K8Lens watches every workload. When health degrades, it correlates events, deploys, and resource pressure into a single explanation - so your team stops guessing why pods crash.
Every team stitches together the same stack to monitor Kubernetes. Five tools later, the answer still isn't on any of their screens.
Endless time-series, zero context. You're hand-writing PromQL at 2am just to find which pod moved.
A dozen dashboards later, none of them actually say which workload broke or why it broke now.
New Relic, Datadog, an agent on every node and it still misses live cluster and scheduler state.
get pods, describe, logs, exec. Every incident, across four terminal tabs, piecing it together by hand.
Plug in your cluster and every signal metrics, logs, events, topology, Helm, cost lands on one screen. When something breaks, the investigation is already done. Open the incident, read the why, ship the fix.
Deploy #482shipped a 30% larger heap at 22:58. The pod's memory limit (512Mi) was left unchanged, so the kubelet OOMKilled it on the first traffic spike. Raising the limit to 768Mi or rolling back #482 clears the loop.
resources.limits.memory: 768Mi in payments-api or roll back Helm release to 4.1.3.Prometheus-grade CPU, memory, and disk with zero scrape configs and no PromQL.
Every pod log and Kubernetes event, searchable and already in context.
One number for the whole cluster, with the exact reasons behind it.
Deployments, services, ingress and namespaces, mapped at a glance.
Track every chart, version and status without ever leaving the dashboard.
Per-namespace spend, no extra agent and no billing export to wire up.
MIT Licensed. Self-hosted or cloud. No black boxes, no vendor lock-in read the code, fork it, ship it.
Drops in with a read-only service account. K8Lens never writes to your cluster.