This is an old revision of the document!
Azure Open AI Chatbot Deployment Checklist
✅ Azure OpenAI Chatbot Deployment Checklist
1. Local Build and Testing
2. Azure Resources Setup
Resource Group created or selected.
Azure Container Registry (ACR) created and admin access enabled (if not using managed identity).
Azure OpenAI Service created with deployed model (e.g., gpt-35-turbo).
3. Docker Image Build and Push
Docker image built with docker build -t <image> ..
Docker image tagged with ACR address.
Docker image pushed to ACR.
4. Deployment Parameters
Parameter values prepared, either inline or in containerapp-params.json.
Verified
API Key, OpenAI endpoint, and deployment ID.
Verified ACR server, username, and password (if required).
5. Bicep Deployment
Bicep template validated with az bicep build.
Deployment executed with either inline parameters or parameter file.
Verified Container App is running in Azure Portal.
6. Testing After Deployment
7. Security and Scaling
Confirmed ACR access is secured.
Configured auto-scaling (if needed).
Applied secure ingress settings (HTTPS, auth if needed).
8. Observability
9. Log Analytics and Monitoring
AppContainerAppConsoleLogs_CL \\
| where ContainerAppName_s == "<your-container-app-name>"\\
| sort by TimeGenerated desc
10. Alerts Setup
Example KQL for HTTP 500 Errors:
AppContainerAppConsoleLogs_CL
| where Log_s contains "500"
| summarize Count = count() by bin(TimeGenerated, 5m)
Example Alert Setup Steps:
Go to Azure Monitor > Alerts.
Click + New Alert Rule.
Select Resource: Your Container App or Log Analytics Workspace.
Condition: Add a metric or log query condition.
Action Group: Create or select action groups (email, Teams, webhook, etc.).
Alert Details: Name and severity.