This documentation is intended for beta users only. This functionality is not available to most Appian customers, and is not part of the released and supported platform. If you are interested in enrolling in the beta program, please contact your Appian representative. |
Elastic Process Execution (EPEx) is a new process execution engine that supports automatic horizontal scaling of unattended executed processes, unlocking high-throughput automation and process orchestration use cases. The new engine makes efficient use of memory and compute resources and supports enhancements to Appian task capabilities.
The beta program will allow selected customers and partners to gain hands-on experience with EPEx ahead of it being generally available (GA).
This program is intended for participants who can build apps that belong to at least one of the following use case categories:
At a minimum, at least one Appian developer needs to be assigned to this project. Depending on the scope of the application, it is possible that other roles such as architects and business stakeholders would need to be involved as well.
No additional licenses will need to be purchased for the beta program.
The time commitment depends on the complexity of the use case and the number of Appian developers that are dedicated to the project. One developer building an application from scratch is going to require more time than a few developers making updates to an application that has already been built.
We also ask that participants in the beta program make time to meet with the Appian product team regularly throughout the process. This will likely be more frequent at the beginning of the project (~1 hour a week) and less frequent as the project goes on (~30 minutes every other week).
The Appian product team will meet with participants regularly to review this feedback and answer any questions. Participants are encouraged to document feedback to bring to these meetings.
Yes, all organizations participating in the beta program must sign a legal agreement, separate from any existing agreements or contracts that are in place.
Appian will provide a summary of customer feedback and other findings from the program.
Appian will provide a new site on the pre-release version where customers have access to build applications from scratch, or import their existing applications. Appian employees will have access to this site. Additionally, Appian employees can create accounts for any user from the participating organization.
The newest beta functionality can be updated as frequently as once a week. Appian will determine when an upgrade is available, and work with the customer to set a maintenance window where the upgrade will take place. Any updates to the beta functionality will require about an hour of downtime on the relevant site. Participants can request the timeframe where this upgrade takes place.
The new process execution architecture will support unattended (automation) loads after the initial release. Process models must only contain supported nodes and configurations in order to run on the new engine. Customers will need to remove any unsupported nodes and configurations from their process models to run them using autoscaling mode.
If your process model contains only supported nodes, you can opt to run the process in the new engine by following these instructions.
Eventually, the plan is to be compliant with most existing process models regardless of the nodes they use. Still, some minor backwards incompatibilities may change existing node behavior. See additional details on changed behavior. All process models need to be tested after enabling autoscaling mode to ensure the results are acceptable for your application.
If the new technology introduces constraints or if we find some existing behavior is at odds with the future benefits for our customers, we will evaluate that behavior on a case-by-case basis. All process models will need to be tested after enabling EPEx to ensure expected results.
Smart service plug-ins are not supported in autoscaling mode.
Activity chaining is not supported because autoscaling mode does not currently allow attended tasks. If you have a use case that you believe requires chaining, let us know so we can evaluate the need for supporting chaining or something similar to address your use case.
We do not have plans to mix compliant and non-compliant nodes in the same process model. Since the execution is in a completely new engine for the entire process model, this is technically not feasible.
When designing your processes, take into account the maximum available connections on your database and the data source connection pool size. Knowing these limits will help you make informed choices about creating and updating data in your process model and help to ensure that Appian does not try make too many active connections.
In the event that all available database connections are being used, any additional threads will wait for the next available connection to be open. This helps process execution continue with little interruption.
All out-of-the-box expression capabilities are compatible. Internal testing suggests that plug-in functions run in script tasks are compatible as well, but you should thoroughly test any third-party plug-in functions used in your processes.
We are working on a new licensing model, which is necessitated by the new elastic scaling of resources. For the same process load that you have today, we don't expect your costs to increase under the new model. However, if you add new processes that use substantially more resources, we expect cost to go up proportionately. Remember, the beta program is not going to cost you, and you can use this opportunity to test your future load.
During the beta program, Appian may not fulfill all scalability requests. We'll do our best to accommodate any scaling requests but can't guarantee that they will be fulfilled for all use cases. For more information, create an Appian support case with EPEx beta
in the title.
To help accommodate these cases, you can work with your POC to load test your processes. Together, you can run tests with low load and review the metrics. This insight will help identify and adapt any bottlenecks, so you request additional scale only if needed.
FAQ