Package Mechanics
The following pages contain information relating to the core logic and processing of our packages, it is not required to understand these concepts to make use of our packages but they are provided for more advanced users who may wish to produce more complex custom models or better understand what's going on under the hood of the packages. We refer to these elements throughout the docs.
Throughout the packages and these docs, we will refer to a session
, or session_identifier
. This is the atomic unit of re-processing, as described in the incremental sessionization logic and by default usually refers to the actual session identifier within your tracker; however this language is in part only used for historic reasons. Thanks to the custom identifiers in our packages, these sessions can refer to any identifier you define and our packages will ensure that all events (within the limits) are processed in a single run. A similar case applies to user identifiers as well.