connection behavior
Separate active, opened, closed, and idle-timeout activity with short local history.
Open full screenshot ↗
04Capacity and diagnostics
See resource pressure, dataplane attachment, failed work, and current status boundaries.
Open full screenshot ↗
05Backend drill-down
Open one backend to inspect current load and its recent connection activity.
Open full screenshot ↗Screens use demo data. Panels marked Coming Soon are not yet available in live mode.
Monitoring questions
What teams need to know.
How quickly does the monitoring data update?
The local status snapshot updates every second. The OTLP export interval is configurable, so teams can choose a one-second high-resolution profile or a lower-volume interval.
Can Grafana or Datadog monitor multiple XLB instances?
Yes. The premade configurations combine instances by service, cluster, namespace, node, or instance for shared dashboards, history, and alerts.
Can the dashboard compare multiple backends?
Yes. Backend identity is included on connection, response, lifetime, and failure metrics so dashboards can expose load imbalance, rising reset rates, slow handshakes, and other backend-specific changes.
Can Kubernetes autoscale XLB from these metrics?
Yes. XLB exports CPU and network utilization metrics that can drive Kubernetes autoscaling policies. The load-balancing tier can add or remove replicas as traffic changes while each instance continues to follow the same Service independently.
See XLB under load
Evaluate performance and visibility together.
We’ll validate throughput, backend behavior, capacity signals, and the monitoring path against your real traffic profile.