agentleFS
ricmmartinsricmmartins/azure-sre-agent-skills/skills/03-capacity-planning/SKILL.md

capacity-planning

Skill70 starsChanged 4 months agoMITView on GitHub

Assess Azure resource capacity, quota utilization, and growth trends to prevent outages from resource exhaustion. Use when asked about capacity, quotas, scaling readiness, load testing prep, Black Friday readiness, or growth projections.

7 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After hitting a quota or throttling limit

    line 19
  2. Distinctive:

    Check current quota usage across all critical resource providers.

    line 31
  3. Distinctive:

    If the user specified a growth factor, multiply current usage and check against limits

    line 116
  4. Distinctive:

    If the user provides a list of accepted exceptions, do not flag those items.

    line 148
  5. Distinctive:

    When exceptions are provided:

    line 158
  6. Very specific:

    Use GetAzCliHelp to validate the command syntax before suggesting

    line 193

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Capacity Planning

Purpose

Analyze current resource utilization, quota consumption, and growth trends to predict capacity risks and recommend scaling actions before limits are hit.

When to use this skill

  • User asks "will our infrastructure handle the traffic spike?"
  • User asks about quotas, limits, or capacity
  • Pre-event planning (Black Friday, product launch, marketing campaign)
  • Quarterly capacity review
  • After hitting a quota or throttling limit

Pre-check

Confirm with the user:

  • Scope: which subscriptions and regions
  • Time horizon: when is the expected load increase? (default: 30 days)
  • Expected growth factor (e.g., 2x, 5x, 10x current load)
  • Critical workloads to prioritize

Assessment procedure

Step 1: Quota utilization

Check current quota usage across all critical resource providers.

az vm list-usage --location <region> -o table
az network list-usages --location <region> -o table
az storage account list --query "length(@)"

Check for:

  • vCPU quotas: Total and per-family (Dv5, Ev5, etc.) — flag if > 70% utilized
  • Public IPs: Usage vs. limit
  • Load balancers: Usage vs. limit
  • Network interfaces: Usage vs. limit
  • Storage accounts per subscription: Usage vs. 250 limit
  • App Service plans per region
  • AKS clusters per subscription

Build a quota utilization table. Use Unicode emoji characters directly (🟢🟡🔴) — never use emoji shortcodes like :green_circle: or :red_circle:.

ResourceRegionUsedLimitUtilizationRisk
Total vCPUsEast US8510085%🔴
Dv5 vCPUsEast US325064%🟡
Public IPsEast US82040%🟢

Step 2: Compute capacity

Analyze current compute utilization and headroom.

  1. VM utilization (last 14 days):
    • Average and P95 CPU across all VMs
    • Average and P95 memory (if available via Log Analytics)
    • Flag VMs consistently > 80% CPU or memory
  2. App Service plan utilization:
    az appservice plan list --query "[].{name:name, sku:sku.name, workers:numberOfWorkers, maxWorkers:maximumElasticWorkerCount}" -o table
    • Current instance count vs. max
    • CPU/memory % of the plan
    • Autoscale rules and current headroom
  3. AKS cluster capacity:
    az aks list --query "[].{name:name, nodeCount:agentPoolProfiles[0].count, maxCount:agentPoolProfiles[0].maxCount, vmSize:agentPoolProfiles[0].vmSize}" -o table
    • Node count vs. max count
    • Pod density per node
    • Cluster autoscaler configuration
  4. Container Apps:
    • Current replicas vs. max replicas
    • Autoscale rules (HTTP concurrency, CPU, custom)

Step 3: Data layer capacity

  1. SQL Database:
    • DTU/vCore utilization (avg and P95)
    • Storage usage vs. max size
    • Connection count vs. limit
  2. Cosmos DB:
    • RU consumption vs. provisioned (or autoscale max)
    • Storage per container
    • 429 (throttling) rate from metrics
  3. Redis Cache:
    • Memory usage %
    • Connection count vs. max
    • Server load %
  4. Storage accounts:
    • Ingress/egress patterns
    • Transaction rates vs. limits
    • Blob count growth rate

Step 4: Networking capacity

  1. ExpressRoute / VPN Gateway: Bandwidth utilization %
  2. Application Gateway / Front Door: Connection count, throughput
  3. NAT Gateway: SNAT port utilization
  4. DNS zones: Query volume trends

Step 5: Growth projection

Based on the data collected:

  1. Plot utilization trends for the last 90 days
  2. Apply linear projection to estimate when each resource hits 80% and 100%
  3. If the user specified a growth factor, multiply current usage and check against limits

Build a risk timeline:

