Skip to main content

Mission Check 4 & 5 - Reconnaissance & Reporting back (RAG + Git agent)

Overview​

In this mission, you'll deploy and interact with specialized agents for knowledge retrieval and version control:

  • 📚 RAG Agent: Use Retrieval-Augmented Generation (RAG) to answer questions about Mars
  • 🔧 Git Agent: Automate git operations like commits, pushes, and repository management

The following diagram shows the architecture of the RAG system:

RAG Architecture

Step 1: Launch RAG + Git Agent Stack​

Make sure you are in the ai-platform-engineering directory:

cd $HOME/work/ai-platform-engineering

Copy over the mission4 environment variables file:

cp $HOME/.env_vars .env.mission4

Start the Docker Compose stack that includes both RAG and Git agents:

docker compose -f $HOME/work/ai-platform-engineering/workshop/docker-compose.mission4.yaml up

👀 Observe: You should see the following in the logs:

...
platform-engineer-p2p  |        SLACK AGENT CONFIG      ====
platform-engineer-p2p  | ===================================
platform-engineer-p2p  | AGENT_URL: http://localhost:8000
platform-engineer-p2p  | ===================================
platform-engineer-p2p  | ===================================8000
platform-engineer-p2p  |        KB-RAG AGENT CONFIG      ===
platform-engineer-p2p  | ===================================
platform-engineer-p2p  | AGENT_URL: http://kb-rag-agent:8000
platform-engineer-p2p  | ===================================
platform-engineer-p2p  | 2025-08-26 12:03:18,817 - INFO - Running connectivity checks for 8 agents (max 4 attempts per agent)...
platform-engineer-p2p  |        SLACK AGENT CONFIG
platform-engineer-p2p  | ===================================
platform-engineer-p2p  | AGENT_URL: http://localhost:8000
platform-engineer-p2p  | ===================================
platform-engineer-p2p  | ===================================
platform-engineer-p2p  |        KB-RAG AGENT CONFIG
platform-engineer-p2p  | ===================================
platform-engineer-p2p  | AGENT_URL: http://kb-rag-agent:8000
platform-engineer-p2p  | ===================================
platform-engineer-p2p  | 2025-08-26 12:03:18,817 - INFO - Running connectivity checks for 8 agents (max 4 attempts per agent)...
kb-rag-agent           | INFO:     172.19.0.9:53190 - "GET /.well-known/agent.json HTTP/1.1" 200 OK
agent-github-p2p       | INFO:     172.19.0.9:47004 - "GET /.well-known/agent.json HTTP/1.1" 200 OK
platform-engineer-p2p  | 2025-08-26 12:03:18,895 - INFO - HTTP Request: GET http://kb-rag-agent:8000/.well-known/agent.json "HTTP/1.1 200 OK"
platform-engineer-p2p  | 2025-08-26 12:03:18,895 - INFO - HTTP Request: GET http://kb-rag-agent:8000/.well-known/agent.json "HTTP/1.1 200 OK"
platform-engineer-p2p  | 2025-08-26 12:03:18,895 - INFO - ✅ Agent kb-rag is reachable at http://kb-rag-agent:8000
platform-engineer-p2p  | 2025-08-26 12:03:18,896 - INFO - HTTP Request: GET http://agent-github-p2p:8000/.well-known/agent.json "HTTP/1.1 200 OK"
platform-engineer-p2p  | 2025-08-26 12:03:18,897 - INFO - ✅ Agent github is reachable at http://agent-github-p2p:8000
...

The docker-compose stack should start the following services:

  • platform-engineer-p2p : The platform engineer supervisor agent
  • agent-github-p2p : The git agent
  • kb-rag-agent : The RAG agent
  • kb-rag-server : The RAG server
  • kb-rag-web: The RAG web UI
  • milvus-standalone: The Milvus vector database
  • etcd: Needed for Milvus
  • minio: Needed for Milvus

🫸 Wait: Wait until this process is completed.

Once you see the logs above, you can verify the supervisor agent is healthy by checking the health endpoint.

💡 Tip: You can open a new terminal by clicking the + button in your terminal window. This allows you to run commands in parallel without stopping your running services.

curl http://localhost:8000/.well-known/agent.json | jq

👀 Observe: The response should be a JSON object (the A2A agent card). If you get an error, wait for a few seconds and try again. The agent might still be starting up.

Step 2: Populate RAG database​

Now, we will populate the RAG with documentation. The docker-compose stack should have started the kb-rag-web service, which is the web ui for the RAG server.

🚀 Open RAG Web UI (Port 6100)

Once the RAG Web UI is open:

  1. Copy the URL for Mars wiki page https://en.wikipedia.org/wiki/Mars and paste it in Ingest URL field
  2. Click Ingest button
  3. 🫸 Wait: Wait for the ingestion to complete

👀 Observe:

  • The status in the RAG Web UI should show as ✅ Successfully processed 1 URL(s).
  • Switch over to the terminal showing your docker compose logs and observe the kb_rag ingestion logs for details about the document processing.

RAG UI

Explanation:​

Here's what happens:

  • The RAG server crawls webpage (it also supports sitemap), and fetches the page.
  • The HTML is parsed and content is extracted.
  • If the page is too large, it is split into chunks, using Recursive Text Splitter.
  • Each chunk is sent to embedding model (LLM in this case) to generate embeddings.
  • The embeddings are stored in a vector store (Milvus), along with metadata (source, title, description etc).

