snoopybot

Mastodon bot that posts lines from Snoopy's novel and other writings from the Peanuts comic strip. Runs as an Azure Container Apps Job on a daily schedule.

Prerequisites

• Go 1.22+
• Azure Developer CLI (azd)
• Docker (for building the container image)
• An existing Azure Container Registry
• A Mastodon account and access token

Local development

Fill in your credentials in .env:

Variable	Description
MASTODON_SERVER	Your Mastodon instance URL (e.g. https://mastodon.social)
MASTODON_ACCESS_TOKEN	Access token with write:statuses scope
AZURE_STORAGE_ACCOUNT	Azure Storage account name
AZURE_STORAGE_ACCESS_KEY	Azure Storage account key
DRY_RUN	Set to true to print the post without sending it

To get a Mastodon access token: go to your instance → Preferences → Development → New Application → enable write:statuses → copy the access token.

Run locally (dry run by default):

go run .

To fire a real post, set DRY_RUN=false in .env before running.

Project structure

main.go                            # Entry point: validates env vars, calls bot.DoWork()
internal/
  bot/
    data.go                        # Novel lines and miscellaneous quotes
    bot.go                         # DoWork(): coin flip, post novel or misc quote
  mastodon/
    client.go                      # Mastodon API client (POST /api/v1/statuses)
  storage/
    state.go                       # Azure Tables: tracks current novel line index
Dockerfile                         # Multi-stage build → scratch-based image (~8 MB)
azure.yaml                         # Azure Developer CLI template definition
infra/
  main.bicep                       # Subscription-scope: resource group + module
  resources.bicep                  # All Azure resources (storage, container apps env, job)
  main.parameters.bicepparam       # Reads values from azd environment
  scripts/
    postprovision.sh               # Hook: updates .env and prints GitHub secret commands
    postprovision.ps1              # Same, for Windows/PowerShell
.github/
  workflows/
    deploy.yml                     # CI/CD: build image → push to ACR → update job

Provision to Azure

Infrastructure is managed with the Azure Developer CLI. A single azd provision creates the resource group, storage account, Container Apps Environment, and Container Apps Job.

1. Set required environment variables

azd handles AZURE_ENV_NAME and AZURE_LOCATION automatically — it will prompt for the location on first run and remember it. Set the remaining values before provisioning:

# Existing Azure Container Registry
azd env set AZURE_CONTAINER_REGISTRY_LOGIN_SERVER  <registry>.azurecr.io
azd env set AZURE_CONTAINER_REGISTRY_USERNAME      <acr-admin-username>
azd env set AZURE_CONTAINER_REGISTRY_PASSWORD      <acr-admin-password>

# Mastodon credentials
azd env set MASTODON_SERVER        https://mastodon.social
azd env set MASTODON_ACCESS_TOKEN  <value>

ACR admin credentials are found in the Azure Portal under your registry → Access keys. Enable the admin user if it is not already on.

2. Provision infrastructure

azd provision

This creates the following resources inside a new rg-<env> resource group:

Resource	Purpose
Storage Account	Hosts the state table tracking the novel line index
Log Analytics Workspace	Container Apps structured logging
Container Apps Environment	Shared environment for the job
Container Apps Job	Runs daily at 5:00 PM UTC; 0.25 vCPU / 0.5 GiB

After provisioning, the postprovision hook automatically updates AZURE_STORAGE_ACCOUNT and AZURE_STORAGE_ACCESS_KEY in your local .env file. It also prints the gh secret set commands needed for CI/CD.

3. Configure GitHub Actions secrets

The deployment workflow (.github/workflows/deploy.yml) requires these repository secrets:

Secret	Where to get it
ACR_LOGIN_SERVER	Your ACR login server (e.g. myregistry.azurecr.io)
ACR_USERNAME	ACR admin username
ACR_PASSWORD	ACR admin password
AZURE_CREDENTIALS	Service principal JSON — see below
AZURE_RG	Printed by postprovision hook
CONTAINERAPPS_JOB_NAME	Printed by postprovision hook

Create the AZURE_CREDENTIALS service principal:

az ad sp create-for-rbac \
  --name snoopybot-deploy \
  --role Contributor \
  --scopes /subscriptions/<subscription-id>/resourceGroups/rg-<env> \
  --sdk-auth

Copy the JSON output as the AZURE_CREDENTIALS secret.

4. Deploy the container image

Push to master to trigger the GitHub Actions workflow, which builds the Docker image, pushes it to ACR, and updates the Container Apps Job. To manually trigger a job execution after deploying:

az containerapp job start --name <job-name> --resource-group rg-<env>

Architecture

GitHub push → Actions workflow → docker build → ACR → Container Apps Job
                                                           │
                                            (daily 5 PM UTC cron)
                                                           │
                                                    Go binary runs
                                                    ┌──────┴──────┐
                                               50% novel     50% misc
                                                    │
                                            Azure Table Storage
                                            (tracks novel index)
                                                    │
                                             Mastodon API v1
                                              (posts status)