How does ml-connector handle Oracle JD Edwards session tokens that expire after 30 to 60 minutes?

ml-connector monitors for HTTP 444 responses (invalid token) and obtains a fresh session token on demand by POST to the tokenrequest endpoint with the stored credentials. This ensures continuous availability without manual intervention. Tokens are tied to a service account license that must remain valid.

Since Oracle JD Edwards has no native webhooks, how does ml-connector stay current with new data?

ml-connector polls Oracle JD Edwards tables on a schedule you define, using date filters on UPMJ (updated date) and DGJ (GL date) to read only changed records. It tracks the last-polled timestamp so unchanged records are skipped. You set the polling cadence to match your business cycle, for example nightly for customer updates or after each sales cycle for invoices.

Oracle JD EdwardsHubSpot

Oracle JD Edwards and HubSpot integration

Oracle JD Edwards runs your financials, procurement, and supply chain. HubSpot runs your sales and customer relationships. Connecting the two keeps your customer master and your CRM in sync. New customers and account changes in Oracle JD Edwards flow to HubSpot contacts and companies, and invoices and payments created in HubSpot post back to Oracle JD Edwards so your accounting stays current without re-entry. ml-connector handles the very different APIs on each side and moves the data on a schedule you control.

How Oracle JD Edwards works

Oracle JD Edwards EnterpriseOne exposes customers, vendors, accounts payable ledger, invoices, purchase orders, items, and GL accounts through REST APIs via its Application Interface Services Server, running at your infrastructure. Authentication uses session tokens obtained by POST with username and password, valid for 30 to 60 minutes and passed in the jde-AIS-Auth header on all calls, or HTTP Basic Auth per request in Tools 9.2.5.2 and later. Oracle JD Edwards has no native outbound webhooks, so data is read by polling the underlying F-tables with date filters on UPMJ (updated date) and DGJ (GL date). Pagination uses maxPageSize (default 100 records) with moreRecords flags for continuation.

How HubSpot works

HubSpot exposes contacts, companies, deals, invoices, line items, payments, and orders through REST APIs at https://api.hubapi.com. For server-to-server integrations it authenticates with a Private App Access Token (Bearer token starting with pat-). Writes to invoices and commerce objects are supported. HubSpot can also push events via webhooks with signature verification, though webhook subscriptions require a separate Public App registration. Batch operations are capped at 100 records per request, cursor-based pagination uses a limit of up to 200 records, and rate limits vary by tier from 100 to 190 requests per 10 seconds.

What moves between them

The main flow is bidirectional. Customer and account data from Oracle JD Edwards Address Book and Customer Master tables flows into HubSpot as contacts and companies. Invoices and payments created in HubSpot flow back to Oracle JD Edwards for posting to accounts payable and the general ledger. Reference data such as item masters and GL account structures is synced in both directions to ensure invoices land on valid accounts and cost dimensions. Oracle JD Edwards is read by polling on a schedule you control since it has no native push capability.

How ml-connector handles it

ml-connector stores the Oracle JD Edwards AIS Server URL and credentials encrypted, obtains session tokens on demand, and refreshes them when a call returns HTTP 444 (invalid token). It polls customer, account, and invoice tables on your sync schedule, tracking the last-polled timestamp to avoid re-reading unchanged records. On the HubSpot side it uses Private App token authentication and posts contacts, companies, and commerce records via batch REST calls. Oracle JD Edwards requires direct association between contacts and invoices via the F0403 Customer Master, and HubSpot requires line items to exist before invoices can be created, so ml-connector builds associations in the correct order. Rate limiting is handled with exponential backoff, session tokens are tied to a service account license that must remain valid, and every record carries a full audit trail so failed postings can be replayed.

A real-world example

A mid-sized distribution company runs Oracle JD Edwards for ERP and HubSpot for sales and customer relationships. Before the integration, the sales team managed customer records in HubSpot while accounting maintained a separate customer master in Oracle JD Edwards, creating friction when customers changed addresses, phone numbers, or payment terms. Invoices were created in HubSpot but required manual re-entry into Oracle JD Edwards for accounting. With Oracle JD Edwards and HubSpot connected, customer changes in Oracle JD Edwards flow to HubSpot automatically, and invoices created by sales in HubSpot post to Oracle JD Edwards without manual re-keying. The two systems stay in agreement and the accounting team's workload drops.

What you can do

Sync customer and account data from Oracle JD Edwards Address Book and Customer Master to HubSpot contacts and companies.
Flow invoices and payments created in HubSpot back to Oracle JD Edwards accounts payable for automatic posting to the general ledger.
Handle Oracle JD Edwards session token expiry and refresh on demand, and manage HubSpot Private App token authentication.
Poll Oracle JD Edwards on your schedule since it has no native webhooks, tracking last-polled timestamp to avoid redundant reads.
Map invoice line items and payments between HubSpot commerce objects and Oracle JD Edwards F-tables with full audit trail and replay on failure.

Questions

Which direction does data move between Oracle JD Edwards and HubSpot?: Both directions. Customer and account data flows from Oracle JD Edwards to HubSpot as contacts and companies. Invoices and payments created in HubSpot flow back to Oracle JD Edwards for posting to accounts payable and the general ledger. Both systems stay synchronized.
How does ml-connector handle Oracle JD Edwards session tokens that expire after 30 to 60 minutes?: ml-connector monitors for HTTP 444 responses (invalid token) and obtains a fresh session token on demand by POST to the tokenrequest endpoint with the stored credentials. This ensures continuous availability without manual intervention. Tokens are tied to a service account license that must remain valid.
Since Oracle JD Edwards has no native webhooks, how does ml-connector stay current with new data?: ml-connector polls Oracle JD Edwards tables on a schedule you define, using date filters on UPMJ (updated date) and DGJ (GL date) to read only changed records. It tracks the last-polled timestamp so unchanged records are skipped. You set the polling cadence to match your business cycle, for example nightly for customer updates or after each sales cycle for invoices.

Connect Oracle JD Edwards and HubSpot

Free to use. Add your credentials, ping your real systems, and see if we fit.

Get started