Lumenbase Portage

Unified Portage wizard

Open Lumenbase Portage in Data Hub, then connect your source CRM.

Dependency-ordered import

Imports pipelines, stages, companies, contacts, deals, then activities so links resolve in one pass.

User owner mapping

Map source CRM users to Lumenbase members; unmapped owners surface as preview warnings.

Entity type selection

Import only what you need; skipped types never touch the target workspace.

Background batch jobs

Track live progress by the summary bar and view the full log in Import History.

Sprint + slow drip

Entity types over 10,000 records import in a paced second phase to protect API budgets and server limits.

Pause & cancel

Pause an in-flight import or cancel a scan; sessions resume from saved checkpoints.

Connection validation

Each origin validates credentials before scan (OAuth token, Attio API token, HubSpot service key).

Workspace membership

Only members of the workspace can start or continue a migration session.

Active session guard

One active Portage session per workspace per origin; prevents overlapping imports for the same CRM.

Preview warnings

Unmapped users, zero-count entity types, and deals-without-companies relationship gaps.

Scan schema version

Full pagination counts with no estimates from partial pages.

Duplicate-safe resume

import_record_changes keyed by session + source_system + source_id skips already-imported rows on continuation.

Provenance-or-rollback

If provenance write fails after insert, the entity row is deleted so resume never creates silent duplicates.

Immutable entity origins

Each imported record gets a migration origin tag (provider, session, external source id) for audit and lineage.

Chunked self-reinvoke

Scan and import edge functions checkpoint progress and re-invoke on wall-clock or early_drop shutdowns.

Adaptive chunk sizing

Import batches shrink after a timeout and grow after successful runs.

Rate-limit awareness

Pipedrive: token budget headers and automatic pause metadata. All origins: stale-detector respects pause flags.

Stalled job recovery

Cron + UI resume restart failed batch jobs when the session is still scanning or importing with a saved checkpoint.

Session rollback

Chunked rollback via execute-import-rollback removes provenance-tracked entities and clears the session.

Import invocation limits

Migration edge functions cap self-reinvoke depth to prevent runaway loops.

Cancel propagation

Cancelled sessions abort continuations and mark linked batch jobs cancelled.

import_record_changes

Every imported row logged with migration_session_id, source_system, source_id, and target record_id for rollback.

entity_origins

First-origin certificate: origin_kind=migration, origin_system=provider, origin_run_id=session, external_source_id=source key.

Batch job operational log

Scan and import milestones appended to batch_jobs metadata for support and recovery.

Per-provider batch job types

Distinct job types (e.g. attio_migration_import) for monitoring and stale-detector rules.