Deploy to Azure

This guide walks you through deploying an Azure Functions Python app with a diagnostics-first workflow. You will run azure-functions-doctor first, confirm the project is healthy, then provision Azure resources and publish.

Who this guide is for

You write Python and use pip, but Azure is new to you. This guide is for first-time Azure users who want copy-paste commands and clear checkpoints.

What you are deploying

azure-functions-doctor is a CLI diagnostics tool, not a runtime framework. It checks Azure Functions project health before deployment so you catch problems early.

In this guide you will:

• Run doctor on a healthy sample project (examples/v2/http-trigger)
• Compare that with a broken sample (examples/v2/broken-missing-host-json)
• Generate text, JSON, SARIF, and JUnit outputs for local and CI use
• Deploy only after checks pass

Azure concepts you need for this guide

New to Azure plans? Read Choose an Azure Functions Hosting Plan.

Term	What it means
Function App	Your deployed Azure Functions application (like a cloud-hosted Python app).
Hosting plan	Compute model that controls scaling, cold starts, timeout, and cost.
Resource Group	A container for related Azure resources; delete it to clean up everything.
Storage Account	Required backing resource for Azure Functions state and operations.

Recommended plan for this repo

Default plan	Flex Consumption
Why	Lowest friction for first deployment, scale-to-zero, and enough timeout for typical Python HTTP samples.
Switch to Premium if	You need faster cold starts, always-warm instances, or very large dependency footprints.

Before you start

Requirement	How to check	Install if missing
Azure account	portal.azure.com	Create free account
Azure CLI	az --version	Install Azure CLI
Azure Functions Core Tools v4	func --version	Install Core Tools
Python 3.10-3.13	python3 --version	python.org
Doctor CLI v0.16.2	pip show azure-functions-doctor	python3 -m pip install azure-functions-doctor==0.16.2

Install doctor if needed:

python3 -m pip install --upgrade pip
python3 -m pip install azure-functions-doctor==0.16.2

Read these warnings before provisioning

1. Storage account names must be globally unique (lowercase letters and numbers only, 3–24 characters).
2. Keep all resources in the same Azure region to avoid avoidable latency and provisioning issues.
3. Local settings do not auto-sync to Azure; use app settings after deployment when needed.
4. First publish can take a few minutes because Azure performs a remote build for Python dependencies.
5. Deleting local files does not delete Azure resources; remove the resource group to stop billing.
6. Doctor validates your function project, not your Azure subscription state; permission, quota, policy, and region availability issues can still fail deployment.

Run doctor before you deploy

This is the core workflow for this repository: diagnose first, deploy second.

Healthy project check:

azure-functions doctor examples/v2/http-trigger --profile minimal

Representative text output:

Azure Functions Doctor
Path: /data/GitHub/azure-functions-doctor/examples/v2/http-trigger

Python Env
[PASS] Python version: Python 3.11.11
[PASS] requirements.txt: found
[PASS] azure-functions package: declared in requirements.txt

Project Structure
[PASS] host.json: found
[PASS] host.json version: "2.0"

Doctor summary (to see all details, run azure-functions doctor -v):
  0 fails, 0 warnings, 5 passed
Exit code: 0

Broken project detection (missing host.json):

azure-functions doctor examples/v2/broken-missing-host-json --profile minimal

Representative text output:

Azure Functions Doctor
Path: /data/GitHub/azure-functions-doctor/examples/v2/broken-missing-host-json

Python Env
[PASS] Python version: Python 3.11.11
[PASS] requirements.txt: found
[PASS] azure-functions package: declared in requirements.txt

Project Structure
[FAIL] host.json: file not found (fail)
[FAIL] host.json version: host.json missing or invalid (fail)

Doctor summary (to see all details, run azure-functions doctor -v):
  2 fails, 0 warnings, 3 passed
Exit code: 1

Machine-readable output formats:

# JSON
azure-functions doctor examples/v2/http-trigger --profile minimal --format json --output doctor-report.json

# SARIF (for code scanning / security tooling)
azure-functions doctor examples/v2/http-trigger --profile minimal --format sarif --output doctor-report.sarif

