Daily monitoring

Outcome

Each morning you read the EDI dashboard, the response-processing queue, and the Issues queue, and you have a clear picture of what flowed yesterday — what's stuck, what's failing, what needs the operator's hand on it today.

Prerequisites

Scope	What it lets you do
edi.dashboard.read	View dashboard panels
edi.transaction.read	Inspect transactions
ops.issue.read	View issues queue

Familiarity with the rest of the EDI manual — daily monitoring is the crosscutting page where every transaction-type chapter eventually points back to.

The dashboard

/dashboard in the edi-app. Fixed header with date range + partner filter; tiled body of panels.

Panel	What it shows	Healthy range
Outbound volume	837 / 270 / 278 / 276 counts last 24h, broken by partner	Steady relative to historical baseline
Inbound ack latency	Median + P95 time from outbound to 999 / 277CA / 271	< 5 min for clearinghouses; < 1 h for direct payers
Stuck transactions	Outbound stuck in submitted past expected ack window	0
Response processing queue depth	How many inbound files await processing	< 10
999 rejection rate	% of outbound functional groups rejected	< 1%
277CA rejection rate	% of claims rejected (A3/A6/A7/A8)	< 5%
Companion guide validation failures	Count of outbound builds halted by validate rules	0
Connection probe health	Per-partner probe status	All green

Click any panel to drill into the underlying list filtered to its context — a stuck-transactions click opens /transactions?status=submitted&age>SLA, for example.

The response-processing queue

/responses shows the queue of inbound files waiting for processing. The queue is normally near-empty; non-zero depth means either:

• A file is currently parsing (transient).
• A parsing failure halted the file (operator action needed).

Each row carries:

Column	Notes
Inbound file	Filename / source ISA13.
Partner	The trading partner.
Status	parsing, parsed, failed.
Tx count	Number of ST-bounded transactions inside.
Failure reason	Populated when failed.

A failed row needs operator attention — see 8.2 — Replay & retry.

Steps for daily triage

1. Open /dashboard. Read the headline numbers vs. yesterday.

2. Scan stuck transactions. Anything past its ack window?

   Transaction	Expected ack	If past
   837 → 999	Minutes (clearinghouse) to hours (direct)	Check transport health on the partner (2.5 — Probe).
   837 → 277CA	Minutes to a day	Check companion guide for new validate rules.
   270 → 271	Seconds (real-time) to minutes (batch)	Real-time: partner-side issue. Batch: confirm response file landed.
   278 → 278 response	Minutes to hours	Some payers pend by default; check the MSG text on response when arrives.

3. Open /responses. Confirm queue depth is low. Any failed rows? Resolve via 8.2.

4. Open /issues in the rcm-app (cross-app — same issues queue). Filter to category EDI. Take the highest-severity row first.

5. Spot-check one rejection from the 277CA rejection rate panel. Open the rejected transaction; read the STC narrative; confirm the rejection class is normal (a typo in the source claim) vs. systemic (a payer rule changed).

6. End-of-day: scan the dashboard for stuck-transaction count trending up or rejection rate climbing. These are usually slow- burn issues that begin gradually.

Looking at one partner

/partners/:id → Activity tab shows partner-scoped versions of the same panels:

Panel	Per-partner detail
Recent transactions	Last 200 transactions with this partner.
Probe history	Probe results over the past 7 days.
Ack latency trend	This partner's median + P95 vs. tenant average.

When a problem is suspected to be partner-specific, start at this view.

Steps to escalate to a partner

1. Confirm the problem is partner-side. Use the dashboard + partner activity to rule out tenant-side issues — bad companion guide, wrong routing rule, expired credential.

2. Capture the evidence: filename / ISA13, expected vs. actual behavior, sample timestamps. Export from /transactions if needed.

3. Use the partner's published support channel. Reference their own ISA13, which their team can look up faster than ours.

4. Track the open ticket in your tenant's issue queue with category EDI + tag partner-issue. Close when the partner confirms resolution.

Validation

Check	Expected
Dashboard renders without errors	Yes.
Stuck-transaction count is 0 by end of business hours	Yes (when partners are responding normally).
Response queue is empty	Yes — anything else demands action.
277CA rejection rate stable vs. baseline	Yes.

Troubleshooting

Symptom	Cause	Fix
Dashboard shows zero outbound for a partner that should be active	Routing rule pointing elsewhere, or partner deactivated	Check routing (3.3) and partner active flag.
Ack latency P95 spike for one partner	Partner-side slowdown	Probe; if probe is fine, the partner is processing slowly — let it ride or escalate.
Response queue depth climbing	Parser repeatedly failing on one file	Open the failed entry; replay or quarantine (8.2).
Companion guide validation failures rising	A bound guide added a new validate rule that fires on real data	Loosen the rule or fix the source claim shape.

Next

8.2 — Replay & retry