This repository contains links to Camunda Platform 8 resources, the official release artifacts (binaries), and supporting config files for running Docker Compose as a local development option.
For production and development environments, we recommend:
- Camunda SaaS for managed hosting
- Helm/Kubernetes for self-managed installations.
For more information about Self-Managed, including additional development installation options, see our documentation.
For production setups we recommend using Helm charts, which can be found at helm.camunda.io.
- Camunda 8 Run: A lightweight, self-contained distribution of Camunda Platform 8 - designed to simplify deployment by bundling all required components into a single package. For more information, see Camunda 8 Run Documentation
- Docker Compose: A local development environment. For more information, see the Docker Compose section below.
- Local Kubernetes Cluster: Deploy Camunda 8 Self-Managed on a local Kubernetes cluster for development purposes. For more information, see Camunda 8 Local Kubernetes documentation
- Documentation
- Camunda Platform SaaS
- Getting Started Guide
- Releases
- Helm Charts
- Zeebe Workflow Engine
- Contact
• Docker Compose: Version 1.27.0+ (supports the [latest compose specification](https://docs.docker.com/compose/compose-file/) ).
• Docker: Version 20.10.16+.
• Add keycloak to resolve to 127.0.0.1 on your local machine, and set KEYCLOAK_HOST to `keycloak` in the `.env` file for token refresh and logout functionality.
To spin up a local Camunda Platform 8 Self-Managed environment, use one of the following docker compose configuration files:
- docker-compose.yaml: The full stack deployment including all Camunda 8 Components: Zeebe, Operate, Tasklist, Connectors, Optimize, Identity, Elasticsearch, Keycloak, Web Modeler, PostgreSQL
- docker-compose-core.yaml: Deploys Camunda 8 Orchestration Cluster Components: Zeebe, Operate, Tasklist, Connectors, Optimize, Identity
- docker-compose-web-modeler.yaml: Contains Camunda 8 Web Modeler for modeling purposes only - without Play or any Orchestration Cluster Components
To start a complete Camunda Platform 8 Self-Managed environment locally:
- Clone this repository.
- Run:
docker compose up -d
- Wait a few minutes for the environment to start up. Monitor the logs, especially the Keycloak container, to ensure the components have started.
Navigate to the different web apps and log in with the user demo
and password demo
:
- Operate: http://localhost:8081
- Tasklist: http://localhost:8082
- Optimize: http://localhost:8083
- Identity: http://localhost:8084
- Elasticsearch: http://localhost:9200
Log in to Keycloak using user: admin
and password: admin
to manage users.
- Keycloak: http://localhost:18080/auth/
The workflow engine Zeebe is available using gRPC at localhost:26500
.
docker compose down -v
Zeebe, Operate, Tasklist, along with Optimize require a separate network from Identity as you'll see in the docker-compose file.
This minimal setup does not include Optimize, Identity, Web Modeler and Keycloak:
docker compose -f docker-compose-core.yaml up -d
In addition to the local environment setup with docker compose, use the Camunda Desktop Modeler to locally model BPMN diagrams for execution and directly deploy them to your local environment. You can also use Web Modeler.
Feedback and updates are welcome!
By default, the Zeebe GRPC API is publicly accessible without requiring any client credentials for development purposes.
You can however enable authentication of GRPC requests in Zeebe by setting the environment variable ZEEBE_AUTHENTICATION_MODE
to identity
, e.g. via running:
ZEEBE_AUTHENTICATION_MODE=identity docker compose up -d
or by modifying the default value in the .env
file.
Both docker-compose files contain our out-of-the-box Connectors.
Refer to the Connector installation guide for details on how to provide the related Connector templates for modeling.
To inject secrets into the Connector runtime they can be added to the
connector-secrets.txt
file inside the repository in the format NAME=VALUE
per line. The secrets will then be available in the Connector runtime with the
format secrets.NAME
.
To add custom Connectors either create a new docker image bundling them as described here.
Alternatively, you can mount new Connector JARs as volumes into the /opt/app
folder by adding this to the docker-compose file. Keep in mind that the Connector JARs need to bring along all necessary dependencies inside the JAR.
A kibana
profile is available in the provided docker compose files to support inspection and exploration of the Camunda Platform 8 data in Elasticsearch.
It can be enabled by adding --profile kibana
to your docker compose command.
In addition to the other components, this profile spins up Kibana.
