Monitoring

Menatic AI provides built-in monitoring on the Dashboard and integrates with standard Kubernetes observability tools for production deployments.

Built-in Monitoring

Menatic AI provides built-in monitoring on the Dashboard:

Node metrics — real-time CPU and memory per node (requires metrics-server)
Workload metrics — per-notebook and per-MCP-server CPU and memory usage
Resource allocation — cluster-wide CPU, memory, and GPU allocation charts

metrics-server

The Kubernetes metrics-server aggregates resource usage from the kubelet on each node. It powers:

kubectl top nodes / kubectl top pods
Dashboard node monitoring
Workload usage columns

Installation

kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

For k3d (disable TLS verification):

kubectl patch deployment metrics-server -n kube-system \
  --type=json \
  -p='[{"op":"add","path":"/spec/template/spec/containers/0/args/-","value":"--kubelet-insecure-tls"}]'

Verification

kubectl top nodes
kubectl top pods -n apps

Prometheus & Grafana (Advanced)

For production deployments, integrate with Prometheus and Grafana for historical metrics, alerting, and dashboards.

Recommended Stack

# Install kube-prometheus-stack (includes Prometheus + Grafana + Alertmanager)
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm install prometheus prometheus-community/kube-prometheus-stack \
  --namespace monitoring --create-namespace

Useful Metrics

Metric	Description
`container_cpu_usage_seconds_total`	CPU usage per container
`container_memory_working_set_bytes`	Memory usage per container
`kube_deployment_status_replicas_ready`	Ready replicas per deployment
`kube_pod_status_phase`	Pod phase (Running, Pending, etc.)

Log Aggregation

Collect logs from all platform components using a standard Kubernetes log stack:

# Install Loki + Grafana Agent
helm install loki grafana/loki-stack \
  --namespace logging --create-namespace

Then query logs from the Grafana UI or with LogQL.

Alerting

Configure alerts for:

Node CPU/memory above 80% for 5 minutes
Any notebook pod in CrashLoopBackOff
Database connection failures
Platform backend not responding