Understanding the Push API Processing Delay

When you use the Push API to add or update one or several items in a Push source, the service normally returns a 202 Accepted response almost instantaneously. This merely means that the items were successfully forwarded to the Push API queue; it does not imply that the items were successfully processed and indexed.

Typically, you must wait between 2 and 5 minutes before your newly pushed items begin appearing in your index. You cannot directly alter or reduce this delay, as it is the sum of the smaller delays which are incurred while the items are transiting through the different stages of the Coveo Cloud V2 indexing pipeline.

The infrastructure automatically scales depending on load variations to maintain the lowest possible processing delays. Moreover, the Coveo team constantly monitors those delays and regularly fine-tunes the infrastructure scaling mechanisms.

The total Push API processing delay may vary depending on a number of factor such as:

  • The number and size of items to process.
  • The current load on the Push API queue (which is multi-tenant).
  • The current load on the document processing manager queue (which is multi-tenant).
  • The index commit frequency (typically every minute, although the index commits immediately when its queue reaches a certain threshold).
  • The Coveo Cloud V2 platform infrastructure scaling delays (e.g., it can take a few seconds to provision an AWS Lambda function container).