Live EDI operations / evidence ledger
Run an EDI control tower that separates delivery from acceptance
Trace the business flow from purchase order through shipment, invoice, transport receipt and functional or application acknowledgement.
Direct answer
Direct answer
An EDI monitoring dashboard should correlate documents by business reference and show each independent state: validation gate, release approval, exact-byte delivery, retry or dead-letter status, functional acknowledgement and application exception. Shipment Sentry uses the same tenant-scoped operational ledger for the live control tower and its exports, so a green transport receipt cannot be mistaken for a 997, 999, 824, TA1, CONTRL or APERAK acceptance.
Why a parser is not enough
A transport-centric dashboard answers whether a file moved. An operations team needs to know whether the correct bytes moved, which order or shipment they belong to, whether the receiving system acknowledged them, and who owns any exception. Keeping those states separate makes retries safer and investigations faster.
What the control should check
- 01Correlate purchase order, acknowledgement, shipment and invoice documents under one business key.
- 02Show the validation gate and immutable content hash before delivery.
- 03Record every managed-SFTP or signed-webhook attempt, response, latency and receipt.
- 04Apply bounded retries and move exhausted work into a visible dead-letter queue.
- 05Match syntax, functional and application acknowledgements by controls and business reference.
- 06Link every KPI and status to the underlying document evidence.
Working example
PO-77821
850 purchase order accepted
855 order response acknowledged
856 shipment notice delivered / 997 accepted
810 invoice retry scheduled
Exception queue 1 owner actionThe timeline is business-key driven. A missing acknowledgement or dead-letter delivery remains visible even when another document in the same flow succeeds.
Failure modes worth catching
- A successful SFTP write is shown as partner acceptance.
- An acknowledgement is matched to the latest document instead of its controls.
- A retry sends changed bytes under the original release identity.
- An aggregate count cannot be traced to a tenant-scoped record.
Put it into the dispatch workflow
- 1Admit and validate the inbound or outbound document.
- 2Correlate it to the business transaction and active partner route.
- 3Release the exact validated bytes through an approved delivery channel.
- 4Record each attempt, receipt, retry and terminal delivery state.
- 5Attach acknowledgements and exceptions without collapsing their meanings.
- 6Use the control tower or export the same ledger for investigation.
Questions
Implementation answers
Is transport delivery the same as EDI acceptance?
No. Managed SFTP or an HTTPS response proves a transport outcome. TA1, 997, 999, 824, CONTRL and APERAK messages express different syntax, functional or application outcomes.
Can users drill into the dashboard counts?
Yes. Flows link to their validation evidence, while the dispatch ledger shows hashes, attempts, receipts and errors.
What happens after all delivery retries fail?
The dispatch moves to a dead-letter state, creates an operational exception and requires an explicit owner replay after the cause is corrected.
Sources and limits
Reviewed 2026-07-20. Public references cannot establish every partner-specific rule. Current implementation guides and agreements remain controlling.
Keep investigating