Metrics, logs, events, and traces from every cluster, node, and pod. Correlated at ingestion. Searchable in one query. No tool-switching required.
Kubernetes topology — clusters, nodes and pods with live resource usage
Pods, nodes, namespaces, and clusters auto-discovered with full cardinality. Every pod label, every replica stored natively. No premium tier for high-cardinality tags.
Out-of-the-box dashboards for pods, nodes, namespaces, and workloads. Prebuilt alerts for CPU pressure, OOM kills, pod restart loops, and disk saturation. No manual setup.
Container metrics join to APM spans and logs in one query. The CPU spike, the OOM-kill log, and the failing span — together.
A platform deep dive: fleet overview, per-resource drill-down, live dependencies, the agents you already run, and the facets that tie it all together.
Four purpose-built views for Pods, Clusters, Nodes, and Namespaces. Each entity is auto-discovered and tagged with Kubernetes metadata the moment it appears — click any entity to drill into its metrics, logs, events, and traces without losing context.
Kubernetes topology — clusters, nodes and pods with live resource usage
Select any pod, node, or namespace and its Metrics tab shows real-time CPU, memory, network, and disk usage. The Logs tab surfaces application and system logs scoped to that resource. The Events tab shows scheduled updates, restarts, warnings, and failed operations.
Pod detail — CPU, memory and restart history for a single workload
The service map renders a live dependency graph for every service in the cluster. New deployments and changed routes appear without manual configuration. Health indicators overlay every edge so you spot degraded paths at a glance.
Kloudfuse ingests from the tools you already run. Existing Grafana dashboards import directly. Teams keep their preferred query language. The only thing that changes is where the data lands.
| Agent you already run | How it connects |
|---|---|
| OpenTelemetry Collector | ✓ DaemonSet · OTLP |
| Datadog Agent | ✓ Native ingest, no code changes |
| Filebeat · Fluent Bit · Fluentd | ✓ Log shippers, as-is |
| Victoria Metrics · Prometheus | ✓ remote_write |
| Elastic APM | ✓ No instrumentation rewrite |
| New Relic agent | ✓ Drop-in |
Every telemetry signal is tagged with nine Kubernetes facets at collection time. Apply a filter once and see results across metrics, logs, events, and traces simultaneously. No manual tagging. No label mapping.
Kloudfuse collects four telemetry types from Kubernetes environments: metrics (via kubeletstats, hostmetrics, and k8s_cluster receivers), logs (from kubelet, containerd, and application containers with auto-extracted metadata), events (pod crashes, scheduling failures, scaling operations), and distributed traces with K8s metadata labels. All signals are automatically tagged with cluster, namespace, node, deployment, and pod-level facets at collection time.
No. Kloudfuse deploys OpenTelemetry Collectors as DaemonSets for telemetry collection. It also supports Datadog Agent, Filebeat, Fluent Bit, Fluentd, Victoria Metrics, Elastic APM, and New Relic agents — all without code changes. Bring what you already run.
Yes. Kloudfuse supports Prometheus remote_write and is fully PromQL-compliant, so teams can point existing Prometheus instances at Kloudfuse or replace the Prometheus server entirely. Existing Grafana dashboards import directly and continue to work with PromQL queries. kube-state-metrics is required for Advanced Analytics features.
Kloudfuse automatically detects new pods, namespaces, and workloads as they appear. Every telemetry signal is tagged with auto-extracted facets including kube_cluster_name, kube_namespace, kube_node, kube_deployment, kube_replica_set, kube_service, pod_name, pod_phase, and pod_status. No manual configuration is needed when clusters scale or workloads change.
All telemetry data stays inside your VPC. Kloudfuse deploys as a Self-SaaS platform in your own cloud environment. No metrics, logs, events, or traces leave your infrastructure. This satisfies data residency requirements and eliminates compliance concerns around telemetry egress.
Deploy OTel Collectors as DaemonSets, connect your clusters, and see pods, nodes, and namespaces in the Kloudfuse UI. No proprietary agents required.
Read the guideHow Kloudfuse built real-time Kubernetes topology maps powered entirely by OpenTelemetry — no proprietary agents, no sidecars.
Read the blogAgent compatibility, data residency, pricing model, and Prometheus support. A side-by-side breakdown for teams evaluating both.
See the comparisonThirty minutes on your telemetry. The cause, before the call ends.