Kure Monitor

Kure Monitor is a Kubernetes observability and security platform for teams that need to diagnose cluster failures. It detects failing pods and provides AI-generated explanations and fixes using local or cloud LLMs. Features include live pod monitoring and logs, security checks for more than 50 misconfigurations, and a cluster topology graph. Teams can configure multiple AI providers, add custom instructions in Markdown, and manage access with granular RBAC or Keycloak SSO. Kure installs with Helm, stores nothing outside the cluster, and is described as open source and air-gapped ready. It focuses on failure diagnosis and security, complementing rather than replacing Prometheus and Grafana.

Kure is a Kubernetes observability platform focused on failure diagnosis and security. It detects pod crashes, image-pull errors, and scheduling issues, then uses AI to provide context-aware explanations and fixes. It also audits clusters for security misconfigurations and provides live pod monitoring, logs, and topology features. Kure is intended to complement Prometheus and Grafana rather than replace them; the page says it does not provide time-series metrics or alerting rules. Users install Kure in a Kubernetes cluster with Helm. It monitors pods and detects failures, then provides AI explanations and copy-paste fixes using local or cloud language models. Users can register multiple AI providers, set a primary provider with automatic failover, or select a model during pod triage. Markdown instructions can provide models with cluster topology and team standards. The platform also streams container logs, scans for security issues, and supports Slack or Teams failure notifications. Kure is for teams operating Kubernetes clusters that need help diagnosing pod failures and checking cluster security. The page highlights team access controls, namespace-based permissions, and Keycloak SSO, along with support for local language models and air-gapped environments. It is positioned as a complement to Prometheus and Grafana, with a focus on failure diagnosis and security rather than metrics, alerting, or log aggregation.