DataStax Admin Console for Apache Pulsar, referred to as Pulsar Admin Console in short here after, is a web based UI that administrates topics, namespaces, sources, sinks and various aspects of Apache Pulsar features.
The admin console is a VueJS application that runs in a browser. It also includes a web server that serves up the files for the admin console as well as providing configuration and authentication services.
To install the Pulsar Admin Console:
-
Ensure node version 16.15 or higher is installed.
-
Download the latest release tarball from the release page.
-
Extract the tarball.
tar xfvz pulsar-admin-console-<version>.tar.gz
- Update the configuration in the
./pulsar-admin-console-<version>/config
directory. You can update thedefault.json
or add an override file namedlocal.json
. See "Configuration" section below.
- Change to the
server
subdirectory of the install directory.
cd ./pulsar-admin-console-<version>/server
- Start the process.
npm start
- Open the port (default 6454) in a browser.
google-chrome-stable http://localhost:6454
The configuration file for the Pulsar admin console is named default.json
and is in the ./config
directory. Additional
configuration files (for example, local.json
) can be placed in this directory to override parameters, as described here.
The default.json
configuration file contains a set of general configs for the admin console plus a server-specific set under server_config
.
The admin console server proxies all requests from the admin console to the Pulsar broker (or Pulsar proxy). You need to configure pulsar_url
to point to one of your brokers or a proxy/loadbalancer in front of the brokers (can be Pulsar proxy).
For the Brokers section of the
admin console to work, the admin console server must be able to directly reach each broker by the IP/hostname that is returned by
the Pulsar CLI command pulsar-admin brokers list <clustername>
. The cluster name can be retrieved with the CLI command pulsar-admin clusters list
.
In a geo-replication configuration, you will want to use the cluster name for the local cluster. You must set the cluster name in the
cluster_name
setting of the configuration file.
If you are using self-signed certificates, it is essential to configure the trusted root CA certificate(s) so that the
Pulsar Admin Console server can securely proxy requests to upstream servers. This can be done by setting this
environment variable, NODE_EXTRA_CA_CERTS
, to the path to the trust ca.crt file. Please see this
official Node.js documentation for additional details.
Setting | Default | Description |
---|---|---|
auth_mode | none | Authentication mode. One of none , user , k8s , or openidconnect . See "Auth Modes for details |
cluster_name | standalone | Name of Pulsar cluster connecting to. The cluster name can be retrieved with the CLI command pulsar-admin clusters list . |
tenant | public | The default Pulsar tenant to view when starting the console. |
oauth_client_id | This is the client ID that the console will use when authenticating with authentication provider | |
server_config.port | 6454 | The listen port for the console server |
server_config.pulsar_url | http://localhost:8080 | URL for connecting to the Pulsar cluster. Should point to either a broker or Pulsar proxy. The console server must be able to reach this URL. |
server_config.websocket_url | https://websocket.example.com:8500 | URL for WebSocket proxy. Used by Test Clients to connect to Pulsar. The console server must be able to reach this URL. |
server_config.function_worker_url | "" |
URL for function worker. The default value of "" means that the server will proxy all function/sink/source calls to the pulsar_url . This only needs to be set when the pulsar_url cannot handle function/sink/source requests. The console server must be able to reach this URL. |
server_config.log_level | info | Log level for the console server |
server_config.token_path | When using user or k8s auth mode, a Pulsar token is used to connect to the Pulsar cluster. This specifies the path to a file that contains the token to use. For full access, a superuser token is recommended. Alternatively, use admin_token . |
|
server_config.admin_token | When using user or k8s auth mode, a Pulsar token is used to connect to the Pulsar cluster. This specifies the token as a string. For full access, a superuser token is recommended. The token_path setting will override this value if present. |
|
server_config.token_secret | Secret used when signing access token for logging into the console. If not specified, a default secret is used | |
server_config.token_options.algorithm | HS256 |
Algorithm to use when signing the JWT (access token). Usually you want to set RS256 for asymmetric authentication (using the private key) and HS256 for the symmetric one (using the secret). |
server_config.token_options.expires_in | 12h |
TTL for the access token. |
server_config.token_options.claim | sub |
Claim field for the JWT access token. |
server_config.token_options.private_key_path | Private key path relative to the container to find the private key for signing the token. If not set, the server_config.token_secret will be used. The private key must be in PEM (.pem) or DER (.key) format. Only PKCS8 key type is supported. |
|
server_config.ssl.verify_certs | true |
Verify TLS certificate is trusted |
server_config.ssl.hostname_validation | true |
Verify hostname matches the TLS certificate |
server_config.ssl.ca_path | Path to the CA certificate. To enable HTTPS, ca_path , cert_path , and key_path must all be set |
|
server_config.ssl.cert_path | Path to the server certificate. To enable HTTPS, ca_path , cert_path , and key_path must all be set |
|
server_config.ssl.key_path | Path to the TLS key. To enable HTTPS, ca_path , cert_path , and key_path must all be set |
|
server_config.kubernetes.k8s_namespace | pulsar | When using k8s auth_mode, Kubernetes namespace that contains the username/password secrets |
server_config.kubernetes.service_host | When using k8s auth_mode, specify a custom Kubernetes host name. |
|
server_config.kubernetes.service_port | When using k8s auth_mode, specify a custom Kubernetes port. |
|
server_config.user_auth.username | When using user auth_mode, the login user name. |
|
server_config.user_auth.password | When using user auth_mode, the login password. |
|
server_config.oauth2.identity_provider_url | "" |
When using auth_mode: openidconnect set to your hostname and port. ex: https://keycloak:8443 |
server_config.oauth2.token_endpoint | "" |
When using auth_mode: openidconnect set to the path you need to forward to to get the token. ex: /token |
server_config.oauth2.grant_type | password |
When using auth_mode: openidconnect set to the grant type. Only password is support at this time. |
polling_interval | 10000 | How often the console polls Pulsar for updated values. In milliseconds. |
ca_certificate | String of CA certificate to display in the console under Credentials. | |
api_version | 2.8.3 | Version of the Pulsar client API to recommend under Samples. |
notice_text | Custom notice to appear at top of console. | |
functions_disabled | false | If functions are not enabled in the cluster, disable the function sections (Functions, Sinks, Sources) |
render_monitoring_tab | false | Enable tab that includes links to Grafana dashboards. |
grafana_url | If render_monitoring_tab is enabled, URL for Grafana. |
|
clients_disabled | false | Disable test clients. Test clients depend on WebSocket proxy, so if this is not running in Pulsar cluster you may want to disable them. |
host_overrides.pulsar | http://localhost:6650" | URL to display in console to connect to Pulsar. |
host_overrides.ws | //localhost:8080" | URL to display in console to connect to WebSocket proxy. |
host_overrides.http | http://localhost:8964" | URL to display in console to connect to Pulsar Beam HTTP proxy. |
There are four available configurations for auth_mode
: none
, user
, k8s
, and openidconnect
.
The user
and k8s
mode support token based authentication to Pulsar. For this to work, you need to configure a token either using token_path
(a file containing the token)
or admin_token
(the token itself). This token will be used to authenticate with Pulsar. For all features of the admin console to work, a superuser token is recommended.
The openidconnect
mode works with an OpenID Connect 1.0 compatible authentication provider (ex Okta) to authentication both the admin console and Pulsar.
In this mode, no login screen is presented. Authentication must be disabled on Pulsar since the admin console will not attempt to authenticate.
The admin console is protected by a login screen. The credentials are configured using the username
and password
settings in the config file. Once authenticated with these credentials, the token for connecting to Pulsar is retrieved from the server (configured using token_path
or admin_token
) and used to authenticate with the Pulsar cluster.
The admin console is protected by a login screen. The credentials are retrieved from a Kubernetes secret. For this to work, the admin console server must have access to a Kubernetes API, either user a kubeconfig file in the environment or by having local access (for example, by running as a pod in a Kubernetes cluster).
The namespace for the secret is set using k8s_namespace
. The secret name must have a prefix of dashboard-user-
followed by the username. For example, the admin user would have a secret name of dashboard-user-admin
. The password must be stored in the secret with a key of password
and a value of the password itself. Multiple secrets with the prefix can be configured to set up multiple users for the admin console. A password can be reset by patching the corresponding Kubernetes secret.
Once the user is authenticated using one of the Kubernetes secrets, the token for connecting to Pulsar is retrieved from the server (configured using token_path
or admin_token
) and used to authenticate with the Pulsar cluster.
In this auth mode, the dashboard will use your login credentials to attempt to retrieve a JWT from an authentication
provider by following the password
grant type. In the DataStax Pulsar Helm Chart, this is implemented by
integrating the Pulsar Admin Console with an identity provider, like Keycloak. Upon successful retrieval of the JWT, the Pulsar Admin Console will
use the retrieved JWT as the bearer token when making calls to Pulsar.
In addition to configuring the auth_mode
, you also need to configure the oauth_client_id
. This is the client id that
the Console will use when authenticating with Keycloak. Note that in Keycloak, it is important that this client exists
and that it has the sub
claim properly mapped to your desired Pulsar subject. Otherwise, the JWT won't work as desired.
When opening the Console, the first page is the login page. When using the openidconnect
auth mode, the auth call
needs to go to the Provider's server. In the current design, nginx must be configured to route the call to the provider.
The DataStax Pulsar Helm Chart does this automatically.
Use these versions of node and npm:
$ npm -v
6.14.15
$ node -v
v14.18.3
In order to easily manage your node and npm versions across developer machines, we use nvm. We set the node version in the .nvmrc file. In order to set up your environment, open a shell, change to the project directory, and run the following:
nvm use
nvm install
When upgrading versions of node, we should make sure to use LTS versions. You can find LTS versions on the node website or by using nvm
.
nvm list
Then, pick which lts/
version to use. Currently, we use fermium
, so you can run the following:
nvm install lts/fermium
When upgrading node versions, make sure to update the Dockerfile, the .nvmrc file, and the CI scripts (i.e. .github/).
cd dashboard
npm install
cd ..
cd server
npm install
Update the version to your desired version of Pulsar
docker run -it -p 6650:6650 -p 8080:8080 --mount source=pulsardata,target=/pulsar/data --mount source=pulsarconf,target=/pulsar/conf apachepulsar/pulsar:2.8. bin/pulsar standalone
- Configure the Admin Console dashboard. By default the settings expect a Pulsar running standalone mode reachable on localhost. Configuration is controlled by these files:
- config/default.json. This specifies the dashboard global configuration, including the Pulsar URL, cluster name and various other settings.
- config/local.json. A override file that can be used for development specific settings.
- Start server.
cd server
npm run serve
- Start the development dashboard server.
cd dashboard
npm run serve
- Connect to the Admin Console dashboard on port 9080. The dashboard server will proxy Pulsar API calls to server which then forwards to the actual Pulsar cluster (local or remote, depending on configuration)
cd dashboard
npm install
npm run serve
cd server
npm install
npm run serve
cd dashboard
npm install
npm install -g @vue/cli
npm run build-standalone