Kibana can be used to explore the records exported by Zeebe into Elasticsearch, or to discover the data in Elasticsearch used by the other components (e.g. Operate).
You can navigate to the Kibana web app and start exploring the data without login credentials:
- Kibana: http://localhost:5601
Note You need to configure the index patterns in Kibana before you can explore the data.
- Go to
Management > Stack Management > Kibana > Index Patterns
.- Create a new index pattern. For example,
zeebe-record-*
matches the exported records.
- If you don't see any indexes then make sure to export some data first (e.g. deploy a process). The indexes of the records are created when the first record of this type is exported.
- Go to
Analytics > Discover
and select the index pattern.
ℹ️ The Desktop Modeler is open source, free to use.
Download the Desktop Modeler and start modeling BPMN, DMN and Camunda Forms on your local machine.
Once you are ready to deploy or execute processes use these settings to deploy to the local Zeebe instance:
- Authentication:
None
- URL:
http://localhost:26500
If you enabled authentication for GRPC requests on Zeebe you need to provide client credentials when deploying and executing processes:
- Authentication:
OAuth
- URL:
http://localhost:26500
- Client ID:
zeebe
- Client secret:
zecret
- OAuth URL:
http://localhost:18080/auth/realms/camunda-platform/protocol/openid-connect/token
- Audience:
zeebe-api
Important
Non-production installations of Web Modeler are restricted to five collaborators per project. Refer to the documentation for more information.
Web Modeler can be run standalone with only Identity, Keycloak and Postgres as dependencies by using the Docker Compose.
Issue the following commands to only start Web Modeler and its dependencies:
docker compose -f docker-compose-web-modeler.yaml up -d
To tear down the whole environment, run the following command (including all the data and volumes):
docker compose -f docker-compose-web-modeler.yaml down -v
Warning
This will also delete any data you created.
Alternatively, if you want to keep the data, run without -v
parameter:
docker compose -f docker-compose-web-modeler.yaml down
You can access Web Modeler and log in with the user demo
and password demo
at http://localhost:8070.
The local Zeebe instance (that is started when using the Docker Compose docker-compose.yaml is pre-configured in Web Modeler.
Once you are ready to deploy or execute a process, you can just use this instance without having to enter the cluster endpoint manually.
The correct authentication type is also preset based on the ZEEBE_AUTHENTICATION_MODE
environment variable.
No additional input is required.
If you enabled authentication for gRPC requests on Zeebe, use the following client credentials when deploying and executing processes:
- Client ID:
zeebe
- Client secret:
zecret
Note
The correct OAuth token URL and audience are preset internally.
The setup includes Mailpit as a test SMTP server. It captures all emails sent by Web Modeler, but does not forward them to the actual recipients.
You can access emails in Mailpit's Web UI at http://localhost:8075.
When submitting an issue on this repository, please make sure your issue is related to the docker compose deployment method of the Camunda Platform. All questions regarding to functionality of the web applications should be instead posted on the Camunda Forum. This is the best way for users to query for existing answers that others have already encountered. We also have a category on that forum specifically for Deployment Related Topics.
When using arm64-based hardware like a M1 or M2 Mac the Keycloak container might not start because Bitnami only provides amd64-based images for versions < 22. You can build and tag an arm-based image locally using the following command. After building and tagging the image you can start the environment as described in Using docker-compose.
$ DOCKER_BUILDKIT=0 docker build -t bitnami/keycloak:19.0.3 "https://github.com/camunda/camunda-platform.git#8.2.15:.keycloak/"
You can control access to specific processes and decision tables in Operate and Tasklist with resource-based authorization.
This feature is disabled by default and can be enabled by setting
RESOURCE_AUTHORIZATIONS_ENABLED
to true
, either via the .env
file or through the command line:
RESOURCE_AUTHORIZATIONS_ENABLED=true docker compose up -d
You can use multi-tenancy to achieve tenant-based isolation.
This feature is disabled by default and can be enabled by setting
MULTI_TENANCY_ENABLED
to true
, either via the .env
file or through the command line:
ZEEBE_AUTHENICATION_MODE=identity MULTI_TENANCY_ENABLED=true docker compose up -d
As seen above the feature also requires you to use identity
as an authentication provider.
Ensure you setup tenants in identity after you start the platform.
Looking for information on Camunda Platform 7? Check out the links below: