skills.rest
alon3153
alon3153COMMUNITY
Explore all skills by this author →
GitHub

gke-observability

Configures GKE Cloud Logging, Cloud Monitoring, and managed Prometheus metrics collection.

Updated May 11, 2026
Search Tags:#monitoring#kubernetes#observability#gcloud#prometheus#gke#cloud logging
One-click install
npx skills add https://github.com/alon3153/upe-social-publisher --skill gke-observability-alon3153
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: gke-observability
Source: https://github.com/alon3153/upe-social-publisher/tree/main/.agents/skills/gke-observability
Command: npx skills add https://github.com/alon3153/upe-social-publisher --skill gke-observability-alon3153

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? GKE clusters ship with minimal observability enabled by default, leaving teams blind to control-plane health, workload metrics, and audit trails until they manually configure logging and monitoring components. ## Core Features & Use Cases - Golden Path Monitoring Configuration: Enable the full monitoring suite including control-plane metrics (API server, scheduler, controller manager) via gcloud commands, with warnings about the replace-not-append flag behavior and per-sample billing. - Managed Prometheus & Alerting: Set up Google Managed Prometheus, query key GKE metrics with PromQL, and design alert thresholds for crash loops, node health, and API latency. - Log Querying with LQL: Query cluster, workload, and audit logs using gcloud logging read and Logging Query Language patterns for troubleshooting OOMKills and scheduling failures. - Use Case: A platform engineer notices missing API server metrics on a new cluster and uses this Skill to enable the full monitoring component list, deploy kube-state-metrics, and build a Cloud Monitoring dashboard. ## Quick Start Ask the assistant to enable full GKE monitoring and managed Prometheus on your cluster, including control-plane metrics and cost considerations.

Frequently Asked Questions about gke-observability

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I enable control-plane metrics on a GKE cluster?▼

Run gcloud container clusters update with --monitoring=SYSTEM,API_SERVER,SCHEDULER,CONTROLLER_MANAGER plus any workload components. Control-plane metrics are off by default and the flag replaces the previous setting entirely, so always pass the full desired list.

How to query GKE cluster logs with gcloud?▼

Use gcloud logging read with a filter like resource.type="k8s_cluster" and resource.labels.cluster_name to fetch system logs. For workload logs, filter on resource.type="k8s_container" with the namespace label, and use logName:"cloudaudit.googleapis.com" for audit logs.

Why are there no API server metrics on my new GKE cluster?▼

API server, scheduler, and controller manager metrics are not enabled by default; only SYSTEM components are. You must explicitly enable them via the --monitoring flag, and note that these metrics bill per sample ingested through Managed Service for Prometheus.

Does GKE collect kube-state-metrics by default?▼

No, GKE does not collect kube-state-metrics by default. Metrics like kube_pod_status_phase and kube_pod_container_status_restarts_total require deploying the Managed Prometheus kube-state-metrics package first.

What is the difference between gcloud monitoring flags and the monitoringConfig API enum?▼

The gcloud flag uses spellings like SYSTEM and API_SERVER, while the monitoringConfig API uses SYSTEM_COMPONENTS and APISERVER. Mixing the two fails the command, which is a common and confusing error.

How can I reduce GKE monitoring costs in non-production clusters?▼

Update the cluster with --monitoring=SYSTEM to collect only system metrics. Cloud Logging charges per GiB ingested beyond the 50 GiB free tier, and Managed Prometheus charges per samples ingested, so limiting components lowers cost.

Related Skills

Explore More Software Engineering Skills→
gke-labsgke-labs
42

gke-cluster-lifecycle

Manage Google Kubernetes Engine cluster lifecycle, upgrades, and decommissioning.

Official
Advanced
HabitaNexusHabitaNexus
1

Alerting & Incident Management

Design alerting and incident management workflows for production services.

Official
Advanced
BagelHoleBagelHole
46

opentelemetry

Instrument applications and infrastructure with OpenTelemetry for unified traces, metrics, and logs.

Community
Intermediate
3horizons3horizons

observability-stack

Deploy Prometheus, Grafana, Loki, and Alertmanager onto Kubernetes with Helm.

Official
Advanced