# JD Edwards Orchestrator Error Handling and Production Readiness Checklist > Source: https://netray.co/tools/jde-orchestrator-error-handling-production-checklist > Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems. > Section: Free Tools ## Short answer Check an EnterpriseOne orchestration before it goes live: what happens when a step fails, who it runs as, how it is monitored and scheduled, and how it is promoted. ## Category oracle ## Erp JD Edwards EnterpriseOne ## Audience - JDE CNC administrator - Orchestrator developer - JDE technical lead ## Tool **Kind:** checklist **Groups:** #### Design and error behaviour **Items:** **Text:** Every step has a decided On Error behaviour in Orchestrator Studio: abort the orchestration, continue and record the error, or abort and run an alternate orchestration or notification **Critical:** true **Text:** The orchestration response tells the caller clearly whether it failed (error output mapped or an error message returned), not a silent empty success **Critical:** true A failure path exists for unattended runs: a notification or a separate error orchestration that alerts a named owner **Text:** Retries are safe: re-running the orchestration after a partial failure does not create duplicate sales orders, vouchers or journal entries **Critical:** true Rules validate inputs (required fields, valid business unit, valid branch/plant, numeric quantities) before any form request writes data Form requests that write data were tested with invalid data so the JDE form error comes back in the response Cross references and white lists of allowed values are maintained in soft coding or cross reference records, not hard coded in steps Custom Groovy or Java steps catch exceptions and return a readable error instead of a stack trace #### Security **Items:** **Text:** The orchestration and every component (form requests, data requests, connectors, rules, cross references) are approved and shared UDOs, not personal objects **Critical:** true Run-time access is granted through UDO security or feature security to the roles that should call it, and denied to others **Text:** The user or service account the caller authenticates as has least-privilege application and row security in JDE **Critical:** true **Text:** Connections to external REST, database and FTP endpoints use connection soft coding per environment, with credentials not stored in step definitions **Critical:** true AIS is reached over HTTPS only and external callers cannot reach the AIS server directly without the agreed gateway or firewall rule Inputs that end up in database connector SQL are bound parameters, not concatenated strings #### Monitoring and scheduling **Items:** **Text:** Someone owns the Orchestrator Monitor and checks it for failed runs on a fixed rhythm **Critical:** true AIS server logs are located in Server Manager and their retention is known **Text:** AIS and HTML server logging is at the production level, not debug, after testing is finished **Critical:** true Scheduled orchestrations have a documented schedule, frequency and owner, and the scheduler is started on the production AIS server Schedules do not collide with the UBE batch window, package deployments or month-end close Notifications have real recipients (role or distribution list) and were tested end to end in the target environment A runbook says how to stop the scheduler, disable one orchestration and re-run a failed one #### Performance **Items:** Read-only lookups use data requests rather than form requests where possible, so no form processing runs just to read data Data requests and form requests return only the columns needed, with a maximum record count set Large reads page through results instead of returning one oversized response AIS session timeout and the caller's HTTP timeout are aligned so long runs are not cut off mid-update Expected call volume was tested against the AIS and HTML server capacity (threads, JVM heap) on a production-sized copy Loops over many rows were timed with realistic data volumes, not five test rows #### Promotion and testing **Items:** **Text:** UDO lifecycle followed: personal, then pending approval, then shared after review; rework requests closed **Critical:** true Components that need OMW (business functions, custom objects, Groovy dependencies) are in an OMW project promoted with the same change **Text:** Environment-specific values (URLs, users, paths, company numbers) are soft coded per environment, not edited inside the orchestration after promotion **Critical:** true The orchestration was regression tested in the target environment using the Studio test feature or a scripted REST call with saved inputs and expected outputs Callers (portal, iPaaS, IoT device, mobile app) were retested after promotion with the production endpoint and credentials Version and change record are documented so the previous version can be restored **Scoring Note:** Critical items are the ones that cause duplicate transactions, silent failures or security exposure in production. Feature names and screens vary by EnterpriseOne tools release; confirm against the Orchestrator Guide for your release. ## Content **Intro:** This checklist reviews a JD Edwards