IFS Cloud and Looker integration

What moves between them

Financial and operational records flow from IFS Cloud into Looker. Supplier invoices, purchase orders, GL accounts, vouchers, and customer orders are polled from IFS Cloud on a daily or weekly schedule aligned with your close cycle or reporting cadence. The extracted records are mapped to Looker's warehouse table structure and written via Looker's API or data load actions, then surfaced in queries and dashboards. Cost allocation dimensions and GL account hierarchies are synchronized to ensure Looker's financial rollups match IFS Cloud. The flow is pull-based on a schedule rather than event-driven, since IFS Cloud does not offer standard webhooks.

How ml-connector handles it

ml-connector stores both OAuth2 credential sets encrypted: the IFS Cloud client_id and client_secret (exchanged for a 60-minute bearer token), and the Looker client credentials (exchanged for a 1-hour token). For each poll cycle, ml-connector queries IFS Cloud's OData API with filters on modified date and state fields, respecting the 5000-element page size recommendation and backing off on HTTP 429 rate-limit responses. Before updating records, ml-connector reads the ETag from IFS Cloud to satisfy the If-Match header requirement for concurrent mutations. Records are transformed and mapped to Looker's warehouse schema, then written via Looker's API. If a Looker token expires mid-request, ml-connector reauthenticates and retries. IFS Company Code is configured per customer, so all GL and financial queries are scoped to the correct legal entity. Every record carry a full audit trail and can be replayed if a downstream API call fails.

A real-world example

A multinational manufacturing company runs IFS Cloud for production, procurement, and general ledger across three plants and regional sales offices. Finance and supply chain teams use Looker dashboards to monitor cash flow, supplier spend, and production costs. Before the integration, teams exported GL and invoice data from IFS Cloud weekly, loaded it manually into the Looker warehouse, and waited for overnight ETL to update dashboards. With IFS Cloud and Looker connected, invoices and GL postings sync automatically every night, Looker dashboards refresh on schedule, and finance teams have current visibility into accruals and commitments without manual intervention.

What you can do

• •Sync IFS Cloud supplier invoices, purchase orders, and GL accounts into Looker on a configurable daily or weekly schedule.
• •Map IFS Cloud GL account hierarchies and cost dimensions to Looker warehouse tables so financial rollups match the source system.
• •Handle IFS Cloud OAuth2 token refresh and ETag-based optimistic concurrency, and manage Looker token expiration and re-authentication.
• •Extract and transform IFS Cloud financial records while respecting OData page size limits and rate limits on both systems.
• •Maintain a complete audit trail of every record sync so failed or stale transactions can be identified and replayed.

Questions

Does Looker store ERP data natively, or does ml-connector write it to a warehouse?

Looker does not natively store ERP data. ml-connector extracts records from IFS Cloud and writes them into Looker's connected warehouse (such as BigQuery, Snowflake, or Redshift) via Looker's API data load actions or by writing directly to warehouse tables if credentials are available. The warehouse then serves as the source for Looker's modeled looks and dashboards.

How does ml-connector handle IFS Cloud's ETag requirement and OData paging?

IFS Cloud requires an ETag header (If-Match) for any mutation operation (PATCH or certain POST actions). ml-connector reads the ETag when fetching a record and includes it in the subsequent update request. For large result sets, ml-connector respects IFS Cloud's 5000-element page size recommendation, paginates using OData skip and top parameters, and backs off on HTTP 429 rate-limit responses with exponential backoff.

Why does the sync use polling instead of webhooks?

IFS Cloud has no standard webhook subscription API. Event Actions exist but require manual per-customer configuration in the IFS admin UI and are not API-driven, making them unsuitable for automated connector integrations. Polling on a schedule tied to your reporting cadence is more reliable and allows ml-connector to sync on a predictable timeline that aligns with your close process and dashboard refresh windows.