google-cloud-recipe-networking-observability

Diagnose Google Cloud networking issues using logs, metrics, and connectivity tests.

Updated Dec 9, 2025
One-click install
npx skills add https://github.com/Aki2022/skills --skill google-cloud-recipe-networking-observability-aki2022
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: google-cloud-recipe-networking-observability
Source: https://github.com/Aki2022/skills/tree/main/google-cloud-recipe-networking-observability
Command: npx skills add https://github.com/Aki2022/skills --skill google-cloud-recipe-networking-observability-aki2022

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Investigating Google Cloud networking problems requires navigating VPC Flow Logs, firewall logs, Cloud NAT logs, threat logs, and scattered metrics, which is slow and error-prone without a structured approach. ## Core Features & Use Cases - Log Analysis: Query VPC Flow Logs, firewall logs, Cloud NAT logs, and threat logs via Cloud Logging MCP or BigQuery aggregation on _AllLogs datasets. - Metrics & Performance Monitoring: Retrieve latency (RTT), packet loss, throughput, and NAT port exhaustion metrics through Cloud Monitoring MCP or curl fallbacks. - Connectivity Tests: Run Network Management path diagnostics to identify firewall or routing misconfigurations between endpoints. - Use Case: When a VM cannot reach an external service, use this Skill to check firewall DENY logs, verify NAT port allocation status, and run a Connectivity Test to pinpoint the blocking rule. ## Quick Start Ask the agent to investigate why traffic from a specific VM is being blocked by analyzing firewall and VPC Flow logs in your Google Cloud project.

Frequently Asked Questions about google-cloud-recipe-networking-observability

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

FAQPage Schema
How do I analyze VPC Flow Logs in Google Cloud?

Query VPC Flow Logs using Cloud Logging filters on compute.googleapis.com/vpc_flows for exploratory analysis, or run BigQuery SQL aggregations on the _AllLogs dataset for high-volume trends like top talkers and traffic volume.

How to find which firewall rule is blocking traffic in GCP?

Filter firewall logs for jsonPayload.rule_details.action="DENY" to identify blocking rules. Aggregate by rule_details.reference in BigQuery to rank the top-blocking firewall rules across your project.

Should I use BigQuery or Cloud Logging for network log analysis?

Use BigQuery linked datasets like _AllLogs for high-volume aggregations and trend analysis. Use Cloud Logging MCP for quick exploratory queries on recent individual log entries when BigQuery datasets are not linked.

How do I troubleshoot Cloud NAT port exhaustion?

Check NAT logs for jsonPayload.allocation_status="DROPPED" to identify VMs failing port allocation. You can also query the compute.googleapis.com/nat/dropped_sent_packets_count metric for historical exhaustion trends.

Why does my BigQuery query fail with an Unrecognized name error?

This usually indicates a schema casing mismatch, such as jsonPayload versus json_payload. Run bq show --schema to verify field names, then use bq query --dry_run to validate the corrected query before execution.

Can a Connectivity Test pass while traffic still fails?

Yes. Connectivity Tests perform static configuration analysis, so a REACHABLE result does not guarantee traffic flows if the source or destination VM is not in RUNNING state. Always verify instance status first.