Camunda RPA Worker

The Camunda RPA Runtime is designed to bridge the gap between Robotic Process Automation (RPA) bots and the Camunda workflow engine, enabling you to seamlessly integrate your task workers with Camunda.

Getting Started

Obtaining the RPA Worker

Downloads and details for container images are available on the Releases page.

Configuration

The RPA Worker can source its configuration either from a configuration file, or environment variables. The RPA Worker will look for a file named rpa-worker.properties in the current directory.

Zeebe

Connecting to a Camunda cluster requires providing the configuration for the Zeebe client. The RPA Worker uses the Spring Zeebe client, and so uses the same Configuration Properties.

Example Camunda local dev configuration

camunda.client.mode=selfmanaged
camunda.client.zeebe.grpc-address=http://zeebe.camunda.local
camunda.client.zeebe.rest-address=http://zeebe.camunda.local

Example Camunda SaaS configuration

camunda.client.mode=saas
camunda.client.auth.client-id=<client id>
camunda.client.auth.client-secret=<client secret>
camunda.client.cloud.cluster-id=<cluster id>
camunda.client.cloud.region=<region>

camunda.rpa.secrets.backend=camunda

Example custom/dev SaaS configuration

camunda.client.mode=selfmanaged

camunda.client.cloud.cluster-id=<cluster id>
camunda.client.cloud.region=<region>

camunda.client.auth.issuer=https://login.cloud.dev.ultrawombat.com/oauth/token
camunda.client.auth.client-id=<client id>
camunda.client.auth.client-secret=<client secret>

camunda.client.zeebe.rest-address=https://${camunda.client.cloud.cluster-id}.${camunda.client.cloud.region}.zeebe.dev.ultrawombat.com
camunda.client.zeebe.grpc-address=https://${camunda.client.cloud.cluster-id}.${camunda.client.cloud.region}.zeebe.dev.ultrawombat.com
camunda.client.zeebe.base-url=https://${camunda.client.cloud.region}.zeebe.dev.ultrawombat.com/${camunda.client.cloud.cluster-id}
camunda.client.zeebe.audience=zeebe.dev.ultrawombat.com

camunda.rpa.zeebe.auth-endpoint=https://login.cloud.dev.ultrawombat.com/oauth

camunda.rpa.secrets.backend=camunda
camunda.rpa.secrets.camunda.token-audience=secrets.dev.ultrawombat.com
camunda.rpa.secrets.camunda.secrets-endpoint=https://cluster-api.cloud.dev.ultrawombat.com

Tip	When creating an API client in the Camunda Cloud Console, the autogenerated configuration file provided contains most of the necessary configuration for connecting to production Camunda SaaS stacks.

Local-only Working

The Zeebe startup check and subscriptions can be disabled by setting camunda.client.zeebe.enabled to false. Although the Zeebe Client itself cannot be disabled and it will still log, the application will keep running allowing local working and these messages can be ignored.

Python

When the RPA Worker sets up its Python environment it will install a default set of dependencies.

Additional dependencies can be installed by providing a supplementary requirements.txt file with the camunda.rpa.python.extra-requirements property.

extra-requirements.txt

numpy==1.26.4
jmespath

rpa-worker.properties

camunda.rpa.python.extra-requirements=extra-requirements.txt

If the extra requirements files changes, the RPA Worker will install the new dependencies at the next restart, although no dependencies will be removed. If the Python environment is deleted then the RPA Worker will recreate it from scratch.

Robot Job Execution

Robot Runtimes

The RPA Worker supports two runtimes for executing Robotframework.

The Python Runtime

The Python runtime creates a dedicated Virtual Environment using an available Python distribution. The Python distribution may already be present on the host, or it may be provisioned automatically by the RPA Worker on platforms where this is supported.

The first time the environment is created the RPA Worker will install the necessary base dependencies into the environment. It will also install, if configured, any additional user-supplied extra requirements. If the base requirements change, such as a new version of the RPA Worker updating these dependencies, or the extra requirements are updated, then the RPA Worker will update the environment with the new dependencies automatically.

As dependencies are fetched from the internet, the Python Runtime requires an internet connection.

The Static Runtime

The Static Runtime is a pre-built Python environment which is bundled with the RPA Worker. It contains a Python interpreter, as well as all the base dependencies, statically linked together in a single executable. As such, when using the Static Runtime it is not possible to install any extra requirements.

As the base dependencies are included in the Static Runtime, it does not require an internet connection.

Choosing a Runtime

The choice of runtime can be configured via the camunda.rpa.python-runtime.type configuration property. It supports three values: auto (the default), static, and python.

When the runtime is set to auto, the RPA Worker will run the following checks to see if the Python Runtime is supported:

• If there is an existing Python environment created by the RPA Worker then this will be used

• If an internet connection is detected, AND:

  • There is a local Python distribution at a supported version (>=3.10,<3.12) then this will be used to create a new Virtual Environment, OR

  • There is no local Python, but it can be provisioned automatically on the current platform (Windows), then Python will be provisioned in this way and used to create a new Virtual Environment

If these conditions are not satisfied then the RPA Worker will fall back to the Static Runtime.

