Definition
To update or modify the definition of an Object Type, choose Components > Agents in the navigational menu, then choose the Object Type. The Object Type detail view will default to the Definition tab, which includes:
Agent ID: required, permanent unique identifier for the Agent, used throughout the Tenant Workspace UI, throughout your solution, and as the Agent API endpoint for an Http Agent.
Name: editable, this required field is used throughout the Tenant Workspace UI
Description: optional, used to describe the function of the agent.
Type: sets the type of the Agent to one of the following allowed types:
HTTP to Flow — HTTP Agent type for APIs or Web Pages
Event to Flow — Event Processing agent that receives messages from the Contextual messaging service.
Flow: The specific flow and version of the flow selected and currently in use for this agent
Entry Points (Event-to-Flow only): If the flow supports multiple entry points (defined by the use of multiple Event-Start nodes in the flow), then the entry point to use for this agent is selectable. This allows a single flow to contain the logic for more than one agent, and for the entry point to be selectable on a per agent basis.
Flow Version: Displays the current version of the flow used by this agent, and if a newer version of the flow is available, provides an "Update to Version nnn" button that allows you to update to the latest version and restart all agent instances to use that version.
Image Version: The version of the agent "image". New agent versions are published periodically that introduce new features. By allowing the image version to be specified, the decision about when to upgrade the agent to the latest version is left in the hands of the developer.
Parallel Instance Scaling Type: The scaling type for this agent. This is related to the agent-type in the sense that some scaling types only work with specific agent types.
Lag-Based Scaling — Lag based scaling allows new instances to be launched when the number of messages waiting to be processed by the agent exceeds a defined threshold. This type only works with "Event to Flow" agents.
CPU-Based Scaling — CPU-based scaling allows new instances to be launched when the CPU utiilization of the current instances consistently exceeds a defined threshold. This type works with any agent type.
The table below shows the definition fields associated with each of the two different scaling types:
| Lag-Based Scaling | CPU-Based Scaling |
|---|---|
Minimum Instances — The minimum number of instances to execute. | Minimum Instances — The minimum number of instances to execute. |
Maximum Instances — The maximum number of instances to execute. | Maximum Instances — The maximum number of instances to execute. |
Event Lag Threshold — The threshold of events waiting to be processed to exceed before launching an additional instance. | CPU Threshold — The threshold of CPU time to exceed before launching an additional instance. |
Instance Compute: The size of the agent to launch in terms of milli-CPUs and memory as shown in the table below.
| Size | mCPU | Memory |
|---|---|---|
Small | 150 | 512 MiB |
Medium | 300 | 1024 MiB |
Large | 1000 | 4096 MiB |
Extra Large (XL) | 2000 | 8192 MiB |
Extra-Extra Large (XXL) | 4000 | 16384 MiB |
Environment Variables: Allows developers to specify environment variables which provide configuration settings to the flow. The environment variables are accessible to function nodes within the flow.
Deleting an Agent
On the Agent Definition tab in the Tenant Workspace UI, there is a Danger Zone to delete the Object Type. Once confirmed, the Agent and all log data associated with that Agent will be immediately deleted, an operation which cannot be undone. However the flow associated with the agent is not deleted, allowing the agent to be recreated if needed.
Last updated