onyx_connector (Resource)

A connector definition: what to index and how often. A connector on its own indexes nothing — pair it with an onyx_credential to start indexing.

Access control is not set here. Onyx applies it when a credential is associated, so it belongs to the connector-credential pair.

Example Usage

# A connector says what to index and how often.
resource "onyx_connector" "docs_site" {
  name       = "docs-site"
  source     = "web"
  input_type = "load_state"

  # Re-index once a day, prune monthly.
  refresh_freq = 24 * 60 * 60
  prune_freq   = 30 * 24 * 60 * 60

  connector_specific_config = jsonencode({
    base_url           = "https://example.com/docs"
    web_connector_type = "recursive"
  })
}

Schema

Required

• connector_specific_config (String) Source-specific settings as a JSON object, e.g. jsonencode({ base_url = "https://example.com", web_connector_type = "recursive" }).
• input_type (String) How the connector reads: load_state, poll, event, or slim_retrieval.
• name (String) Connector name. Must be unique for the source.
• source (String) Source system, lowercase, e.g. web, confluence, google_drive. Onyx rejects sources excluded by ENABLED_CONNECTOR_TYPES.

Optional

• indexing_start (String) Earliest document timestamp to index, RFC 3339, e.g. 2026-01-01T00:00:00Z. Onyx ignores it on update, so changing it replaces the connector.
• prune_freq (Number) Seconds between pruning runs. Onyx rewrites an unset value to its default of 604800 (7 days) on the first update, and Terraform then keeps that value.
• refresh_freq (Number) Seconds between index runs. Unset means index once, with no refresh.

Read-Only

• credential_ids (List of Number) Ids of the credentials paired with this connector.
• id (String) Numeric connector id.

Import

Import is supported using the following syntax:

The terraform import command can be used, for example:

#!/bin/sh
# Import by numeric connector id.
terraform import onyx_connector.docs_site 5