RAG ingestion

Step 3: Verify RAG System is working​

We will now use the UI to query the RAG system and verify it is working.

Open RAG Web UI again (Port 6100)

Type: Axial tilt in the query box, then Click: Search button

👀 Observe: The response should return relevant document chunks. The chunks may not be formatted in a way that is easy to read. As long as some document chunks are returned, the RAG system is working.

Step 4: Using the RAG Agent​

💡 Tip: You can open a new terminal by clicking the + button in your terminal window. This allows you to run commands in parallel without stopping your running services.

docker run -it --network=host ghcr.io/cnoe-io/agent-chat-cli:stable

💡 Tip: When asked to 💬 Enter token (optional): , just press enter ⏎.

In production, your system will use a JWT or Bearer token for authentication here.

chatcli token

This should open a CLI chat client. You can now interact with the supervisor agent.

Ask the agent:

What is the axial tilt of Mars?

👀 Observe: The agent should respond with the axial tilt of Mars, and cite the source of the information.

Explanation:​

Here's what happens:

  • The RAG agent embeds the question using the same embedding model used to embed the documents.
  • The agent then uses the vector store to find the most similar documents.
  • The agent uses the retrieved document chunks to answer the question.

RAG Agent Architecture

Step 5: Multi-agent interaction​

Now, we will test the multi-agent interaction by asking the supervisor agent to:

  • search for information about Mars (reconnaissance)
  • commit the steps to git repository (report back).

In the CLI chat client, ask the agent:

Research and write a report on mars surface in markdown format, then commit/upload it as a markdown file named '%%LABNAME%%-report.md' with commit message "mars-surface-report" to repo https://github.com/outshiftcaipe/mission-mars-colony on the `main` branch.

💡 Tip: Switch to your logs terminal to watch the agents collaborate.

👀 Observe: While the agent is working, in the first terminal tab, you should see the ai platform engineering supervisor agent delegate tasks to kb-rag and github agents. Please note that this step might take a bit of time as the agents go back and forth to figure out the task.

👀 Observe: Back in the CLI chat client, the agent should have:

  • Created a report with name: %%LABNAME%%-report.md
  • Committed it to the workshop git repository with commit message "mars-surface-report".

Explanation:​

Here's what happens:

  • The supervisor agent determines what needs to be done, and delegates the tasks to the sub agents.
  • It first asks the RAG agent to search for information about Mars surface.
  • The RAG agent uses the vector store to find relavant information, and write a report.
  • The supervisor agent then asks the git agent to commit the report as a text file to the repository.

Final step: Tear down the mult-agent system​

Run the following in one of the terminals:

💡 Tip: You can open a new terminal by clicking the + button in your terminal window. This allows you to run commands in parallel without stopping your running services.

🛑 Before You Proceed: Bring Down Your Docker Containers

  • Important: Run docker compose down in your terminal to stop and remove all running containers for this demo before moving on to the next steps.
  • This ensures a clean environment and prevents port conflicts or resource issues.
docker compose -f $HOME/work/ai-platform-engineering/workshop/docker-compose.mission4.yaml down

Then close all the tabs that were opened for this mission.

Mission Checks​

🚀 Reconnaissance & Reporting Mission Checklist​

  • Launch RAG + Git agent stack successfully
  • Import Mars wiki into RAG
  • Verify RAG queries return relevant documentation
  • Connect chat client to supervisor agent
  • Test knowledge retrieval: "What is the axial tilt of Mars?"
  • Test multi-agent interaction: Documentation search + git operations

📖 Further Reading​

RAG Agent​

More information on the RAG agent can be found here.

Graph RAG​

The AI platform engineering repository also includes Graph RAG, which uses Neo4J to create knowledge graphs from structured data (K8s objects, AWS resources). See the Graph RAG documentation for more details, and how to run it in your local environment.

Unified Knowledge Base​

Currently there is a open discussion on how to unify different RAG systems, and create a unified knowledge base thats specialized for AI platform engineering. Please follow the discussion here.

Github agent​

The Github agent can be used as is. Please refer to the Github agent documentation for more details.

🛠 Troubleshooting​

Port already in use​

If you get a port already in use error from docker, you may still be running containers from previous missions. You can run docker ps to see what containers are running, and docker stop <container_id> to stop them.

🛑 Do not stop essential containers

  • Important: DO NOT STOP containers with name etilab and localdev-control-plane. These are needed to run the workshop.

Terminal refreshed and can't see logs​

If your terminal refreshed and you can no longer see the docker compose logs:

docker compose -f $HOME/work/ai-platform-engineering/workshop/docker-compose.mission4.yaml logs -f --tail 50

If you want to see individual container logs, you can use:

docker logs agent-github-p2p
docker logs kb-rag-agent
docker logs kb-rag-server
docker logs kb-rag-web
docker logs platform-engineer-p2p

Could not find message from source agent​

If you see an error similar to this:

Could not find message from source agent github_tools_agent. Found names: {'platform_engineer_supervisor'}

There might be an issue with the communication between the agents. Try restarting the docker-compose stack by running:

docker compose -f $HOME/work/ai-platform-engineering/workshop/docker-compose.mission4.yaml down
docker compose -f $HOME/work/ai-platform-engineering/workshop/docker-compose.mission4.yaml up

General errors​

If you have issues with any agents not responding properly, try the prompt again after a few seconds, or restart the docker-compose stack as shown above.