Agent Skill · Qovery

qovery-optimize

Reduces Kubernetes cluster and application costs on Qovery. Analyzes historical resource consumption, factors in the user's business context (seasonal patterns, growth stage, reliability requirements), estimates external resource costs from public cloud pricing, and proposes right-sizing, autoscaling, environment scheduling, spot instances, and database mode changes. Generates a cost report with CSV export and applies changes via CLI+API or Terraform. Use when the user asks to reduce costs, right-size, or optimize Qovery resource spend.

View SKILL.md on GitHub → Source repository Provider profile
Provider: Qovery Path in repo: qovery-optimize/SKILL.md

Skill body

Qovery Optimize Skill

This skill reduces Kubernetes infrastructure costs on Qovery via right-sizing, autoscaling, environment scheduling, spot instances, and database mode changes.

It is NOT about blindly reducing everything to minimum. It is intelligent optimization that:

When to Use This Skill

Trigger phrases:

For slow deployments use qovery-speedup. For deployment failures use qovery-troubleshoot.

Workflow checklist

Cost Optimization Progress:
- [ ] Phase 1 — Context gathering (auth, business context, resource metrics)
- [ ] Phase 2 — Analysis across 7 optimization dimensions
- [ ] Phase 3 — Generate cost report (Markdown + CSV)
- [ ] Phase 4 — Apply approved changes (CLI+API or Terraform) + USER CONFIRMATION per change
- [ ] Phase 5 — Set up ongoing monitoring & follow-up
- [ ] Phase 6 — Document seasonal & special considerations

Reference materials (load on demand)

Phase File Purpose
Console URL reference/console-url-detection.md Extract IDs from a Qovery Console URL
Auth reference/auth.md API token flow
Phase 1 reference/phase1-context-gathering.md Inventory, business context, resource metrics queries
Phase 2 reference/phase2-optimization-dimensions.md Right-sizing, autoscaling, DB mode, scheduling, cluster, build, external resources
Phase 3 reference/phase3-cost-report.md Report template (Markdown + CSV)
Phase 4 reference/phase4-apply-changes.md CLI+API + Terraform application paths, with per-change confirmation
Phase 5 reference/phase5-monitoring.md Ongoing dashboards, alert thresholds, follow-up cadence
Phase 6 reference/phase6-seasonal.md Black Friday / EOY / launch handling

Cloud pricing reference

For external-resource cost estimation, consult the per-provider pricing tables. These are point-in-time public list prices — verify against the provider’s pricing page before quoting savings.

Provider File
AWS (EC2 / EKS / RDS / ElastiCache / infra fixed) reference/pricing/aws.md
GCP (GKE / Cloud SQL / Cloud NAT / disks) reference/pricing/gcp.md
Azure (AKS / VMs / Postgres / LB) reference/pricing/azure.md
Scaleway (Kapsule / DB / instances) reference/pricing/scaleway.md

Quick reference

MCP queries

# Cost analysis
"Show me monthly spending"
"What are my highest cost services?"
"Compare costs this month vs last month"

# Resource analysis
"Show CPU usage across all services"
"Find over-provisioned services"
"Find underutilized resources"

# Inactive resources
"Show environments inactive for 24 hours"
"List unused databases"
"Find idle applications"

# Actions
"Stop all development environments for the weekend"
"Scale down the backend to 250m CPU"
"Stop all non-production environments"

# Recommendations
"Should I scale up or down?"
"Optimize resource allocation for my-api"

API endpoints

# Base URL: https://api.qovery.com   Auth: Authorization: Token $QOVERY_API_TOKEN

# Costs
GET /organization/{orgId}/currentCost
GET /organization/{orgId}/cluster/{clusterId}/currentCost
GET /organization/{orgId}/invoice

# Metrics
GET /cluster/{clusterId}/metrics?query={promql}&range={duration}&step={interval}

# Service configuration
GET /application/{appId}                                    CPU, memory, instances
PUT /application/{appId}                                    Update CPU, memory, instances

# Cloud provider instance types
GET /organization/{orgId}/cloudProvider/aws/instanceType
GET /organization/{orgId}/cloudProvider/gcp/instanceType
GET /organization/{orgId}/cloudProvider/azure/instanceType
GET /organization/{orgId}/cloudProvider/scaleway/instanceType

# Environment management
POST /environment/{envId}/deploy                            Redeploy after changes
POST /environment/{envId}/stop
POST /environment/{envId}/restart

CLI commands

qovery cluster list                  # Cluster overview
qovery service list                  # Service overview with resource info
qovery status                        # Current status
qovery application env list          # Check env var configuration

Skill frontmatter

license: MIT compatibility: opencode metadata: {"audience"=>"developers", "workflow"=>"optimization"}