Make scenario error monitoring.

What Make tells you on its own

When an error stops a run, Make sends an email, then a digest every 5 minutes, and another one when it disables a scenario.¹ It disables a scenario after three consecutive errors, and scenarios with an instant trigger (webhook) after the first.² Incomplete executions, which hold the error message and the run's data, are only stored when you switch them on per scenario.³

That works for three scenarios. With thirty, spread across teams and clients, the mails get lost, and nobody notices that the same error has happened every night for a week.

What Fehlo adds

• One issue per cause. Fehlo groups by scenario, module and kind of error. IDs, numbers, links and email addresses in the message are ignored for grouping, so a module that fails with a different record ID each time stays one issue.
• Frequency and history. How often, since when, last seen, like any other issue in Fehlo.
• Straight to the run. Each issue links the failed run and the scenario in Make.
• Alerts where you want them: new issues, spikes and regressions by email, Microsoft Teams, Discord, Google Chat, ntfy, Zulip or webhook.
• Next to your software. When a scenario fails because your API threw an error, both issues are in Fehlo.

Option 1: through the Make API (recommended)

Every 5 minutes Fehlo reads the failed runs of all scenarios in a Make organisation. Your scenarios stay as they are.

1. Create a project in Fehlo and pick "Make" as the platform.
2. Create an API token in Make: profile, "API" tab, "Add token". Give it the scopes scenarios:read, organizations:read, teams:read and dlqs:read.
3. Paste the token into Fehlo. Fehlo finds your account's zone (eu1, eu2, us1, us2) by itself. If the token reaches several organisations, pick one. You can limit monitoring to some teams.
4. Test it: make a scenario fail and click "Check now" instead of waiting 5 minutes.

Why dlqs:read: Make's run logs carry no error text. Only incomplete executions do.⁴ So Fehlo reads both: failed runs (GET /scenarios/{id}/logs with status error) and open incomplete executions (GET /dlqs).⁵ Without the scope, runs still come through but without a message, and Fehlo shows a hint. For incomplete executions to exist, "Store incomplete executions" must be on in the scenario.

Fehlo stores the token encrypted and only reads with it. Disconnect, and Fehlo deletes it.

Option 2: an error handler

Without an API token, use an error handler with an HTTP module. You set it up in each scenario:

1. Right-click the module that can fail, then "Add error handler".
2. "HTTP", "Make a request": method POST, the webhook URL from Fehlo, body type "Raw", content type "JSON (application/json)".
3. Paste this template as request content, replace MODULE with the module name, and map error type and message from the error bundle into TYPE and MESSAGE:

{
  "scenario": { "id": "{{var.scenario.id}}", "name": "{{var.scenario.name}}" },
  "execution": { "id": "{{var.execution.id}}" },
  "module": { "name": "MODULE" },
  "error": { "type": "TYPE", "message": "MESSAGE" }
}

The webhook URL always answers 202, even when Fehlo drops a report, so the error handler itself never fails.

Limits and cost

Each failed run counts as one event against your plan's quota; the Free plan includes 5,000 a month. Fehlo accepts at most 300 errors per hour for each Make connection and webhook URL. The first check looks back 24 hours, later checks only at new runs. Admins, managers and owners can connect and change it. If a scenario runs without errors but too rarely, use a heartbeat monitor: the scenario calls a URL at the end, and Fehlo alerts when the call doesn't come.

Questions