Monitoring · synthetic journeys · alert routing · incident review

Charles Trinh — Delivery Monitoring and Alerting

A service card reading DELIVERY MONITORING — receipts, journeys and alerts that reach a person
Charles Trinh — Delivery Monitoring and Alerting

Monitoring that measures what reached the patient or clinician rather than what left the server: delivery receipts, synthetic journeys, and alerts that reach a human.

What this is

Almost every monitoring stack answers the question the server asks: did the request return two hundred. Almost no outage is announced that way.

The gap between sent and delivered, or between rendered and usable, is where the real incidents live, and it is invisible to a request counter.

This page is about measuring the last mile: receipts where they exist, synthetic journeys where they do not, and an alert path that ends at a person with an action.

Sent is not delivered

Count what arrived, not what was dispatched; the gap between them is the entire problem.

Where a receipt exists, use it. Where it does not, use a downstream observation such as a portal login or a confirmation click as the proxy, and say plainly that it is a proxy.

Synthetic journeys

A scripted end-to-end path every few minutes catches the failures logs cannot see.

Journeys should touch the real dependencies and stop short of real patients: a test account, a test number, a harmless record. A journey that runs in isolation proves very little.

Alert routing

An alert nobody owns is a log line; assign every alert an owner and a first action.

Each alert names the service, the first diagnostic step and the person on call. If nobody can act on it at three in the morning, it belongs in a report, not in a pager.

Noise budgets

Alert fatigue is a design failure; tune thresholds against real incidents, not against fear.

Review every alert that fired last month against what actually happened. Alerts that never preceded an incident get demoted or deleted, and the page rate becomes a metric in its own right.

Dependency visibility

Monitor the third parties you depend on, because their outage is your outage and their status page lags.

Probe the dependency from your own network on your own schedule. A status page is a press release; a probe is evidence.

Retention for investigation

The window you can investigate is the window you retain.

Set retention from the worst case you must reconstruct, not from disk cost. The incident that needs a month of history will not announce itself in advance.

Post-incident review

A timeline reconstructed afterwards is the only reliable account; write it while it is fresh.

Write the timeline within a day, from logs rather than memory, and separate what happened from what should change. The second list needs owners and dates or it does not exist.

Dashboards for decisions

If a dashboard never changes a decision, delete it and keep the one that does.

Fewer, sharper boards get read. Each panel should name the decision it supports, and a panel that only reassures is the first candidate for removal.

How I do it

I start by listing what the business believes it measures and comparing it to what the systems actually count. The gap is usually one or two signals that nobody owns and one dashboard that everybody reads and nobody acts on.

Then the last mile is instrumented: receipts where they exist, a stated proxy where they do not, and synthetic journeys over the real dependencies. Alerts are written with an owner and a first action, and anything that cannot be acted on is demoted to a report.

The engagement closes with a review cadence: a monthly alert review against real incidents and a post-incident template that separates what happened from what changes. Monitoring that is not reviewed becomes decoration within a quarter.

Artifacts you keep

  • A sent-versus-delivered measure per channel with a stated proxy where needed
  • Synthetic journeys covering the critical paths, run on a short interval
  • An alert register with owner, first action and page classification
  • A monthly alert review comparing firings to real incidents
  • A post-incident timeline template with owners and dates for changes

What to monitor

Monitor the outcome, not the process: did the message reach the person, did the claim reach the payer, did the record reach the exchange. Process metrics tell you the system ran; outcome metrics tell you the work happened.

Pick one outcome per journey and alert on it. A dashboard of process metrics trains people to look at green tiles while the outcome degrades.

Alert design

An alert needs an owner, a first action and a threshold that means something. An alert without an owner is noise, and noise is how a real signal gets missed.

Route alerts to a person rather than a channel, and record the acknowledgement. The record is what shows whether the on-call process operates or merely exists.

Quiet failures

The dangerous failures are silent: a queue that stops draining, a token that expired, a feed that returns empty rather than an error. Detect emptiness explicitly, because an empty result and a correct empty result are indistinguishable without a check.

Alert on the absence of expected events rather than only on errors. Most of the incidents worth preventing produce no error at all.

