Skip to content

quota_check.md

Latest commit

 

History

History
129 lines (85 loc) · 3.67 KB

quota_check.md

File metadata and controls

129 lines (85 loc) · 3.67 KB
Raw

Check Quota Availability Before Deployment

Before deploying the accelerator, ensure sufficient quota availability for the required models.

For the GPT chat model — set the capacity to at least 150k tokens for optimal performance. For the embedding model, set at least 80k tokens.

Login if you have not done so already

az login

If using VS Code Web:

az login --use-device-code

📌 Default Models & Capacities

The quota check script (infra/scripts/pre-provision/quota_check_params.sh) checks the following models by default:

gpt-5.2:150, gpt-4o-mini:150, gpt-4:150, text-embedding-3-small:80

Note: This solution deploys gpt-5.2 (version 2025-12-11) for chat/insights and text-embedding-3-small for embeddings, both on the GlobalStandard SKU. The additional models in the default check are included for convenience when planning capacity.

📌 Default Regions

australiaeast, swedencentral, southeastasia

Usage Scenarios

  • No parameters passed → Default models and capacities will be checked in default regions.
  • Only model(s) provided → The script will check for those models in the default regions.
  • Only region(s) provided → The script will check default models in the specified regions.
  • Both models and regions provided → The script will check those models in the specified regions.
  • --verbose passed → Enables detailed logging output for debugging and traceability.

Input Formats

Use the --models, --regions, and --verbose options for parameter handling:

✔️ Run without parameters to check default models & regions:

./quota_check_params.sh

✔️ Enable verbose logging:

./quota_check_params.sh --verbose

✔️ Check specific model(s) in default regions:

./quota_check_params.sh --models gpt-5.2:150

✔️ Check default models in specific region(s):

./quota_check_params.sh --regions southeastasia,swedencentral

✔️ Passing both models and regions:

./quota_check_params.sh --models gpt-5.2:150 --regions southeastasia,swedencentral

✔️ Multiple models with a single region:

./quota_check_params.sh --models gpt-5.2:150,text-embedding-3-small:80 --regions swedencentral --verbose

Running the Quota Check

If using Azure Portal and Cloud Shell

  1. Navigate to the Azure Portal.

  2. Click on Azure Cloud Shell in the top right navigation menu.

  3. Download and run the quota check script:

    curl -L -o quota_check_params.sh "https://raw.githubusercontent.com/microsoft/Conversation-Knowledge-Mining-Solution-Accelerator/main/infra/scripts/pre-provision/quota_check_params.sh"
chmod +x quota_check_params.sh
./quota_check_params.sh

If using VS Code or Codespaces

  1. Open the terminal in VS Code or Codespaces.

  2. If you're using VS Code, select Git Bash / bash from the terminal dropdown.

  3. Navigate to the script folder and make the script executable:

    cd infra/scripts/pre-provision
chmod +x quota_check_params.sh

  4. Run the script:

    ./quota_check_params.sh

  5. If you see the error bash: az: command not found, install Azure CLI:

    curl -sL https://aka.ms/InstallAzureCLIDeb | sudo bash
az login

    Note: Use az login --use-device-code in VS Code Web.

  6. Rerun the script after installing Azure CLI.

The final output lists regions with available quota. You can select any of these regions for deployment.

Next Steps

Return to the Deployment Guide to continue with your deployment.