# JUnit (for CI test dashboards)
azure-functions doctor examples/v2/http-trigger --profile minimal --format junit --output doctor-report.xml

# Summary JSON sidecar (quick pass/fail counts)
azure-functions doctor examples/v2/http-trigger --profile minimal --summary-json doctor-summary.json

Representative JSON output (doctor-report.json):

{
  "metadata": {
    "tool_version": "0.16.2",
    "target_path": "/data/GitHub/azure-functions-doctor/examples/v2/http-trigger"
  },
  "results": [
    {
      "title": "Python Env",
      "status": "pass",
      "items": [
        {
          "label": "Python version",
          "status": "pass"
        }
      ]
    }
  ]
}

Representative SARIF output (doctor-report.sarif):

{
  "version": "2.1.0",
  "$schema": "https://json.schemastore.org/sarif-2.1.0.json",
  "runs": [
    {
      "tool": {
        "driver": {
          "name": "azure-functions-doctor",
          "version": "0.16.2"
        }
      },
      "results": []
    }
  ]
}

Representative JUnit output (doctor-report.xml):

<?xml version="1.0" encoding="utf-8"?>
<testsuite name="func-doctor" tests="5" failures="0" skipped="0" time="0.021">
  <testcase classname="Python Env" name="Python version" />
  <testcase classname="Project Structure" name="host.json" />
</testsuite>

Representative summary sidecar content:

{"passed":5,"warned":0,"failed":0}

When checks fail, doctor exits with code 1. Use that exit code to block deployment.

Deploy the healthy example

This section deploys examples/v2/http-trigger on Flex Consumption.

Step 1 — Set deployment variables

SUBSCRIPTION_ID="00000000-0000-0000-0000-000000000000"
RESOURCE_GROUP="rg-doctor-http"
LOCATION="koreacentral"
STORAGE_ACCOUNT="stdoctorhttp$(date +%s | tail -c 6)"
FUNCTIONAPP_NAME="func-doctor-http"

Step 2 — Sign in and select subscription

az login
az account set --subscription "$SUBSCRIPTION_ID"

Step 3 — Confirm Flex region support

az functionapp list-flexconsumption-locations -o table

Step 4 — Re-run doctor on the healthy sample (deployment gate)

azure-functions doctor examples/v2/http-trigger --profile minimal

Proceed only if the command exits with 0.

Step 5 — Create resource group

az group create --name "$RESOURCE_GROUP" --location "$LOCATION"

Step 6 — Create storage account

az storage account create \
  --name "$STORAGE_ACCOUNT" \
  --resource-group "$RESOURCE_GROUP" \
  --location "$LOCATION" \
  --sku Standard_LRS

Step 7 — Create Function App (Flex Consumption)

az functionapp create \
  --name "$FUNCTIONAPP_NAME" \
  --resource-group "$RESOURCE_GROUP" \
  --storage-account "$STORAGE_ACCOUNT" \
  --flexconsumption-location "$LOCATION" \
  --runtime python \
  --runtime-version 3.11

Step 8 — Publish the healthy example

func azure functionapp publish "$FUNCTIONAPP_NAME" --python --source examples/v2/http-trigger

If your Core Tools version does not support --source, run this equivalent command instead:

(cd examples/v2/http-trigger && func azure functionapp publish "$FUNCTIONAPP_NAME" --python)

Step 9 — Verify the deployed endpoint

BASE_URL="https://$FUNCTIONAPP_NAME.azurewebsites.net"
curl -i "$BASE_URL/api/HttpExample?name=Azure"
curl -i "$BASE_URL/api/HttpExample"

Expected behavior: both calls return HTTP 200 with plain-text responses.

Use doctor in CI

Run diagnostics before publish, and fail the pipeline when doctor finds required check failures.

