Busola is a web-based UI for managing resources within a Kubernetes cluster. It's based on the ReactJS library.
Busola project contains additional sub-projects:
Backend
- A kind of a proxy between Busola and the Kubernetes clusterTests
- Acceptance, regression and integration tests
Busola supports:
To install dependencies for the root and backend projects, and to prepare symlinks for local libraries within this repository, run the following command:
npm install
Read Install Kyma Dashboard manually to learn how to install the Dashboard with Istio Ingress and how to install it on a Kyma cluster.
Run the npm start
command.
Learn about the default configuration in Busola and how to change it.
Busola is delivered with the following default settings:
Parameter | Comment | Default Value |
---|---|---|
features |
Switches a set of Busola features on and off. Use selectors to configure conditions for the features. To switch them off, set isEnabled=false . |
isEnabled=true |
version |
Configuration version. Don’t edit this. Can be empty. | the most recent release |
Busola configuration is the product of gathering and merging the configurations from several individual sources. The following list presents the sources in the order of precedence:
Backend:
- Busola backend default cluster configuration, acquired from the defaultConfig.yaml file.
- Busola cluster configuration, available on the Busola cluster in the ConfigMap "busola/busola-config" under the key "config".
This data is mounted to the Busola
web
andbackend
Pods, and during the local development, the defaultConfig.yaml file is used.
Frontend:
- Built-in, hardcoded defaults.
- Busola frontend default cluster configuration, acquired from the defaultConfig.yaml file.
- Busola cluster configuration, available on the Busola cluster in the ConfigMap "busola/busola-config" under the key "config".
This data is mounted to the Busola
web
andbackend
Pods, and during the local development, the defaultConfig.yaml file is used. - Target cluster configuration, available on the target cluster in ConfigMap "kube-public/busola-config" under the key "config". Busola performs a request for that resource during the bootstrap process.
If you have the required authorizations and access to the kubeconfig, you can change the settings for the Busola cluster configuration and the target cluster configuration.
With the feature
toggles, you can switch each Busola feature on or off and configure them to fit your needs.
Features comprise the following elements:
FEATURE_ID
: Unique identifier, as defined in the Busola source codeselector
: The k8s resources that can activate the featureisEnabled
: Activates or deactivates the feature, overwriting the status set byselector
config
: Provides additional configuration options as needed for each feature. For details, see the README in the specific component or feature.
See the available Busola feature flags for more information.
Use the following command to run Busola locally:
npm start
After a while, open the http://localhost:8080 address in your browser, and provide your kubeconfig in the Connect cluster wizard.
Once you started Busola locally, you can begin the development. All modules have the hot-reload feature enabled, therefore, you can edit the code in real-time and see the changes in your browser.
The apps you started run at the following addresses:
Busola
- http://localhost:8080Backend
- http://localhost:3001
When developing new features in Busola, adhere to the following rules. This will help you to mitigate any security-related threats.
-
Prevent cross-site request forgery (XSRF).
- Do not store the authentication token as a cookie. Make sure the token is sent to Busola backend as a bearer token.
- Make sure that the state-changing operations (
POST
,PUT
,DELETE
, andUPDATE
requests) are only triggered upon explicit user interactions, such as form submissions. - Keep in mind that UI rendering in response to the user navigating between views is only allowed to trigger read-only operations (
GET
requests) without any data mutations.
-
Protect against cross-site scripting (XSS).
- It is recommended to use JS frameworks that have built-in XSS prevention mechanisms, such as ReactJS, Vue.js, or Angular.
- As a rule of thumb, you cannot perceive user input to be 100% safe. Get familiar with prevention mechanisms included in the framework of your choice. Make sure the user input is sanitized before it is embedded in the DOM tree.
- Get familiar with the most common XSS bypasses and potential dangers. Keep them in mind when writing or reviewing the code.
- Enable the
Content-security-policy
header for all new micro frontends to ensure in-depth XSS prevention. Do not allow forunsafe-eval
policy.
For the information on how to run tests and configure them, go to the tests
directory.
-
If you run Busola in Docker, you can mount your kubeconfig as a bind mount for Busola container. Execute the following command:
docker run --rm -it -p 3001:3001 -v <path to your kubeconfig>:/app/core-ui/kubeconfig/<your kubeconfig file name> --pid=host --name busola europe-docker.pkg.dev/kyma-project/prod/busola:latest
-
When you open Busola in your browser, go to
http://localhost:3001?kubeconfigID={YOUR_KUBECONFIG_FILE_NAME}
. Busola will try to download that file and add it for your Busola instance.
TIP: To solve most of the problems with Busola development, clear the browser cache or do a hard refresh of the website.
You are experiencing connectivity problems with Busola in Docker against a k3d cluster.
When the k3d cluster's API Server is exposed on the 0.0.0.0
address on your machine, Busola in Docker interprets 0.0.0.0
as its internal Docker address, routing the requests to the wrong endpoint.
-
For Docker Desktop for Mac and Windows, pass
DOCKER_DESKTOP_CLUSTER=true
on dockerized Busola startup. This way,0.0.0.0
is automatically replaced withhost.docker.internal
, which is resolved to 'routable' IP address of a Docker Desktop virtual machine.docker run --rm -it -p 3001:3001 -e DOCKER_DESKTOP_CLUSTER=true --pid=host --name busola europe-docker.pkg.dev/kyma-project/prod/busola:latest
-
For Linux, run Busola with
--net=host
(omitting the-p
parameter).docker run --rm -it --net=host --pid=host --name busola europe-docker.pkg.dev/kyma-project/prod/busola:latest
When you run Busola in Docker on macOS, it can't connect to the k3d cluster. The container log contains the following errors:
Error [ERR_TLS_CERT_ALTNAME_INVALID]: Hostname/IP does not match certificate's altnames: Host: host.docker.internal. is not in the cert's altnames: DNS:k3d-k3s-default-server-0, DNS:k3d-k3s-default-serverlb, DNS:kubernetes, DNS:kubernetes.default, DNS:kubernetes.default.svc, DNS:kubernetes.default.svc.cluster.local, DNS:localhost, IP Address:0.0.0.0, IP Address:10.43.0.1, IP Address:127.0.0.1, IP Address:172.28.0.3, IP Address:0:0:0:0:0:0:0:1
Busola run in a Docker container with the environment variable DOCKER_DESKTOP_CLUSTER=true
replaces the IP 0.0.0.0
in the API Server URL with host.docker.internal
. Kubernetes is not aware of that host name, so its API Server doesn't have it in the SSL certificate, which results in the above error.
Furthermore, this behavior has changed in the recent k3d versions, which is a result of this fix for this security issue.
Clusters created by k3d use a listener that extracts SNI host names from requests sent to the API server. If a new host name is requested, then the SSL certificate is regenerated, and the new host name is added to the list of Subject Alternative Names. Unfortunately, the security fix limits this mechanism only to the expected host names, like those related to Kubernetes nodes. This makes it useless for the host.docker.internal
case.
Provide the host.docker.internal
host name upfront during k3d cluster creation:
k3d cluster create kyma --k3s-arg '--tls-san=host.docker.internal@server:*'
A cluster created in such a way has the host.docker.internal
set as Subject Alternative Name in the SSL Certificate of the API Server since the very beginning.