Reporting upward

Three numbers for a sponsor: delivery rate on the journey, median time to delivery, and the count of quiet failures caught by monitoring rather than reported by a user. The third is the only evidence the monitoring works.

Trend them monthly with the same definitions, because a change in definition hides a change in performance.

T

he dashboard nobody reads

A

dashboard is not monitoring. If nothing pages, nothing is monitored, and the wallboard is decoration. Pick the journeys, alert on them, and keep the rest for the incident.

R

eview monthly which alerts fired and which were actionable. Retire the ones that never were, because a noisy alert set is the same as no alert set.

The test message

Send a real message through the real path on a schedule from outside the estate, and alert if it does not arrive. Platform-level monitoring cannot see a carrier block or a template rejection, and those are the failures that reach users.

Rotate the destination rather than using one address, because a single destination develops its own delivery quirks.

Record the last successful delivery per journey per destination, so a silent stop is visible as an age rather than as a gap in a chart. Age is the metric that survives a dashboard being rebuilt.

Alert on the absence of expected traffic during business hours, because a journey that stops entirely produces no errors and no alerts unless absence itself is instrumented.

Keep the alert history and review which alerts were actionable each month. A set of alerts that never required action is a set that will be ignored when one does.

Assign each journey an owner and review the journey list with them quarterly, because journeys are added by projects and retired by nobody.

Keep the journey definitions with the alert rules, so adding a journey without an alert is visible as an omission rather than a preference.

Review the alert thresholds against the last six months of history, because a threshold set once drifts into either noise or silence as the estate changes.

The review takes an hour and it is the difference between a monitoring system and a set of rules nobody adjusts.

Proof

  • ONC (g)(10) self-test PASS, 317 of 317 subtests, April 2026 — public suite URL
  • 67 public hostnames, 51 serving, one AWS account
  • 104 clearinghouse enrolments across 53 jurisdictions
  • USCDI v7 proposal — Patient Digital Identity Credential — listed at isp.healthit.gov
  • SAM.gov UEI MVKQW9DKKGB8 · CAGE 20KM5 · NPI 1194645564
  • AWS Business Associate Addendum active since 18 August 2026

Where this work is delivered

Engagements run from Las Vegas and reach across the valley: clients in Las Vegas, NV; Las Vegas, Nevada; Henderson; Clark County, Nevada; and North Las Vegas, Nevada, plus remote delivery nationwide.

How an engagement runs

1. Trace one real case

Follow a single patient or claim end to end with a timestamp at every hop, before any document is written.

2. Written findings

A findings list naming the defect, the system, the owner and the date.

3. Monthly review

Review the list monthly and check whether the dates held.

Frequently asked

What is this service?

This page describes charles trinh — delivery monitoring and alerting — the scope, the failure modes it prevents and how the work is measured. It sits in the AI and Automation category at trinh.io.

Where is the work delivered?

Engagements run from Las Vegas, Nevada across the valley — Las Vegas, Henderson, North Las Vegas, Spring Valley, Paradise, Winchester, Sunrise Manor, Enterprise, Summerlin, Clark County — and remotely anywhere in the United States. On-site work in the Las Vegas valley, remote delivery elsewhere.

What does it cost?

Standard rate $195 per hour, committed $165 per hour at twenty hours or more per week for three months or more; $1,600 standard day, $1,350 committed day. Scoped in writing before work starts, fixed fee where the deliverable is defined.

Who is Charles Trinh?

A technology architect in Las Vegas, Nevada, working in healthcare identity, consent and interoperability. Not a physician, not a medical practice, and not affiliated with any other Charles Trinh.

How does an engagement start?

With one trace of a real case and a written findings list naming the defect, the system, the owner and the date. The first call is twenty minutes and free: 702-559-3489.

Engagement

Scoped in writing before work starts; fixed fee where the deliverable is defined. First call is 20 minutes and free. 702-559-3489 · charles@trinh.io

Talk to Charles

TRINH — Charles Trinh, technology architect
TRINH

Charles Trinh

Fractional CTO/CIO and Technology Architect — Healthcare Data, Identity & Regulated Infrastructure · Las Vegas, NV

Las Vegas, Nevada · US citizen · Fractional and project engagements

(702) 559-3489 Email Charles Trinh