ResourceCurrentAt 2x LoadAt 5x LoadHits LimitAction Needed
vCPUs East US85/100170 ❌425 ❌NowQuota increase
SQL DTU60%120% ❌300% ❌At 1.7xScale up tier
AKS nodes5/1010/10 ⚠️25 ❌At 2xIncrease max
Redis memory45%90% ⚠️225% ❌At 2.2xUpgrade SKU

Step 6: Quota increase requests

For any quota that needs increasing, generate the correct request command:

# Request a quota increase via the Azure portal Support + Troubleshooting flow,
# or use the REST API directly:
az rest --method patch \
  --url "https://management.azure.com/subscriptions/<sub-id>/providers/Microsoft.Capacity/resourceProviders/Microsoft.Compute/locations/<region>/serviceLimits/StandardDv5Family?api-version=2020-10-25" \
  --body '{"properties":{"limit":{"limitObjectType":"LimitValue","value":<new-limit>}}}'

Note: The az quota CLI extension is in preview and syntax may change. The REST API approach above is stable. Alternatively, guide users to submit quota requests via the Azure portal: Help + support > New support request > Service and subscription limits (quotas).

Step 7: Recommendations

For each risk identified:

  1. Immediate: Quota increase requests, SKU upgrades
  2. Short-term (1-2 weeks): Autoscale configuration, caching layers
  3. Long-term: Architecture changes (sharding, async patterns, CDN)

Accepted exceptions (optional)

If the user provides a list of accepted exceptions, do not flag those items. Instead, note them in the report as Accepted Exception with the reason provided.

Example format the user may provide:

CheckReason
1.1 vCPU quota utilizationQuota increase already submitted, awaiting approval
2.3 AKS cluster capacityCluster is scheduled for decommission next month
4.2 Application Gateway throughputTraffic is seasonal, current capacity is intentional

When exceptions are provided:

  • Skip the flagged checks in scoring
  • List them in a separate "Accepted Exceptions" section at the end of the report
  • Recalculate the overall score excluding excepted checks

Expected output

Report header (mandatory — use this exact format)

Capacity Planning Report

FieldValue
Subscription(name + ID)
Assessment DateYYYY-MM-DD
Region(s) Assessed(list)

Capacity summary

MetricValue
🏗️ Resources scannedX
⚠️ Resources at risk (>70% capacity)Y
🔴 Resources critical (>90% capacity)Z
📈 Projected limit breach within 30 daysW resources

Risk matrix

Table with all resources, current utilization, projected utilization at specified growth factor, and recommended action.

Scaling runbook

Step-by-step actions to prepare for the expected load increase, ordered by priority and dependency.

Remediation guidance

For each capacity risk identified, include in the output:

  1. The specific az CLI command to remediate (quota increase, scale-up, autoscale config)
  2. Use GetAzCliHelp to validate the command syntax before suggesting
  3. The official Microsoft Learn documentation link

References

Sample output

The following is a redacted example of what the report looks like when run against a subscription.

Capacity Planning Report

FieldValue
Subscriptioncontoso-prod-001 (a1b2c3d4-e5f6-7890-abcd-ef1234567890)
Assessment Date2026-07-15
Region(s) AssessedEast US, West Europe

Capacity summary

MetricValue
🏗️ Resources scanned47
⚠️ Resources at risk (>70% capacity)5
🔴 Resources critical (>90% capacity)2
📈 Projected limit breach within 30 days1 resource

Risk matrix (sample)

ResourceRegionCurrentAt 2x LoadHits LimitAction Needed
Total vCPUsEast US85/100 (85%)170 ❌NowQuota increase
Dv5 vCPUsEast US32/50 (64%)64 ⚠️At 1.6xQuota increase
SQL DTU sql-contoso-prodEast US78% avg156% ❌At 1.3xScale up tier
AKS nodes aks-app-prodWest Europe7/1014 ❌At 1.4xIncrease max count
Public IPsEast US8/20 (40%)16 🟢At 2.5xMonitor

Scaling runbook (sample)

PriorityActionResourceEstimated Time
1Request vCPU quota increase to 200East US subscription1–3 business days
2Scale SQL DB from S3 to S4sql-contoso-prod15 min (online)
3Increase AKS max node count to 20aks-app-prod5 min

Remediation guidance (sample)

# Request vCPU quota increase via REST API
az rest --method patch \
  --url "https://management.azure.com/subscriptions/a1b2c3d4-.../providers/Microsoft.Capacity/resourceProviders/Microsoft.Compute/locations/eastus/serviceLimits/StandardDv5Family?api-version=2020-10-25" \
  --body '{"properties":{"limit":{"limitObjectType":"LimitValue","value":200}}}'

# Scale up SQL Database
az sql db update --resource-group rg-data-prod --server sql-contoso-prod --name appdb --service-objective S4

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.