question-mark
Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Add a --continue-after-error option

See original GitHub issue

Describe the feature

A new flag to attach to a specific step(s) of a run whose failure should not block the rest of the job from completing. Consider this job in our project, which separates fragile from non-fragile tests - if a normally robust test starts failing, I’m much more concerned compared to the fragile ones which tend to self-heal once the pipelines catch up.

dbt run
dbt test --exclude tag:fragile
dbt test -m tag:fragile

If a robust test fails, the fragile tests don’t get a chance to be tested at all.

Describe alternatives you’ve considered

The current workaround is to make multiple jobs. This doesn’t work with the exposure metadata tile beta, which needs all steps to be in a single dbt Cloud job:

dbt source snapshot-freshness
dbt test -m +exposure:*

If a single source is stale, all tests will be skipped. The workaround to that would be to make a separate job for each exposure, something like:

dbt source snapshot-freshness --select +exposure:my_first_exposure
dbt test -m +exposure:my_first_exposure

but that would get intractable after more than a few exposures.

Additional context

In my heart, I think of this as the --soldier-on flag, but something like --continue-after-error is a more practical name

Who will this benefit?

  • People using the new metadata tiles
  • People running separate jobs who could flatten them down to a single job again, which simplifies scheduling to avoid race conditions (currently I schedule the second test step like an hour after the first job, to make completely sure that everything I’m testing is actually finished building)

Are you interested in contributing this feature?

Interested, but unlikely to make a move on it anytime soon

Issue Analytics

  • State:closed
  • Created 3 years ago
  • Comments:8 (5 by maintainers)

github_iconTop GitHub Comments

3reactions
joellabescommented, Feb 1, 2022

@mrogove you’re welcome to re-open issues! Closed doesn’t mean “we’ll never do this”, it just helps to prevent issues whose moment has passed from sticking around forever.

In this case though, it’s not actually a change that can be made in dbt core, so it makes more sense for you to submit a feature request to the dbt Cloud support team using the chat bubble inside the app

2reactions
Crimsonabysscommented, Nov 16, 2022

hi, just following up this issue. has anyone raised the feature request yet? if not I am happy to do it.

Read more comments on GitHub >

github_iconTop Results From Across the Web

Add a --continue-after-error option · Issue #3015 · dbt-labs/dbt ...
Describe the feature A new flag to attach to a specific step(s) of a run whose failure should not block the rest of...
Read more >
How to continue flow execution after an error - It Ain't Boring
When handling errors in the Flows, we can add parallel branches so one of those would run on success, the other one would...
Read more >
How to continue the INSERT even after the error
I have millions of rows to insert in SQL Server, but some of them are in error, I would like to know if...
Read more >
On-Error Components | MuleSoft Documentation
When an error occurs in a Mule app, an Error Handler component routes the error to the first On-Error component (On Error Continue...
Read more >
Continue Package Execution After Error in SSIS - Tim Mitchell
You'll notice that there are no tasks in the error event handler for the data flow task (DFT Load File). Although you can...
Read more >

github_iconTop Related Medium Post

No results found

github_iconTop Related StackOverflow Question

No results found

github_iconTroubleshoot Live Code

Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start Free

github_iconTop Related Reddit Thread

No results found

github_iconTop Related Hackernoon Post

No results found

github_iconTop Related Tweet

No results found

github_iconTop Related Dev.to Post

No results found

github_iconTop Related Hashnode Post

No results found