🎨 Build Inspiration & Tutorials¶

Ready to build? Here are 8 interactive tutorials to get you started with ANSAI.

🚀 NEW: Interactive, executable tutorials now available →

Infrastructure Automation¶

Auto-scale based on error rates

Goal: Automatically scale up resources when error rates spike, then scale down when stable.

Blocks Used: - 📊 Log Analysis: Monitor Nginx/Apache logs for 500 errors. - 🤖 Orchestration: Trigger scaling action based on threshold. - 💬 Notification: Alert team of scaling event.

How to Build: 1. Create Log Monitor: Use ansai-log-analyzer to tail logs and count 5xx errors/minute. 2. Set Threshold: If errors > 50/min, trigger scale_up.yml playbook. 3. Scale Up: Playbook spins up new container/instance or increases resource limits. 4. Scale Down: If errors < 10/min for 15 mins, trigger scale_down.yml.

Code Snippet:

- name: Check Error Rate
  shell: "tail -n 1000 /var/log/nginx/access.log | grep ' 500 ' | wc -l"
  register: error_count

- name: Scale Up
  command: docker compose up -d --scale web=5
  when: error_count.stdout | int > 50

ChatOps for infrastructure management

Goal: Control your infrastructure directly from Slack/Discord.

Blocks Used: - 💬 Natural Language: Interpret "Restart web server" commands. - 🤖 Orchestration: Execute Ansible playbooks securely. - 🛡️ Security: Validate user permissions.

How to Build: 1. Setup Webhook: Create a simple Flask/FastAPI endpoint to receive Slack slash commands. 2. Parse Intent: Use LiteLLM to parse "Fix the database" into specific intent db_restart. 3. Execute: Trigger ansible-playbook playbooks/db-restart.yml. 4. Reply: Send "Database restarting..." back to Slack.

Code Snippet:

@app.route('/slack/command', methods=['POST'])
def handle_command():
    command = request.form.get('text')
    # AI parses intent
    action = ai.parse_intent(command) 
    if action == 'restart_web':
        subprocess.run(['ansible-playbook', 'restart-web.yml'])
        return "Restarting web server..."

Multi-cloud orchestration with fallback

Goal: Deploy to AWS; if it fails, automatically deploy to DigitalOcean.

Blocks Used: - 🎯 Orchestration: Workflow logic with error handling. - ☁️ Cloud Modules: AWS and DigitalOcean Ansible collections. - 📧 Notification: Report fallback status.

How to Build: 1. Try Primary: Run deploy-aws.yml inside a block. 2. Catch Failure: Use rescue to catch deployment errors. 3. Execute Fallback: Run deploy-do.yml in the rescue block. 4. Verify: Check health of fallback endpoint.

Code Snippet:

- block:
    - name: Deploy to AWS
      include_tasks: aws-deploy.yml
  rescue:
    - name: Fallback to DigitalOcean
      include_tasks: do-deploy.yml
      vars:
        region: nyc1
    - name: Notify Team
      slack:
        msg: "AWS Deployment failed. Failed over to DigitalOcean."

Cost optimization with intelligent scheduling

Goal: Shutdown dev environments at 7 PM and start them at 8 AM to save 50% costs.

Blocks Used: - ⏰ Scheduling: Cron or Systemd timers. - ☁️ Cloud Modules: AWS/GCP instance state management. - 🤖 AI Prediction: (Optional) Don't shutdown if devs are still active.

How to Build: 1. Define Schedule: Create variables for start_time and stop_time. 2. Check Activity: (Optional) Check SSH logs for active sessions. 3. Stop Instances: Run playbook to stop tagged env:dev instances. 4. Start Instances: Run playbook to start them in the morning.

Code Snippet:

- name: Stop Dev Instances
  ec2_instance:
    state: stopped
    filters:
      "tag:Environment": dev
  when: 
    - ansible_date_time.hour == "19"
    - active_ssh_sessions == 0

Security and Compliance¶

Compliance-as-code with auto-remediation

Goal: Automatically fix security misconfigurations (e.g., open SSH ports, root login).

Blocks Used: - 🔍 Audit: Scan config files for violations. - 🤖 Orchestration: Apply hardening roles. - 📊 Reporting: Generate compliance report.

How to Build: 1. Scan: Check /etc/ssh/sshd_config for PermitRootLogin yes. 2. Remediate: Use lineinfile to set it to no. 3. Reload: Restart SSH service. 4. Report: Log the remediation action.

Code Snippet:

- name: Secure SSH
  lineinfile:
    path: /etc/ssh/sshd_config
    regexp: '^PermitRootLogin'
    line: 'PermitRootLogin no'
  notify: Restart SSH

Automated disaster recovery testing

Goal: Randomly kill a service in staging to test if self-healing works (Chaos Engineering).

Blocks Used: - 🎲 Chaos: Randomly select and stop services. - 🛡️ Self-Healing: Verify the system recovers. - 📊 Reporting: Measure Time To Recovery (TTR).

How to Build: 1. Select Target: Randomly pick web, db, or cache. 2. Disrupt: systemctl stop <target>. 3. Wait: Wait 2 minutes. 4. Verify: Check if service is back up (self-healing should have kicked in). 5. Alert: If not back up, fail the test.

Code Snippet:

- name: Chaos Monkey - Stop Service
  service:
    name: "{{ item }}"
    state: stopped
  with_random_choice:
    - nginx
    - postgresql

Advanced AI Ops¶

Self-optimizing database tuning

Goal: Adjust Postgres/MySQL config based on actual load patterns.

Blocks Used: - 📊 Metrics: Prometheus/CloudWatch data. - 🤖 AI Analysis: Analyze performance vs config. - ⚙️ Config Mgmt: Update postgresql.conf.

How to Build: 1. Gather Data: Collect cache hit ratio, active connections, memory usage. 2. AI Analyze: Ask LLM "Given these metrics and 32GB RAM, suggest work_mem setting." 3. Tune: Apply new config if different from current. 4. Restart: Gracefully reload DB.

Code Snippet:

- name: AI Tune Config
  shell: "ansai-tune-db --metrics current_metrics.json"
  register: new_config

- name: Apply Config
  template:
    src: postgresql.conf.j2
    dest: /etc/postgresql/13/main/postgresql.conf

Predictive maintenance with ML

Goal: Detect disk filling up trends and expand storage BEFORE it hits 100%.

Blocks Used: - 📈 Trend Analysis: Linear regression on disk usage. - ☁️ Cloud Modules: Expand EBS volume. - 📂 Filesystem: Resize file system online.

How to Build: 1. Analyze: Check disk usage over last 7 days. 2. Predict: "Will usage hit 95% in next 24 hours?" 3. Expand: If yes, modify AWS volume size +20%. 4. Resize: Run resize2fs.

Code Snippet:

- name: Predict Disk Full
  script: predict_disk_growth.py
  register: prediction

- name: Expand Volume
  ec2_vol:
    id: "{{ volume_id }}"
    size: "{{ current_size * 1.2 }}"
  when: prediction.stdout == "critical"

🧱 Ready to build?¶

Download the Core Building Blocks