/capacity-check

description: Capacity forecast for cloud resources, scoped to a resource type or covering everything connected
argument-hint: "[resource_type]"
arguments: [resource_type]

Capacity Check

Cross-vendor cloud capacity forecast: identifies over-/under-provisioned resources and projects near-term capacity risk based on observed growth trends — pulled from whatever cloud platforms (Azure, DigitalOcean) the org has connected through the gateway.

Prerequisites

• WYRE MCP Gateway connected (`conduit`) with at least one cloud platform

connector. Without one, there is no capacity data to report.

Steps

1. **Discover available tools.** Call `conduit__search_tools` to determine which cloud platform connector(s) are live and their actual tool names (e.g. `azure-mcp__quota`, `azure-mcp__monitor`, `digitalocean__list_droplets`, `digitalocean__list_kubernetes_clusters`, `digitalocean__list_databases`). Never assume a specific vendor's tool surface — cover every connected platform.

2. **Resolve scope.** Parse `resource_type` (optional; default: all resource types). Accepted values include `compute`, `storage`, `database`, or any other resource category the connected platform(s) expose. If omitted, cover everything connected.

3. **Pull resource inventory and utilization.** Per the `cloud-capacity-planning` skill's per-platform signal mapping, classify each resource: over-provisioned, right-sized, under-provisioned, or insufficient-data.

4. **Apply trend discipline.** Pull the longest available utilization history and require a sustained multi-window trend before treating anything as a genuine capacity risk — a single spike is not a finding. State the observation window used.

5. **Forecast near-term risks.** For resources trending toward a critical threshold, project an exhaustion window as a range (not a false-precision date) and classify it near-term (inside the planning horizon) or longer-horizon (worth tracking, not urgent).

6. **Return the report**, separating near-term risks, longer-horizon trends, over-provisioned/right-sizing opportunities, and right-sized resources.

Arguments

• `resource_type` (optional; default: all) — Scope the forecast to a

resource category, e.g. `compute`, `storage`, `database`. Omit to cover every resource type the connected platform(s) expose.

Examples

Full capacity forecast

/cloudops-pack:capacity-check

Scoped to compute

/cloudops-pack:capacity-check compute

Scoped to database

/cloudops-pack:capacity-check database

Output

================================================================================
Capacity Forecast — scope: [resource_type or "all"]
================================================================================

NEAR-TERM CAPACITY RISKS ([N])
--------------------------------------------------------------------------------
[resource] ([platform]) — [current utilization], trending [N]%/week over [N]
weeks, projected exhaustion: [date range]

LONGER-HORIZON / MONITOR ([N])
--------------------------------------------------------------------------------
[same shape, outside the near-term planning horizon]

OVER-PROVISIONED — RIGHT-SIZING OPPORTUNITY ([N])
--------------------------------------------------------------------------------
[resource] ([platform]) — [current utilization], suggested action

RIGHT-SIZED ([N])
--------------------------------------------------------------------------------
[N] resources confirmed correctly sized across [platforms]
================================================================================

Error Handling

• **No cloud platform connector connected:** Report plainly that a capacity

forecast can't be produced without one, and stop rather than fabricating resource data.

• **Invalid or unrecognized `resource_type`:** Note the mismatch and report

what resource types are actually available from the connected platform(s) instead of silently returning an empty result.

• **Platform connected but no historical/trend data exposed:** Report

current utilization as a point-in-time snapshot and state plainly that a trend-based forecast wasn't possible.

Related Commands

• `/cloudops-pack:network-sweep` - Network device/link health, the

network-side counterpart to this cloud-side forecast

• `/cloudops-pack:cost-report [window]` - Cloud spend anomaly and

reclaimable-cost report — a related but distinct judgment from capacity