Activiti 7 Kick Off Roadmap

Showing results for 
Search instead for 
Did you mean: 

Activiti 7 Kick Off Roadmap

Senior Member
7 7 7,733

If you were looking at the Activiti/Activiti repository, you might have noticed that we are restructuring the project. Activiti is taking a new direction towards microservices architectures and we are planning to make big design upgrades.

This new direction will give users the flexibility to open up the architecture, replace components as needed and to scale the engine and your applications independently. In combination with containerization (Docker), Microservice architectures with event-driven endpoints provide a cloud-native approach to deploy and interact with the engine in a distributed fashion.

However, we are well aware that users are currently embedding Activiti 5.x and 6.x in their applications. As we move toward the final release of the next major version of Activiti, we will provide a set of compatibility and migration tools. Embedding of the Activiti engine will continue to be an important feature for the project. You won’t need to re-architect your application to use this version of Activiti. However, new microservices architectures implementations will be able to take advantage of this new approach.

We understand that not everyone is using microservices yet but for those that are (and those that are going to) we need to make a real upgrade of the technology behind the engine and how the internal components integrate with each other. The Activiti Process Engine was designed several years ago, this means that there are some components, extension points, and mechanisms that have a high impedance with modern architectures and environments.

We recognize that most Open Source process engines out there claim that they are following microservices practices, but we haven’t seen the major refactoring in how the process engine works or how it integrates with cloud providers and other microservices that would be required to be a truly microservice architecture. Most of these other projects treat other microservices as REST endpoints that you can interact with. By doing so, you neglect the infrastructure that you are using and we want to leverage that. We also recognize the need for playing nicely with Containers and other technology stacks. For that reason, we need to upgrade how the engine and other services deal with versioning and immutability. With microservice architectures, if you want to integrate nicely in fast evolving environments you need to make sure that other technology stacks integrate well with your APIs and data types. For that reason, we are designing our new APIs to provide consistent data types and HAL (Hypermedia Applications Language) APIs to help you to build modern applications with cutting edge technologies.

We have created a short term roadmap for this initial refactoring, which will provide the stepping stones for more advanced services that will come towards the end of the year such as:

  • Distributed Process Execution and Coordination
  • Contextual Event Driven Case Management  
  • Decision Management and Inference Service
  • Blockchain integrated Audit Logs
  • Polyglot Process & Decision Runtime targeting IoT and mobile platforms

In the short term we want to focus into the process engine APIs and data types as well as our deployment model using containers and orchestrators such as Docker and Kubernetes.

Key points

Below a list of key points that we are going to cover in the short term:

  • Code Quality, Coding Standards, Code Maturity and Code Modularity
  • Infrastructural changes and tools that we are going to use to maintain the project healthy
  • New Rest HAL APIs, new Messaging endpoints with JSON payloads, new Event Listeners with JSON payloads
  • We will decouple all the components that are not part of the core process engine into new repositories and well-focused services
  • We will provide alignment with the Spring Community (Spring Boot 2 / Spring Cloud), targeting AWS + Kubernetes + Docker as our main deployment strategy
  • We will reuse as much of Spring Cloud as possible to make sure that the Process Engine doesn’t overlap with services provided by the infrastructure

By covering all these key points we plan to provide a set of services that you will be able to use as building blocks for your implementations, making sure that the process engine is only responsible for automating your business processes, and the impedance mismatch with your infrastructure is minimal. The following is the proposed roadmap for the next three months. We aim to test our release process by the end of July 2017, so the first one might be delayed a little bit.