Note	The Static Runtime is not supported in the JAR distribution of the RPA Worker

Built-in Variables

The RPA Worker makes the following environment variables available to the Robot execution:

Variable	Description
RPA_WORKSPACE_ID	The workspace ID of the current job
RPA_WORKSPACE	The full path to the workspace of the current job
RPA_SCRIPT	The name of the currently executing script
RPA_EXECUTION_KEY	The execution key of the currently executing script
RPA_ZEEBE_JOB_KEY	The Zeebe Job Key which triggered the Robot execution
RPA_ZEEBE_JOB_TYPE	The Zeebe Job type
RPA_ZEEBE_BPMN_PROCESS_ID	The current BPMN process ID
RPA_ZEEBE_PROCESS_INSTANCE_KEY	The current process instance key

Configuration Reference

Property	Description	Default Value
Python Properties
camunda.rpa.python.path	The directory in which the Python environment will be created	${user.dir}/python/venv/
camunda.rpa.python.download-url	The URL from which the portable Python will be downloaded if no system Python is available (Windows only)	https://github.com/winpython/winpython/releases/download/11.2.20241228final/Winpython64-3.13.1.0dot.zip
camunda.rpa.python.download-hash	The expected SHA-256 hash of the file at download-url	47b9a4ce75efb29d78dda80716d6c35f9a13621efd3a89ef8242a114ef8001a3
camunda.rpa.python.extra-requirements	The path to additional Python requirements that should be installed when setting up the Python environment	None
camunda.rpa.python.interpreter	Path to a specific Python interpreter to use, rather than searching the path	None
camunda.rpa.python-runtime.type	Which runtime to use for executing Robot. Can be auto, python, or static	auto
Robot Properties
camunda.rpa.robot.default-timeout	The default timeout to use when no other timeout is specified (ISO 8601 duration string)	PT5M (5 minutes)
camunda.rpa.robot.fail-fast	Stops task execution if any task fails	true
Sandbox Properties
camunda.rpa.sandbox.enabled	Whether the script evaluation API should be available	true
Zeebe Properties
camunda.client.zeebe.enabled	Whether the Zeebe integration should be enabled	true
camunda.rpa.zeebe.rpa-task-prefix	The prefix used to construct the task subscription key	camunda::RPA-Task::
camunda.rpa.zeebe.worker-tags	The worker tag(s) this worker should subscribe to (comma-separated)	default
camunda.rpa.zeebe.max-concurrent-jobs	How many Robot jobs from Zeebe to run concurrently	1
camunda.rpa.zeebe.auth-endpoint	The authentication endpoint that should be used for authenticating before sending requests	https://login.cloud.camunda.io/oauth
camunda.client.zeebe.audience	The token audience to use when authenticating for Zeebe	zeebe.camunda.io
camunda.client.cloud.region	The region for Camunda SaaS	None
camunda.client.cloud.cluster-id	The cluster ID for Camunda SaaS	None
camunda.client.zeebe.base-url	The region for Camunda SaaS	https://${camunda.client.cloud.region}.zeebe.camunda.io/${camunda.client.cloud.cluster-id}
camunda.client.auth.client-id	The client ID to use when authenticating with Zeebe	None
camunda.client.auth.client-secret	The client secret to use when authenticating with Zeebe	None
See the Spring Zeebe Configuration Reference for other supported properties
Secrets Properties - General
camunda.rpa.secrets.backend	The secrets provider backend	none (Secrets disabled)
Secrets Properties - Camunda (…​backend=camunda)
camunda.rpa.secrets.camunda.secrets-endpoint	The endpoint providing the Camunda secrets service	https://cluster-api.cloud.camunda.io
camunda.rpa.secrets.camunda.secrets.token-audience	The token audience to use when authenticating for secrets	secrets.camunda.io
General Properties
server.address	Network address to which the server should bind	localhost for standalone, any for container
logging.file.name	Output file for log file (non-containerised)	rpa-worker.jsonl.log

Supported Configuration

The RPA Worker will source its configuration from configuration files and/or environment variables. The precedence order is documented in Spring’s Externalized Configuration documentation.

Configuration Files

A configuration file can be a Properties or a YAML document. It can have the name application or rpa-worker. It can be placed in the working directory of the Worker, or in a directory named config therein. The following are all valid paths to RPA Worker configuration files:

• rpa-worker.properties

• application.properties

• rpa-worker.yaml

• application.yaml

• config/rpa-worker.properties

• config/application.properties

• etc

Environment Variables

Preferable for container-based deployments, the RPA Worker can source its configuration from environment variables. The name of the environment variables follow Spring’s Binding Rules, which can be briefly summarised as:

1. Convert the property name to upper-case

2. Replace dots with underscores

3. Remove hyphens

For example camunda.rpa.zeebe.worker-tags would become CAMUNDA_RPA_ZEEBE_WORKERTAGS.

Launching a dev container with Zeebe connection details, mapping the Worker’s listener port to the host, might look like:

$ docker run --rm -it \
    -p 36227:36227 \
    -e CAMUNDA_CLIENT_MODE=saas \
    -e CAMUNDA_CLIENT_AUTH_CLIENTID=abc123 \
    -e CAMUNDA_CLIENT_AUTH_CLIENTSECRET=def234 \
    -e CAMUNDA_CLICNE_CLUSTERID=ghi345 \
    -e CAMUNDA_CLIENT_REGION=lpp-1 \
    ghcr.io/camunda/rpa-worker:<version>

License

These source files are made available under the Camunda License Version 1.0.