EnterpriseOne orchestration before production: error behaviour per step, security, monitoring, scheduling, performance and UDO promotion. It targets Orchestrator Studio on EnterpriseOne Tools 9.2 releases. **Sections:** #### How to use it **Body:** Run the checklist once per orchestration, not once per project. Sit with the developer and open the orchestration in Orchestrator Studio, then tick items only when you can see the evidence: the step setting, the UDO status, the soft coding record, the Monitor entry from a failed test run. Critical items are blockers. Export the result to PDF and attach it to the change request so the CNC team, the business owner and audit see the same record. **Bullets:** - Test one deliberate failure per write step before ticking the design group - Record the service account each caller uses - Keep the exported checklist with the OMW project or change ticket #### Where error handling lives in Orchestrator Studio **Body:** Each orchestration step has error handling options that decide whether a failure stops the orchestration or lets it continue. Choosing continue is only safe when a later step or the response reports the failure; otherwise a caller sees success while a sales order or voucher was never written. Form request steps return the same errors a user would see on the JDE form, so map them into the orchestration output. For unattended work, pair the orchestration with a notification so a person learns about failures, and use the Orchestrator Monitor to review runs. Scheduled orchestrations run under the AIS server scheduler, so its status and schedule belong in the runbook. #### Worked example: inbound order orchestration **Body:** A customer portal calls an orchestration that validates a ship-to, creates a sales order through a form request on P4210 and returns the order number. The review found three gaps. The validation step continued on error, so bad ship-tos reached the form request; it was changed to stop. The portal retried on timeout, creating duplicate orders; a lookup on the customer PO number was added before the create. The REST connector to the freight API had a test URL inside the step; it moved to connection soft coding per environment. All three were critical items. #### Version differences and gotchas **Body:** Orchestrator capability grew across Tools 9.2 releases: data requests, notifications, the scheduler, connectors and the Monitor arrived at different points, and screens moved. If an item refers to a feature your tools release does not have, mark it not applicable and note the compensating control. A common gotcha is approving the orchestration but not its components: a shared orchestration that calls a personal form request fails for every other user. Another is leaving debug logging on after testing, which fills disk and can log payload data. **Bullets:** - Check every component's UDO status, not just the orchestration - Confirm feature names in the Orchestrator Guide for your tools release - Turn logging back to production level after go-live **Cta:** If you want your team to ask JD Edwards questions in plain English through orchestrations running on your own servers, we can design and build that layer with you. **Faq:** #### How does error handling work in JDE Orchestrator? **Answer:** Each orchestration step has an On Error setting: abort the orchestration, continue processing, or abort and run an alternate orchestration or notification. Custom error text is returned in the response or exception and written to the AIS Server log. Form request errors come back as the JDE form error. Good practice is to stop on any failed write, return the error in the response, notify an owner for unattended runs and review failures in the Orchestrator Monitor. #### How do I schedule a JD Edwards orchestration? **Answer:** Orchestrator Studio includes a scheduler that runs orchestrations on the AIS server at a set frequency, with start and stop controls. Document each schedule, its owner and its window, keep it clear of the UBE batch window and package deployments, and test it in the target environment after promotion. #### What is the Orchestrator Monitor used for? **Answer:** The Orchestrator Monitor shows orchestration runs and their outcome so an administrator can find failures, see how long runs take and spot unusual volume. It works best with a named owner checking it daily, alongside AIS server logs in Server Manager for detail on a specific failure. #### Why does my orchestration work for me but fail for other users? **Answer:** Usually one component is still a personal UDO or is not shared, or UDO and feature security does not grant the other user access. Check every form request, data request, rule, connector and cross reference used by the orchestration is approved and shared, and that the caller's role has security to run it and the underlying applications. ## Search terms error handling in jde orchestrator, jde orchestrator error handling, jde orchestrator scheduler, jde orchestrator monitor, jde orchestrator production checklist, jde orchestrator udo approval, jde ais server orchestration security, orchestrator studio best practices --- Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json