Dispatch Policy

The dispatch policy in Signals determines how the events are delivered to the target systems after the publisher has emitted them. It provides options to handle them either individually or collectively that cater to a multitude of use cases and preferences. You can configure this policy based on your requirements, ensuring efficient and timely processing of events within the system.

Instant

In this policy, events are delivered to the target immediately after they occur in the publisher application. This enables continuous monitoring of publisher activities and performance metrics. It results in triggering instant actions such as processing payments, updating inventories, and can also help in preventing potential security breaches.

By default, Signals retains events for up to 24 hours between the publisher and target. This Time to Live (TTL) period can be customized in hours or minutes to suit your workflow needs.

In some scenarios, events may be received in bulk. By default, these are instantly delivered as a batch. If you prefer to deliver each event individually, you can enable the Send as Single Event option under the Instant ploicy.

Batch

This type helps in the collective delivery of events to the target systems. It enables you to gain insights from overall events, particularly in scenarios such as campaign responses and log management to trigger downstream workflows accordingly. The Time To Live (TTL) period for each event within Signals is 24 hours by default and cannot be changed.

Signals offer four dispatch types to define the event collection that must be delivered to the target. Following are the available dispatch types:

• By Count - The Signals wait until the configured events count is reached before delivering them to the targets as a batch. Following are a couple of exceptional cases in this type:

  • If the events reach the maximum size of 1MB before reaching the configured count, Signals collect the remaining events in a new batch and deliver them to the targets as multiple batches.
  • If the configured count is not reached within the TTL period, the events will be delivered to the target systems based on the One-Time Dispatch policy.

• By Size - In this type, events are delivered when their combined size matches the configured size.

If the configured size is not reached within the TTL period, then the events are delivered to the targets based on the One-Time Dispatch policy.

• By Interval - Events are delivered in regular intervals to the target. If events reach the maximum size of 1mb before the interval time, they are dispatched in multiple batches at the configured interval.

• By Schedule - Events are delivered daily at the configured time, and it can be even customized to your timezone.

If events reach the maximum size of 1MB before the scheduled time, they are dispatched in multiple batches at the configured scheduled time.

One-Time Dispatch

The One-Time Dispatch policy of Signals helps in preventing data loss in specific scenarios. The retry counts configured for the Target will not be applicable, and failed events cannot be retrieved again.

It occurs at the following two different instances:

• Before Clean-Up - It takes place at the end of everyday to clear the data of failed events. Before cleaning-up, Signals checks for any batch events that have not met the dispatch conditions within 24 hours. If any are found, they are delivered to the target.

• Instant - If any changes have been made to the dispatch types of an existing rule, events accumulated within Signals according to the previous dispatch type are instantly delivered to the target systems.