Run analysis
Evaluates an analysis configured in the event type against the current event and publishes the result to the following conditions.
| Configuration | Content |
|---|---|
| Analysis | analysis existing in the event type |
| Setting: Result | name under which the result is published; default: the analysis name |
The result is published as a temporary variable, with the keys listed under Published keys:
| Key | Label in the panel | Content | Presence |
|---|---|---|---|
score | Score | calculated score | when the run evaluates at least one score cell, even with a zero value |
score_cell | Score cell | cell the score came from | when the run assigns the event to a cell |
triggered | Triggered | whether any cell triggered | always |
triggered_count | Triggered cell count | how many cells triggered | always; zero triggers is a reading, not an absence |
triggered_cells | Triggered cells | names of the triggered cells | always; empty when no cell triggers |
version | Version | version of the analysis that ran | always |
Condition 1: $amount > 0 → Run analysis: bot
Condition 2: %bot.triggered_count >= 2 → Set field: review = "true"
Condition 3: "email_velocity" in %bot.triggered_cells → Add to list: $user_email to @blocked_emails
Under the Cells heading, the list also shows one key per cell of the analysis, %<result>.cells.<cell>.value, with the raw value the cell measured in the window. The action does not publish this key when the cell is not evaluated. A cell with no trigger delivers its number through this key, equally available to the other cells.
Writing to the event
The action writes nothing into the event. To make a reading queryable in metrics and monitors, add a set field after the action, as described in Published results. The triggered_cells set is written into a text field with join:
Set field: bot_cells = join(%bot.triggered_cells, ",")
Set field: bot_version = %bot.version
Writing the version key separates, later, the events decided by one version of the analysis from those decided by the next. A rule that wrote the version through the former Version destination no longer fills in that field. To fill it in again, add a set field with %<result>.version, as in the example above.
Result name
The result name only needs to be provided when the same rule runs the same analysis twice. Each run is its own result, and the result name is unique within the rule.
This action only reads: it does not feed the analysis’s windows. To feed them, use record analysis.