Understanding Caching in the NetBox Importer

Caching plays a crucial role in the NetBox to Nautobot data import process. This document explains the caching mechanisms, their purpose, and provides practical examples.

Source Identifiers to Nautobot Primary Keys Mapping

Maps source identifiers to Nautobot primary keys, ensuring consistent references. This caching is separate for each SourceModelWrapper instance and is defined in _uid_to_pk_cache attribute.

It's possible to manually pre-cache particular source IDs to Nautobot primary keys. This is necessary to link existing Nautobot objects to the source data.

To cache particular source record to an existing Nautobot primary key, use SourceModelWrapper.cache_record_uids method. This is used e.g. to properly link existing Nautobot Roles to the source data, as NetBox has multiple models for separate roles, while Nautobot has a single model for all roles.

For more information about primary keys mapping see UID to PK Mapping.

Records Caching

Caching Record Data

This is the base caching method used for other records caching techniques. It caches data for optional later use. If cached record is referenced by imported data, it will be imported automatically; otherwise, it will be ignored.

To cache record data, use SourceModelWrapper.cache_record method.

Setting the Default Reference

It's possible to specify the default reference for a model. This affects relation fields referencing this model, but with missing or (NULL) source data. The record is cached, when the default reference is set.

The following example creates a default reference for the dcim.manufacturer model. All empty manufacturer field values will be replaced with reference to Unknown manufacturer.

def configure_importer(adapter):
    manufacturer = adapter.configure_model(
        "dcim.manufacturer",
        default_reference={
            "id": "Unknown",
            "name": "Unknown",
        },
    )