Skip to content

Mindbox connector

Mindbox support is implemented as a canonical Pull API and async export provider through source.type: api and api_type: mindbox.

Supported resources

Typical resources include:

  • orders
  • customers
  • actions
  • promo_actions
  • points_changes

Each resource defines its own date-window semantics and output schema.

Authentication

Recommended Vault secret:

{
  "endpoint": "https://api.mindbox.ru",
  "token": "...",
  "endpoint_id": "..."
}

api_key is supported as a backward-compatible alias, but token is preferred.

Manifest example

source:
  type: api
  api_type: mindbox
  connection_id: mindbox
  connection_type: vault
  resource: actions
  options:
    since_datetime_utc: "2026-01-01T00:00:00Z"
    till_datetime_utc: "2026-01-01T01:00:00Z"

sink:
  type: bigquery
  connection_id: bigquery_landing
  table:
    dataset: landing
    name: landing__mindbox__actions
  strategy:
    mode: replace

Window semantics

  • Event-like resources use sinceDateTimeUtc and tillDateTimeUtc.
  • Day-based resources use sinceDate and tillDate.
  • If explicit datetime bounds are omitted, the runtime falls back to day-based windows controlled by utc_boundary_time.

Default utc_boundary_time keeps upstream-compatible behavior for business-day boundaries.

Dependencies

Streaming JSON parsing uses ijson when available. Without it, the connector falls back to standard JSON loading, which is acceptable for tests but not recommended for large production exports.

Testing

Mindbox live tests require vendor credentials and should run as manual or scheduled certification jobs.

DPONE_RUN_INTEGRATION=1 \
DPONE_RUN_INTEGRATION_LIVE=1 \
uv run pytest tests/integration -m integration_live -k mindbox