All Integrations
ContainersAny K8s Distribution

Kubernetes Integration

Full-stack Kubernetes observability — nodes, pods, namespaces, deployments, HPA, and workload logs — deployed in 60 seconds via Helm. AI-powered OOMKill detection, autoscaling analysis, and cost attribution included.

Setup

How It Works

01

Add the Helm Repository

Add the TigerOps Helm chart repository to your local Helm installation. One command and the repository is ready for use across all clusters.

02

Install the DaemonSet

Deploy the TigerOps agent as a DaemonSet via Helm. It auto-discovers all nodes, namespaces, and workloads — no per-service configuration required.

03

Metrics Flow Instantly

Within 60 seconds, node CPU/memory, pod restarts, deployment rollout status, PVC usage, and HPA scaling events appear in TigerOps dashboards.

04

AI Learns Your Workloads

The AI SRE builds behavioral baselines per namespace and deployment. It detects abnormal pod evictions, memory leaks, OOMKills, and CrashLoopBackOffs automatically.

Capabilities

What You Get Out of the Box

Node & Pod Metrics

CPU throttling, memory pressure, OOMKill events, pod restarts, and node conditions — correlated per deployment, namespace, and label selector.

Workload Health Tracking

Deployment rollout status, ReplicaSet availability, StatefulSet partition progress, and DaemonSet desired vs. ready counts.

HPA & Autoscaling Visibility

Track HPA scale-up and scale-down events, current vs. desired replica counts, and the metrics driving autoscaling decisions.

PersistentVolume Monitoring

PVC capacity utilization, inode exhaustion warnings, and storage class performance degradation alerts before your pods start failing.

Multi-Cluster Dashboard

Manage any number of clusters — GKE, EKS, AKS, bare-metal — in a single TigerOps workspace with per-cluster and cross-cluster incident correlation.

Namespace Cost Attribution

Break down compute cost by namespace, deployment, and team label. Identify overprovisioned workloads and right-size recommendations automatically.

Configuration

Helm Install

Deploy the TigerOps agent to your cluster in under a minute with Helm.

helm-install.sh
# Add the TigerOps Helm repository
helm repo add tigerops https://charts.tigerops.io
helm repo update

# Install into the monitoring namespace
helm install tigerops-agent tigerops/tigerops-agent \
  --namespace tigerops-monitoring \
  --create-namespace \
  --set apiKey="<YOUR_API_KEY>" \
  --set cluster.name="production-cluster" \
  --set logs.enabled=true \
  --set apm.enabled=true \
  --set costAttribution.enabled=true

# Verify the DaemonSet is running on all nodes
kubectl get daemonset -n tigerops-monitoring
# NAME              DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE
# tigerops-agent    12        12        12      12           12

# Check agent status
kubectl logs -n tigerops-monitoring -l app=tigerops-agent --tail=20
FAQ

Common Questions

Which Kubernetes distributions does TigerOps support?

TigerOps works with any CNCF-conformant Kubernetes distribution: GKE, EKS, AKS, Rancher, OpenShift, k3s, k0s, and self-managed clusters on bare metal or VMs. The Helm chart is distribution-agnostic.

What RBAC permissions does the TigerOps agent need?

The DaemonSet requires a ClusterRole with read access to nodes, pods, deployments, replicasets, statefulsets, daemonsets, namespaces, events, and horizontalpodautoscalers. Our Helm chart creates this ClusterRole automatically. No write permissions are required.

Does TigerOps integrate with kube-state-metrics?

Yes. If you already have kube-state-metrics deployed, TigerOps can scrape it directly via Prometheus remote-write or the TigerOps collector. If you do not have it, the TigerOps DaemonSet bundles equivalent metrics collection without a separate deployment.

Can I monitor Kubernetes logs as well as metrics?

Yes. The TigerOps DaemonSet includes a log collection component that tails container stdout/stderr from all pods, enriches logs with Kubernetes metadata (pod name, namespace, deployment, node), and forwards them to TigerOps log management — all via the same Helm install.

How does TigerOps handle CrashLoopBackOff and OOMKill detection?

The AI SRE watches pod restart counts, OOMKill events from kernel cgroups, and container exit codes in real time. When a CrashLoopBackOff is detected, it automatically correlates the restart events with log output, memory trends, and recent deployments to suggest a root cause within seconds.

Get Started

Deploy to Kubernetes in 60 Seconds

No credit card required. One Helm command. Full cluster visibility.