AI Automation with n8n, Ollama, LiteLLM, RAG, and ComfyUI Beginner lab 1
A learner provisions an isolated ai automation with n8n, ollama, litellm, rag, and comfyui environment and completes a guided operational task.
Business context
Ultiblob uses this exercise to train ai automation engineer candidates on realistic private-cloud lab operations rather than static videos.
Technical objective
Configure the core ai automation with n8n, ollama, litellm, rag, and comfyui services, verify health, and record the result in the lab progress view.
Student instructions
- 1Open the lab workspace and review the topology map.
- 2Launch the required templates and wait for all provisioning checks to complete.
- 3Complete the configuration task in the course module.
- 4Run validation and capture the result for instructor review.
- 5Create a snapshot before any risky troubleshooting or failure exercise.
Troubleshooting
- If access fails, confirm the bastion session is active and the instance is not expired.
- If validation fails, inspect the lab event log before rerunning the check.
- If configuration drifts, restore the latest clean snapshot and repeat the module task.
Cleanup
- Export notes or reports required by the instructor.
- Restore or delete temporary snapshots created during the exercise.
- Use the teardown action when the module is complete or allow the TTL policy to expire the lab.
Provisioning readiness
Click Launch lab to start the provisioning flow and watch each stage complete.
- Request accepted
- Capacity reserved
- Templates queued
- Validation running
- Workspace ready
Required templates
- n8n automation node - defined
- Ollama/LiteLLM client node - defined
- Vector database/RAG node - defined
- ComfyUI workflow node - defined
Validation checks
- AI endpoint reachable: The model gateway returns at least one allowed model for the tenant.
- Docker running: Docker daemon responds and can run a hello-world container.
- VM reachable: The VM reports boot complete and responds through the tenant bastion path.
Expected result
The lab reaches Healthy state for AI endpoint reachable, Docker running, VM reachable.
Reset policy: Student can reset to the last clean snapshot; instructor can force reset from admin view. Teardown policy: Automatic teardown at TTL expiry with manual instructor override for cohorts.