The ask is simple. You have workflows doing real work — lead routing, follow-up email, deal stage updates — and you want to hear about it when one of them stops working. You should not have to find out from a sales rep asking why nobody has assigned them a lead for two weeks.
The answer is less simple, because it depends on your subscription tier. Some of the alerting you would expect does exist, but not on every tier. And the part that does exist on Professional is not really alerting. Here is what ships today, and what you can build on top of it.
HubSpot shipped Workflow Issue Notifications in December 2023. Its product team announced them on the Ideas board as live “for all Enterprise customers”, deliverable through Slack, email, Teams, mobile or in-app, and configured per workflow. In the settings UI the same area is labelled performance notifications. They came out of a long-running idea in the HubSpot Community, where customers had been asking for workflow failure alerts since 2017.
The important detail, quoted from HubSpot’s own documentation: “An Enterprise subscription is required to configure performance notifications.” If you are on Professional, you do not get them. That is the most common surprise for people searching for this feature. They read a release note or a community thread, go looking for the setting, and cannot find it.
Professional and Enterprise accounts both get a Workflow Health view in the workflows tool. It looks across your workflows and flags the ones that appear to be doing nothing. In particular it surfaces a workflow as Unused when it has had zero enrollments for 90 days.
That is genuinely useful for cleanup. It is not alerting. It is a page you have to remember to visit, and nobody remembers to visit it. The 90-day window also means a workflow can be dead for nearly three months before it shows up there, which is a long time for lead routing to be broken.
HubSpot has in-app and email notification preferences covering a range of events, and some workflow-related items appear among them. What those preferences do not cover is the case that actually hurts: a workflow that quietly stops enrolling anyone. Nothing errored. Nothing failed. The workflow is switched on and reports as fine. It simply has nobody going into it.
You can approximate alerting inside HubSpot itself. Create an active list whose criteria describe the contacts that should be flowing through a given workflow — contacts created in the last 7 days with a particular lifecycle stage and source, for example. Then build a second workflow that sends an internal email or in-app notification when that population behaves in a way you do not expect.
This works, and it costs nothing extra. The limits are real, though:
Use this for your two or three most business-critical workflows. Do not try to cover thirty of them this way.
HubSpot's automation endpoints expose more than the workflow tool shows you at a glance. Conceptually, you can read a workflow's validation state — whether HubSpot considers its configuration valid — and you can read its enrollment performance over time. That is enough to build the check that matters: compare this week's enrollments against last week's, and alert when a workflow that was busy goes quiet.
A small scheduled script can do this. Once a day, list the active workflows, pull the numbers, diff them against the previous period, and post anything suspicious to Slack or email.
The costs are also real, and worth naming before you start:
If you have engineering time and a handful of workflows, this is a weekend project. If you do not, it becomes the abandoned script nobody trusts.
The third option is to let someone else host option 2. You trade some control and grant a read scope, and in exchange you skip the token handling, the storage, the retries and the upkeep. When you evaluate one, the questions worth asking are narrow. Does it read only, or can it modify workflows? Does it need access to contact records, or only to workflow metadata? Where does it send alerts? And does it work on your tier, or does it quietly assume Enterprise the way the native feature does?
Most people go looking for error notifications. Error notifications are the wrong target, or at least an incomplete one, because the expensive failures raise no error at all.
Consider what actually breaks in a mature portal:
In every one of these, the workflow reports as healthy. There is no error to notify you about. Enrollments simply go to zero, and the only visible symptom is an absence — leads that never got routed, emails that never went out — which is exactly the kind of symptom people are bad at noticing.
So the signal worth alerting on is not "did this error." It is "did the volume change." A workflow that enrolled forty people last week and nobody this week is the alert you want, whether or not anything technically failed.
Workflow Watchdog is option 2, hosted. Once a day it reads your active workflows and posts to Slack only when something looks wrong: zero enrollments in the last 7 days when the prior 7 days had some, enrollments down sharply week over week, validation errors, and workflows with no edits in 90+ days — with the last editor named, so you can tell whether that person still works there. Quiet days produce no message.
It is read-only. It never modifies a workflow, and it does not read contact records. It works on Professional and Enterprise. Free for up to 15 active workflows.
workflow-watchdog-site.onpwr.workers.dev
Plainly: it is built by an independent developer. It is not HubSpot-certified, and it is not on the HubSpot App Marketplace yet.