get-capacity-recommendations

X-McKay's avatarfrom X-McKay

Analyze cluster capacity and provide scaling recommendations. Forecasts resource needs, identifies imbalances, and suggests actions. Use for capacity planning and proactive scaling. Keywords: capacity, planning, scaling, forecast, recommendation, resources.

⭐1stars🔀0forks📁View on GitHub🕐Updated Jan 11, 2026⬇️1downloads
[Cloud Infrastructure]

When & Why to Use This Skill

This Claude skill automates cluster capacity analysis and resource forecasting to ensure optimal infrastructure performance. By identifying resource imbalances and predicting future needs for CPU, memory, and storage, it provides actionable scaling recommendations that help prevent bottlenecks, reduce downtime, and optimize cloud costs through proactive planning.

Use Cases

  • Proactive Capacity Planning: Forecast resource exhaustion dates (e.g., 30 or 90-day outlook) based on historical growth rates to plan infrastructure upgrades before they become critical.
  • Automated Scaling Recommendations: Receive specific, data-driven advice on when to add nodes or increase resource limits (CPU/Memory) based on real-time utilization trends.
  • Cluster Rebalancing: Identify imbalanced nodes where resource distribution is uneven, allowing SREs to redistribute workloads and improve overall fleet efficiency.
  • Budget & Resource Forecasting: Project future infrastructure requirements to assist in financial planning and cloud commitment strategies.
nameget-capacity-recommendations
description>
and proactive scaling. Keywordscapacity, planning, scaling, forecast,
domaingeneral
categoryanalytics
requires-approvalfalse
confidence0.85
mcp-servers[]

Get Capacity Recommendations

Preconditions

Before applying this skill, verify:

  • Resource usage data available for nodes
  • Historical data exists (for forecasting)
  • Cluster metrics accessible

Actions

1. Collect Current Resource Usage

Gather usage data for all resources:

resources:
  - type: cpu
    unit: cores
    nodes: [node1, node2, node3]
  - type: memory
    unit: GB
    nodes: [node1, node2, node3]
  - type: storage
    unit: GB
    nodes: [node1]
  - type: pods
    unit: count
    nodes: [node1, node2, node3]

2. Calculate Cluster Totals

Aggregate across nodes:

total_capacity = sum(node.capacity for node in nodes)
total_used = sum(node.used for node in nodes)
utilization = total_used / total_capacity

3. Forecast Future Usage

Project usage based on growth rate:

days_of_data = (now - earliest_data).days
growth_rate = (latest_usage - earliest_usage) / days_of_data

forecast_30d = current_usage + (growth_rate * 30)
forecast_90d = current_usage + (growth_rate * 90)

4. Calculate Days Until Threshold

Project when thresholds will be reached:

warning_threshold = 0.80 * total_capacity
critical_threshold = 0.90 * total_capacity

days_to_warning = (warning_threshold - current_usage) / growth_rate
days_to_critical = (critical_threshold - current_usage) / growth_rate

5. Check Node Balance

Identify imbalanced nodes:

avg_utilization = mean(node.utilization for node in nodes)
imbalanced = [n for n in nodes if abs(n.utilization - avg_utilization) > 0.2]

6. Generate Recommendations

Create actionable recommendations:

recommendations:
  - type: scale_up
    urgency: high
    resource: memory
    reason: "Days to critical: 15"
  - type: rebalance
    urgency: medium
    nodes: [node1, node3]
    reason: "Utilization imbalance > 20%"

Success Criteria

The skill succeeds when:

  • All resource types analyzed
  • Forecast calculated with confidence
  • Recommendations prioritized by urgency

Failure Handling

If analysis fails:

  1. Insufficient data: Return only current state
  2. Missing metrics: Use available data, note gaps
  3. Calculation error: Return partial recommendations

Examples

Input Context:

{
  "cluster": "production",
  "forecast_days": 30
}

Expected Output:

{
  "current_state": {
    "cpu_utilization": 0.65,
    "memory_utilization": 0.78,
    "storage_utilization": 0.45,
    "pod_count": 150
  },
  "forecast_30d": {
    "memory_utilization": 0.92,
    "days_to_warning": 12,
    "days_to_critical": 25
  },
  "recommendations": [
    {
      "type": "scale_up",
      "resource": "memory",
      "urgency": "high",
      "confidence": 0.85,
      "action": "Add 32GB memory or 1 new node",
      "reason": "Memory will reach 80% in 12 days at current growth"
    }
  ],
  "node_balance": {
    "balanced": false,
    "imbalanced_nodes": ["worker-3"],
    "suggestion": "Rebalance workloads from worker-3"
  }
}