skill-gcp-playbooks

name	skill-gcp-playbooks
description	SRE investigator for executing established GCP/GKE playbooks for infrastructure discovery and mitigation.
author	Riccardo
version	0.0.3
status	draft

GCP SRE Playbooks

Overview

This skill provides a collection of playbooks and reference materials for SREs to investigate and mitigate incidents on Google Cloud Platform. It focuses on mapping generic SRE concepts (discovery, mitigation, postmortem) to specific GCP/GKE actions.

Naming Convention

Playbooks follow a deterministic naming convention based on the GCP API service name, located in the {SKILL_DIR}/references/googleapis.com/ directory. Example: run.md (for run.googleapis.com). Each file should also explicitly list the full API name at the top.

When to Use

• When starting an investigation and needing to discover the infrastructure.
• When following a specific mitigation strategy (e.g., rollback, traffic shifting).
• When a user asks for "GCP playbooks" or "SRE playbooks".

Available Playbooks

1. Infrastructure Discovery

Located at skills/gcp-playbooks/references/googleapis.com/README.md. Use this to systematically identify the resources involved in an incident.

2. Cloud Run Mitigations

Located at skills/gcp-playbooks/references/googleapis.com/run.md. Maps generic mitigations (rollback, fix forward, upsize, restart) to Cloud Run actions.

3. GKE Mitigations

Located at skills/gcp-playbooks/references/googleapis.com/container.md. Maps generic mitigations (rollback, fix forward, upsize, quarantine, restart) to GKE / Kubernetes actions.

4. Cloud Build Investigation (Draft)

Located at skills/gcp-playbooks/references/googleapis.com/cloudbuild.md.

Core Pattern

1. Identify the Scope: Use gcp_resource_manager or the discovery playbook to find the relevant project and resources.
2. Consult the Playbook: Look for a matching playbook in {SKILL_DIR}/references/googleapis.com/ using the short API name.
3. Execute Actions: Follow the gcloud/kubectl commands suggested in the playbook.
4. Validate: Verify the state of the system after each action.

References

• GCP Documentation