`session-trigger` (Processing Node)

The session-trigger node reprocesses a target node in another session based on metadata matching rules.

Use this node to propagate updates from a source session (for example, an individual user session) into another session (for example, a profile or collective session) by triggering reprocessing on a specific node.

Configuration Schema

Property	Type	Required	Description
`target-node`	`string`	Yes	Node name in the matched target session to reprocess.
`searchCriteria`	`object`	Yes	Criteria used to locate the target session. Must contain `session` and `expression`.

searchCriteria fields:

Field	Type	Required	Description
`session`	`string`	Yes	Target session name to search for in the project.
`expression`	`string`	Yes	JMESPath expression applied to target session `metadata_raw` to extract comparison value.

How it works

At runtime, the processor:

Reads latest upstream inputs and expects dashboard payload under key dashboard.
Parses dashboard JSON and reads dashboard_users.
Lists sessions in the current project.
Filters sessions whose name equals searchCriteria.session.
Parses each candidate metadata_raw JSON.
Evaluates searchCriteria.expression against metadata.
When a match is found, loads node instances for that session and finds target-node.
Calls reprocess_node on the first matching node instance and returns:
- {"reprocessed_node_id": "<node-instance-id>"}.

Intended cross-session usage

Typical pattern:

Source session: per-user session emits dashboard payload.
Target session: profile-level or collective session holds a node that should refresh when user data changes.
session-trigger bridges them by selecting the target session and forcing reprocessing on target-node.

Runtime behavior and caveats

Fails when dashboard input is missing.
Fails when dashboard_users is missing in parsed dashboard payload.
Fails when exceptions occur during lookup or reprocess flow.
If no matching node instance is found in the target session, it logs and continues scanning.
If no match is found, it returns None without reprocessing.

Current implementation details to be aware of:

Input lookup is currently hard-wired to input_map.get("dashboard").
Matching compares searchCriteria.expression result with dashboard_users value from payload.
Validation enforces target-node string, searchCriteria object with session/expression, and valid JMESPath syntax.

Example configuration


{
  "target-node": "profile-aggregator",
  "searchCriteria": {
    "session": "user-profile-session",
    "expression": "user.externalId"
  }
}

Integration in a session template

Place session-trigger after a node that produces dashboard JSON including dashboard_users.
Configure searchCriteria.session to the target session name.
Configure searchCriteria.expression to extract the comparison key from target session metadata.
Set target-node to the node name you want to reprocess in the matched session.

session-trigger (Processing Node)