Skip to main content
Do not forget, when upgrading to a new platform version, always ensure that your installed component versions match the versions specified in the release notes. To verify this, navigate to FLOWX.AI Designer > Platform Status.
After updating to 3.3.0 FLOWX.AI release, it is not possible to import old process definitions into the new platform release (available for exports from releases < 3.3.0).

Component versions

With the release of FLOWX.AI 3.0, there have been some changes that you need to be aware when upgrading to the latest version:
  • The flowx-process-renderer has been migrated to @flowx\ui-sdk.
  • As of FlowX 4.0, the paperflow-web-components library will be deprecated. Instead, the new components can be found in @flowx/ui-toolkit.
FlowX.AI supports any version of the third-party components listed as prerequisites.For optimal performance and reliability, our internal QA process validates new releases using specific versions as indicated in the provided table. While exploring alternative versions that suit your company’s specific requirements, we recommend referring to the compatibility matrix for guidance.In the unlikely event that you encounter any compatibility issues with FlowX.AI, please open a support ticket here, and our dedicated team will address and resolve any identified bugs following our standard support process.Compatibility Matrix:
  • FLOWX.AI Platform: Recommended and tested versions
  • Third-Party Components: Supported versions based on specific requirements and client preferences

Additional configuration

This section describes the additional configuration that is required to use the new features in FlowX.AI.

Process engine

Process Instance Indexing through Kafka transport

Introducing a new Kafka transport strategy for sending details about process instances to be indexed in Elasticsearch. To enable indexing of process instances in Elasticsearch through Kafka, configure the following environment variables:
  • FLOWX_INDEXING_ENABLED
  • FLOWX_INDEXING_PROCESSINSTANCE_INDEXING_TYPE
For Kafka indexing, the Kafka Connect with Elastic Search Sink Connector must be deployed in the infrastructure.
  • FLOWX_INDEXING_PROCESSINSTANCE_INDEX_NAME: specify the name of the index used for process instances
  • FLOWX_INDEXING_PROCESSINSTANCE_SHARDS: set the number of shards for the index
  • FLOWX_INDEXING_PROCESSINSTANCE_REPLICAS: set the number of replicas for the index
Process engine new Kafka topics
For more details please check the following section: Process Instance Indexing through Kafka transport

New service account

Added a new service account flowx-process-engine-sa. This service account is needed so the use of Start Catch Event node is possible. Service accounts

Events gateway

Added a new events-gateway microservice, which requires the following configuration. The events-gateway is designed specifically for handling events. Previously, each process-engine pod had a WebSocket (WS) server, and the front-end (FE) would connect to the process-engine to receive messages. Now, instead of a server holding the messages, they are stored in Redis. However, the process-engine sends the messages to the events-gateway, which is responsible for sending them to Redis. Users connect to the events-gateway using an HTTP request and wait for Server-Sent Events (SSE) to flow in that request. They keep the request open for as long as they want SSE on a specific instance.

Events-gateway Kafka topics

New Kafka topics have been added for the events-gateway. These topics are used to send and receive messages between the events-gateway and the process-engine. New Kafka topics that should be added in the events-gateway configuration. Events gateway Events gateway setup guide Process engine topics

SSE

Starting with the 3.3 platform release, the WebSocket protocol has been removed. Therefore, if you are using socket.io-client, you will need to make some changes. Here’s what you should do:
  1. Uninstall socket.io-client:
Before proceeding, ensure that you uninstall socket.io-client from your project. You can do this using the following command:
  1. Install event-source-polyfill@1.0.31:
To replace the functionality provided by socket.io-client, you will need to use a new package called event-source-polyfill@1.0.31 (as mentioned in the Installing the library section). This package serves as a polyfill for the EventSource API, which enables servers to send events to clients over HTTP. The EventSource API is commonly used for server-sent events (SSE) and real-time web applications.

Message events

New Kafka topics that should be added in the process-engine configuration.

Bulk updates

New Kafka topics that should be added in the process-engine configuration, related to task management plugin - bulk updates.

Example

Process engine topics

Migration Steps

To upgrade to FLOWX.AI 3.3.0, follow these steps:
  • Make sure you have taken a backup of your current platform and database configurations.
  • Verify that your current installed component versions match the versions specified in the release notes.
  • Update the FLOWX.AI platform and all related components to the recommended versions.
  • Update the necessary configuration files according to the additional configuration requirements.
  • Restart the FLOWX.AI platform and related services.
  • Verify that the platform is running correctly and all processes are functioning as expected.
  • If you encounter any issues or errors during the upgrade process, refer to the troubleshooting section in the release notes or contact FLOWX.AI support for assistance.

Troubleshooting

If you encounter any issues during the upgrade process or while running the FLOWX.AI platform, refer to the troubleshooting section in the release notes or contact FLOWX.AI support for assistance.
Last modified on July 7, 2026