Thinkube Kubernetes
Install and remove optional components
Add a database, a vector store or an inference backend to the platform when you need it, and take it out again
- Level
- beginner
- Time
- 30 min
- Risk
- low
- Updated
- 2026-10-04
Overview
Basic idea
The platform installs its core services once. Everything else is an optional component you add when a project needs it: vector stores, databases, observability, inference backends.
-
Installed on request. An install sets up the component’s pods, storage and address.
-
Requirements first. Each component names what it needs. The list says whether those are there. An install runs only when they are.
-
One at a time. Installs run one at a time; the agent follows each to the end.
-
Removed with its data kept. An uninstall removes the pods, the addresses and the configuration. The data volume stays.
What you’ll accomplish
You list the optional components, install Weaviate, check that it answers, and uninstall it.
What to know before starting
Required
-
Asking your agent in Thinkube IDE to do things on the platform.
Optional
-
Components catalog: every component and what it needs.
-
Check and restart services from your agent: turning an installed component off and on.
Instructions
Step 1. List the components
Ask your agent:
› which optional components are installed, and which could I add?
Or by hand: in Thinkube Control open Optional Components.
Expected output, from the reference run, shortened:
installed: Prometheus, Perses, OpenSearch, PgAdmin, Qdrant, Knative, Chroma, Valkey,
ClickHouse, NATS, Langfuse, Ollama, Langflow, vLLM, TensorRT-LLM, Text Embeddings
not installed: Weaviate, Argilla, CVAT
Weaviate: requirements_met: true
Step 2. Install one
Ask your agent:
› install Weaviate
Or by hand: on Optional Components, choose Install on the component.
Expected output, from the reference run:
deployment_id: ca152774-dbb7-44df-83db-f2442d119537 status: queued message: Installation of Weaviate queued at position 1; poll get_deployment_status with the deployment id
The status names the step in progress, such as Create PersistentVolumeClaim for Weaviate data or Wait for Weaviate StatefulSet to be ready, and at the end:
status: success output: Component weaviate installed successfully started_at: 00:17:07 completed_at: 00:20:15
Step 3. Check that it answers
Ask your agent:
› is Weaviate running?
Expected output, from the reference run:
status: running message: All 1 pods running
On the reference run https://weaviate.<your domain>/v1/.well-known/ready answered 200.
Step 4. Remove it
Ask your agent:
› uninstall Weaviate
Or by hand: on Optional Components, choose Uninstall on the component.
Expected output, from the reference run:
status: success output: Component weaviate uninstalled successfully started_at: 00:21:59 completed_at: 00:23:08
and then Weaviate’s status is:
installed: false status: not_installed message: No pods found
The namespace weaviate and its data volume weaviate-data-pvc, 10 GiB, stayed.
Step 5. Next steps
To undo: Uninstall the component.
-
Check and restart services from your agent: health, restart, and turning a component off.
-
Deploy a serverless service: an app that uses the Knative component.
-
Components catalog: what each component provides.