A green “Running” status is not health
Journey Builder reports whether a journey is active. It does not report whether it is behaving correctly. A journey can sit at Running for weeks while silently dropping a third of its entries, sending twice to some contacts, and routing others down the wrong path.
None of the twenty-three failures in this audit will appear in your sandbox. They require volume, real data quality, and time — which is to say, they require production.
Entry is where most of the damage happens
The first cluster covers contacts that never enter, or enter catastrophically. The full-audience dump — activating a journey and having the entire source population enter at once rather than on a schedule — is the most expensive single mistake available in the product.
Subscriber Key mismatches are the quietest: contacts simply never appear, with no error anywhere. And API entry events can return HTTP 200 while injecting nothing at all, which means your integration monitoring reports success indefinitely.
The re-entry assumption
One belief causes more confusion than any other: that updating a row in the entry source will bring a contact back into the journey. It will not. Re-entry is governed by explicit re-entry settings, and the default behaviour surprises almost everyone the first time.
Version pinning sits alongside it. Contacts already in a journey continue on the version they entered on, so a fix you publish today does not apply to anyone currently mid-flight. Both settings are covered with their real behaviour rather than the assumed one.
Failures with no error
The most dangerous entries produce no alert at all. A single AMPScript error in one email kills the render for the whole send. Status exclusions silently remove contacts mid-journey. A misconfigured sender profile or reply-to address delivers mail that technically sends and practically fails.
Each of the twenty-three entries pairs the trap with the exact fix and, where one exists, the real error code to search for — so a symptom can be traced to a cause rather than guessed at.
What’s inside
- The full-audience dump on activation
- “It’ll re-enter when I update the row” — it will not, and here is why
- Subscriber Key mismatches: contacts that never enter at all
- API entry event failures that return 200 and do nothing
- Volume backlog stalling advancement mid-journey
- One AMPScript error killing the entire render
- Silent status exclusions, broken sender profiles and reply-to traps
- Re-entry rules and version pinning — the two most misunderstood settings
