---
title: "Kubernetes Stacks"
description: "A Stack defines a set of platform configuration options that will get
applied to each Node-RED instance when created."
canonical_url: "https://flowfuse.com/docs/install/kubernetes/stacks/"
---

> Index of every page on this site: [llms.txt](https://flowfuse.com/llms.txt)

# Kubernetes Stacks

A Stack defines a set of platform configuration options that will get
applied to each Node-RED instance when created.

For container based deployment models, this covers three things:

- `memory` - the amount of memory (in MB) to limit container to. Recommended minimum: `256`.
- `cpu` - a value between 1 and 100 that is the % of a CPU core the container should be allowed to consume.
- `container location` - this is the fully qualified name of the container to use. The default container built when following the install instructions is named `flowfuse/node-red:latest`

As of FlowFuse v3.2.0 both `memory` and `cpu` allow you to set both min and max values. This allows for over commiting of Instance pods to nodes. If you are unsure, set min and max to the same values
to ensure they always get a dedicated allocation.s

If you wish to use different Node-RED version, you need to specify the name of the container and the version you want to use. For example, if you want to use Node-RED v3.1.x, you should enter `flowfuse/node-red:latest-3.1.x` in the `container location` section of the Stack configuration.
Full list of available pre-built containers can be found on [Docker Hub](https://hub.docker.com/r/flowfuse/node-red/tags).

## Creating Own Containers

As mentioned in the previous paragraph, we encourage to use our pre-built containers in your stacks.
However, if you want to create your own container, you can do so by creating a `Dockerfile` and `package.json` files.
There is an example `Dockerfile` and `package.json` in the [node-red-container](https://github.com/FlowFuse/helm/tree/main/node-red-container)
directory of the [helm](https://github.com/FlowFuse/helm) project. This will start with `nodered/node-red:latest`
as it's base and then add the required FlowFuse components.

Builds of this container for amd64, arm64 and armv7 are built for every release and published to Docker hub as [flowfuse/node-red](https://hub.docker.com/r/flowfuse/node-red). These can be used as a base to build custom stacks.

If you wanted to pin at Node-RED v3.0.2 you would change the first line to:

```docker
FROM nodered/node-red:3.0.2

ARG REGISTRY
RUN if [[ ! -z "$REGISTRY" ]] ; then npm config set @flowfuse:registry "$REGISTRY"; fi

COPY package.json /data
...
```

To add nodes to the default image you can extend the supplied container.
The following Dockerfile will install the node-red-dashboard

```docker
FROM flowfuse/node-red

WORKDIR /usr/src/node-red
RUN npm install node-red-dashboard

WORKDIR /usr/src/flowforge-nr-launcher
```

To build the container run the following:

```shell
docker build node-red-container/Dockerfile-dashboard -t [your.container.registry]/flowfuse/node-red-dashboard:3.0.2
docker push [your.container.registry]/flowfuse/node-red-dashboard:3.0.2
```

You would then enter `[your.container.registry]/flowforge/node-red-dashboard:3.0.2` in the `container` section
of the Stack configuration.

Stacks can be changed on a per Node-RED instance basis, see also the
[user stack documentation](https://flowfuse.com/docs/user/changestack).

## Sitemap

See the full [sitemap](https://flowfuse.com/sitemap.md) for all pages.
