Multi-Part Jobs

​

What the hell is a Job Part?

A Job can be split up into multiple parts depending on the intended functionality. This can help spread out large data sets or long tasks across many separate listeners. When a job is split up into parts, the original part is considered the “parent” part and it’s children are considered “parts”. Each Part is considered it’s own separate Job and has the full ecosystem of events attached to it. When all Job Parts have been completed, a new event job:parts-completed is emitted, so that you can listen for that event to complete the Parent Job. This event has a payload that has summary statistics about the parts.

{
  "payload": {
    "parts": {
      "total": 10,
      "completed": 10,
      "failed": 0,
      "canceled": 0
    }
  }
}

By inspecting this payload, you can determine if the all have completed successfully or not. Based on that information, you are expected to complete or fail the parent Job.

Here’s an example of a listener that creates parts and waits for the job:parts-completed event:

Here’s an example of a listener that does logic on each Part: