Google Cloud GKE Integration
Full Kubernetes visibility for GKE clusters including node, pod, and workload metrics. Get AI-powered anomaly detection and root cause analysis for your container workloads.
How It Works
Create a GCP Service Account
Create a service account in your GCP project with the Monitoring Viewer and Container Cluster Viewer roles. Download the JSON key for use with TigerOps.
Enable Cloud Monitoring API
Enable the Cloud Monitoring API and Kubernetes Engine API in your GCP project. TigerOps uses these APIs to pull node, pod, and workload metrics from GKE.
Configure TigerOps GKE Integration
Add your service account credentials and GCP project ID to TigerOps. Select the GKE clusters to monitor and set your desired scrape interval.
Set Alerts and Dashboards
TigerOps auto-provisions GKE dashboards and alert policies for node CPU, pod restarts, OOMKills, and deployment rollout failures. Customize thresholds as needed.
What You Get Out of the Box
Node & Pod Resource Metrics
Track CPU requests vs. limits, memory utilization, pod restarts, and OOMKill events at the node and pod level. TigerOps alerts on resource pressure before nodes become unschedulable.
Workload Health Monitoring
Monitor Deployment, StatefulSet, DaemonSet, and Job health across all GKE clusters. Track rollout progress, replica counts, and failed pod conditions in real time.
Cluster Autoscaler Visibility
Observe node pool scale-up and scale-down events, pending pod counts that trigger autoscaling, and autoscaler decision latency to diagnose capacity issues.
Network & Ingress Metrics
GKE ingress request rates, error rates, and latency percentiles. Monitor pod-to-pod network throughput and detect abnormal egress patterns across namespaces.
Persistent Volume Monitoring
Track PVC utilization, volume attach/detach latency, and I/O throughput for GCE persistent disks attached to GKE workloads. Alert before volumes reach capacity.
AI Incident Correlation
When a GKE workload degrades, TigerOps AI correlates pod restart loops, node pressure, HPA scaling events, and upstream service errors to surface the root cause instantly.
GKE Service Account Setup
Provision the GCP service account and connect your GKE clusters to TigerOps.
# TigerOps GKE Integration Setup
# Prerequisites: gcloud CLI authenticated, kubectl configured
PROJECT_ID="your-gcp-project-id"
SA_NAME="tigerops-monitoring"
SA_EMAIL="${SA_NAME}@${PROJECT_ID}.iam.gserviceaccount.com"
# 1. Enable required APIs
gcloud services enable monitoring.googleapis.com \
container.googleapis.com \
--project=${PROJECT_ID}
# 2. Create service account
gcloud iam service-accounts create ${SA_NAME} \
--display-name="TigerOps Monitoring" \
--project=${PROJECT_ID}
# 3. Bind required roles
gcloud projects add-iam-policy-binding ${PROJECT_ID} \
--member="serviceAccount:${SA_EMAIL}" \
--role="roles/monitoring.viewer"
gcloud projects add-iam-policy-binding ${PROJECT_ID} \
--member="serviceAccount:${SA_EMAIL}" \
--role="roles/container.clusterViewer"
# 4. Download key
gcloud iam service-accounts keys create tigerops-key.json \
--iam-account=${SA_EMAIL} \
--project=${PROJECT_ID}
# 5. Configure TigerOps (tigerops.yaml)
cat > tigerops-gke.yaml << EOF
integrations:
gcp_gke:
project_id: "${PROJECT_ID}"
credentials_file: "./tigerops-key.json"
clusters:
- name: prod-cluster
location: us-central1
- name: staging-cluster
location: us-east1
scrape_interval: 60s
metrics:
- kubernetes.io/container/cpu/request_utilization
- kubernetes.io/container/memory/used_bytes
- kubernetes.io/node/cpu/allocatable_utilization
- kubernetes.io/pod/network/received_bytes_count
EOFCommon Questions
Does TigerOps support both Standard and Autopilot GKE clusters?
Yes. TigerOps integrates with both Standard and Autopilot GKE clusters via the Cloud Monitoring API. For Autopilot clusters, node-level metrics are surfaced in aggregate since individual node access is managed by Google.
Which GCP IAM roles are required for the service account?
The service account requires the roles/monitoring.viewer and roles/container.clusterViewer roles at minimum. If you want workload identity mapping, you can also bind the service account to a Kubernetes service account in your cluster.
Can TigerOps monitor multiple GKE clusters across different GCP projects?
Yes. TigerOps supports multi-project GCP monitoring. Add a service account for each project or use a single service account with cross-project IAM bindings. All clusters appear in a unified dashboard.
How does TigerOps handle GKE cluster upgrades?
TigerOps detects node pool version changes and flags them in the incident timeline. If a cluster upgrade correlates with a workload regression, the upgrade event is surfaced as a potential root cause in the AI analysis.
What is the minimum scrape interval for GKE metrics?
The Cloud Monitoring API surfaces GKE metrics at 60-second resolution by default. TigerOps supports custom scrape intervals down to 60 seconds for GKE. For finer granularity, deploy the TigerOps in-cluster agent which scrapes the Kubernetes metrics API directly.
Stop Debugging GKE Incidents in the Dark
Node pressure, pod restarts, OOMKills, and workload failures — all correlated by AI in one view. Connect in minutes.