Example GitHub Actions workflow:

name: deploy-with-doctor

on:
  push:
    branches: [main]

jobs:
  verify-and-deploy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4

      - uses: actions/setup-python@v5
        with:
          python-version: '3.11'

      - name: Install tooling
        run: |
          python -m pip install --upgrade pip
          pip install azure-functions-doctor==0.16.2

      - name: Run doctor (text)
        run: azure-functions doctor . --profile minimal

      - name: Run doctor (JSON)
        run: azure-functions doctor . --profile minimal --format json --output doctor-report.json

      - name: Run doctor (SARIF)
        run: azure-functions doctor . --profile minimal --format sarif --output doctor-report.sarif

      - name: Run doctor (JUnit)
        run: azure-functions doctor . --profile minimal --format junit --output doctor-report.xml

      - name: Run doctor (summary sidecar)
        run: azure-functions doctor . --profile minimal --summary-json doctor-summary.json

      - name: Publish
        if: success()
        run: func azure functionapp publish "$FUNCTIONAPP_NAME" --python

If you need a different plan

This guide uses Flex Consumption by default. If you need Premium or Dedicated, keep the same diagnostics workflow and replace only the Function App provisioning commands.

See Choose an Azure Functions Hosting Plan for plan-by-plan command blocks.

Troubleshooting

Doctor says pass, but deployment fails

Symptom	Usually means	How to fix
AuthorizationFailed or SubscriptionNotFound	Subscription/permissions issue	Confirm account and subscription: az account show, then az account set --subscription "$SUBSCRIPTION_ID"
StorageAccountAlreadyTaken	Storage account name collision	Generate a new globally unique value for $STORAGE_ACCOUNT
LocationNotAvailableForResourceType	Flex unavailable in that region	Check available regions: az functionapp list-flexconsumption-locations -o table
Publish fails with remote build errors	Azure-side build issue or missing dependency metadata	Verify requirements.txt, retry publish, then inspect deployment logs

Doctor validates project structure and configuration; it does not validate Azure account state, RBAC, quota, policies, or regional service availability.

Doctor says fail, but project works locally

Symptom	Usually means	How to fix
Missing host.json or invalid schema warning	Local run path is different from deployment path	Run doctor against the exact folder being deployed
Dependency check failures	Local environment has packages installed globally/venv, but project metadata is incomplete	Pin required packages in requirements.txt and rerun doctor
Fails on one machine, passes in CI	Environment drift	Use --profile minimal, compare versions, and keep Python/CLI versions aligned

Useful verification commands:

python3 --version
az --version
func --version
pip show azure-functions-doctor
azure-functions doctor . --profile minimal -v

Before opening an issue

If you're stuck, please include the following when opening a GitHub issue:

# 1. Azure CLI version
az --version

# 2. Functions Core Tools version
func --version

# 3. Python version
python --version

# 4. Doctor version
pip show azure-functions-doctor

# 5. Doctor output on your project
azure-functions doctor . -v

# 6. Function App status (if deployed)
az functionapp show \
  --name "$FUNCTIONAPP_NAME" \
  --resource-group "$RESOURCE_GROUP" \
  --query "{state:state, runtime:siteConfig.linuxFxVersion}"

# 7. Recent logs (if deployed)
func azure functionapp logstream "$FUNCTIONAPP_NAME"

---

Clean up resources

Delete the resource group to remove all related Azure resources and stop billing.

az group delete --name "$RESOURCE_GROUP" --yes --no-wait

Confirm cleanup:

az group show --name "$RESOURCE_GROUP"

If deletion has completed, the command returns a not-found error.

Sources

• Azure Functions Python quickstart
• Azure Functions Core Tools reference
• Azure Functions hosting plans
• Flex Consumption plan
• SARIF v2.1.0 specification
• JUnit XML format reference

See Also

• Choose an Azure Functions Hosting Plan — Plan selection guide with decision tree
• README
• Examples: healthy HTTP trigger
• Examples: broken missing host.json
• azure-functions-scaffold
• azure-functions-validation
• azure-functions-openapi
• azure-functions-logging
• azure-functions-langgraph