dbt Cloud job failure alerts, and runs that never started
A dbt Cloud job builds the models behind your dashboards every morning. You want failures in Microsoft Teams or by email, together with the alerts from your other jobs, and you also want to know when the scheduled run never started. Here is what dbt's own notifications do today, and where an outside monitor still adds something.
What dbt sends on its own
dbt's documentation on job notifications (updated October 2026) lists three channels: email, Slack and Microsoft Teams. Each can be triggered when a run succeeds, warns, fails or is cancelled. A few limits are worth knowing:
- Teams messages go only to channels the linked dbt user belongs to; Slack works in public channels, and account-level Slack connects to one workspace.
- Webhooks, for anything else, are available on the Starter, Enterprise and Enterprise+ plans, with events for a run that started, completed or errored.
- All of these are about runs. A scheduled run that never starts — for example because the job's schedule was switched off during a change and never switched back on — produces no run, so no notification is sent. That is our reading of the documentation: there is no event for a run that didn't happen.
So if you only need failures in Teams, dbt does that natively now. The gaps are missing runs, runs that take far longer than usual, and having dbt's alerts in the same place as the Python, Power BI and file jobs around it.
Watch dbt Cloud jobs with TaskBeacon
TaskBeacon's dbt Cloud integration reads the latest runs of each job you choose, with a read-only service token. No code changes in your project. For each dbt job:
- a successful run is OK with its duration;
- an errored or cancelled run is Failed with dbt's status message;
- a run in progress is Running, and Late when it takes much longer than usual;
- with the job's schedule set in TaskBeacon like the one in dbt, a run that didn't start is Missed.
The alerts go to the same places as all your other jobs: email (also to colleagues outside the dbt account), Slack, Teams, Discord, Google Chat, ntfy or a webhook, right away or once a day. The dbt Cloud page in the Docs lists the token permissions per dbt plan. Integrations are part of Pro, open to everyone for now, and this one is marked beta: not yet run on a real installation — tell us how it went.
Fresh runs, stale data
A dbt run that succeeds can still build models from old source data. dbt's own answer is freshness on sources (run with dbt freshness), which fails when loaded_at_field is older than error_after. Add it as a step of the scheduled job and the "stale" case becomes a failed run that TaskBeacon reports like any other. For data loaded outside dbt, see data freshness checks.
Without TaskBeacon
Turn on dbt's Teams or email notifications for failures and cancellations; for most teams that covers the failed-run case. For missed runs, a webhook receiver or a small scheduled script can call the dbt Cloud API each morning, look at the job's last run, and alert when it is older than expected. Or chain a "canary" job after the main one and have something outside dbt expect it, which is the heartbeat idea again.
Connected Claude or ChatGPT to TaskBeacon? Ask it "what failed last night, and why?" and it answers from your jobs. How that works.
Related
In the Docs: dbt Cloud jobs · Integrations · Schedules.
- Check that a daily table or file is freshA load that "works" but brings yesterday's data again.
- Power BI refresh failure notifications for the whole teamFailure emails only, only for scheduled refreshes, only inside your tenant.
- Get an alert when a job loads 0 rows, or far fewer than usualExit code 0 says nothing about how many rows arrived.
TaskBeacon is free to start, with every feature open for now and no credit card. Create a job, paste the code, and the first ping shows up on the job page.
Start free View live demoSources
- dbt docs: Job notifications (email, Slack, Microsoft Teams) (updated 2026-10-15)
- dbt docs: Webhooks for your jobs (Starter, Enterprise and Enterprise+ plans) (updated 2026-07-23)
- dbt docs: freshness (updated 2026-10-12)
Page checked against its sources on 2026-11-09. Menus and limits of other products change; if something here no longer matches, write to support@taskbeacon.io.