Quota Throttle Expert
Diagnose AOAI deployment TPM exhaustion behind Foundry hosted agents — pull capacity vs utilization, identify burst patterns, recommend scale-up or PTU migration.
Install
$ agentstack add skill-aiappsgbb-awesome-gbb-quota-throttle-expert ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Links to your public security report.[](https://agentstack.voostack.com/security/report/skill-aiappsgbb-awesome-gbb-quota-throttle-expert) Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
quotathrottleexpert
When to use
- A user reports 429s from a Foundry hosted agent's underlying model
- App Insights shows
Quota exceeded/OperationLimitExceededfor a deployment - A scheduled task flags TPM utilization > 80% for a deployment
Investigation flow
- Identify the deployment:
``bash az cognitiveservices account deployment show \ --name --resource-group \ --deployment-name -o json ` Capture sku.capacity (TPM in thousands), sku.name (e.g. Standard, GlobalStandard, ProvisionedManaged`).
- Pull TPM utilization for the deployment over the failing window
from App Insights (assumes foundry-observability is wired): ``kql customMetrics | where timestamp > ago(2h) | where name == "gen_ai.client.token.usage" | where customDimensions["gen_ai.system"] == "az.ai.openai" | where customDimensions["gen_ai.response.model"] == "" | summarize sum(valueSum) by bin(timestamp, 1m) | order by timestamp asc ``
- Compare against the deployment's capacity:
sku.capacityof 100 → 100k TPM- Multiply by 60 → 6M tokens/minute capacity
- Identify peak minutes vs limit
- Classify the throttle:
| Pattern | Cause | Recommendation | |---|---|---| | Sustained peak > 80% capacity | Workload outgrew baseline | Increase sku.capacity | | Spiky peaks 200%+ for 1-2 min, calm baseline | Burst pattern | Consider PTU (Provisioned Managed) for predictable burst headroom | | One client dominates | Single noisy neighbor | Add per-spoke rate-limit at Citadel APIM gateway (hand off to apim_throttle_expert) | | Region cap hit (Standard SKU) | Regional quota | Request quota increase via portal or migrate to GlobalStandard |
- Cross-check with Azure Monitor's
TokenTransactionmetric (if the
user has the Microsoft.CognitiveServices/accounts resource in the monitored RG).
- Output: peak TPM, capacity, % utilization, classification, ONE recommended action with the exact CLI command (do NOT execute — review mode).
Tools
RunAzCliReadCommandsQueryAppInsightsByAppId
Safety
- Never modify the deployment SKU —
sku.capacitychanges are reviewed by the human (cost impact) - Never read AOAI account keys
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: aiappsgbb
- Source: aiappsgbb/awesome-gbb
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.