Milestone #0 - 31 July 2017

  • Clean up Activity
    • Repository cleanup & restructuring
    • Dependency upgrades/Alignment with Spring 5 & Spring Boot 2
    • Infrastructure
      • GIT / Travis / Bamboo / Maven Central
      • Daily Snapshots
    • Test Coverage review
    • Testing frameworks review
    • Logging Frameworks review
  • Domain API + HAL API + Runtime Bundle 
    • Process Definitions
    • Process Runtime
    • Task Runtime
    • Source/Producer of Audit Events
  • Event Store for Audit Information - Initial Implementation
    • We should be able to query for all the events generated by multiple process engines
  • Identity Management and SSO (KeyCloak research in progress) - initial integration
  • First Release Process

Milestone #1 - 28th August 2017

  • Domain API + HAL API + Runtime Bundle
    • Source/Producer of Integration Events
    • (Source/Producer of Async Job Executions)
  • Event Store for Runtime Information - initial implementation
  • Gateway (Zuul research in progress) - Initial configuration and infrastructure
  • Application & Service Registry (Eureka research in progress) - Initial configuration and infrastructure
  • Tracer (Zipkin research in progress) - Initial configuration and infrastructure
  • AWS example demonstrating all these services in action

Milestone #2 - 29th September 2017

  • Event Store for Runtime Information - initial version
    • Decoupled query module based on events sourcing and rolling snapshots
    • We should be able to query the current state of processes and tasks without talking with the engine
  • Application Service / Case Management Features - initial version
    • Provide basic case management constructs
    • Provide coordination between different process runtime bundles
    • Deal with versioning, upgrades and case management like behaviors

Feedback are welcome

We encourage all users to start trying our milestones as soon as possible so you can get a gist of how these services are going to fit in your implementations.

We are looking forward for comments, questions, concerns, and collaborations. We will be working in a truly open source way, meaning that we want you to participate in the development of this and future roadmaps. We are looking for contributors from all levels, so if you feel attached to one of these topics, you want to learn new things or to participate in the conversation please get in touch.

Where to find more resources

We are using a Gitter channel here to discuss the roadmap, plan and collaborate every day, feel free to join:

We also have set up the following public services that you might find of interest:

Travis CI for continuous integration:

Codecov for code coverage reports:

Codacy for code style, security, duplication, complexity, and coverage:

These services will keep track of our code quality progress and release cycle.

There will be more announcements and blog posts about the topics mentioned here. We will be sharing progress and engaging community members regularly, stay tuned!

Original blog post: Activiti 7 Kick Off Roadmap – Salaboy (Open Source Knowledge) 

Active Member II

Hello! I'm curious about the Event Store enhancement. Would you mind describing a bit more what this entails?

Senior Member

Hi Greer,

Sure, I will write a blog post about this topic, but briefly, it will be a decoupled component in charge of storing events generated by the process engine, in the same way, that these events are stored in the History tables, but now in a separate component, so we can scale it independently. By doing this, we can provide different event stores depending on how do we want to consume/aggregate this information. 

Does this make sense? 


PS: feel free to join us in the Gitter channel where we have all these discussions and planning: Activiti/Activiti7 - Gitter 

Senior Member

As said above for the new release on 31st july 2017. Where can we download the new release? Can anyone provide the link?

Senior Member

Hi Amruta,

We are just starting with the development of Activiti 7 and we are moving towards monthly snapshots of our new services. During July as stated in this blog post, we were organizing ourselves to create these new services and refactoring the existing bits. We have gone through a release process but we have decided to not deploy this initial version to maven central just yet. You can, of course, use the source code in GitHub or the latest snapshots in alfresco maven repositories if you want to start testing the new services. 

Feedback is more than appreciated. 

Senior Member

Thank you for your reply. I thought there will be new release again. Thanks once again

Member II


Sure, I will write a blog post about this topic, but briefly, it will be a decoupled component in charge of storing events generated by the process engine, in the same way, that these events are stored in the History tables, but now in a separate component, so we can scale it independently. By doing this, we can provide different event stores depending on how do we want to consume/aggregate this information. 

Senior Member

‌ Not sure to understand your comment here. Feel free to ask a question, but copying and pasting a fragment of my response doesn't help.