Extremely Confused - External Trigger not starting Job

[hollow idol]

I have 2 jobs that use a custom trigger of ExternalSourceTrigger (for which I have also created a matching ExternalSource to register the webhooks). They have the same exact trigger configuration, myIntegration.customEvent(). This creates the ExternalSource registering the webhook and should create the ExternalSourceTrigger that causes both jobs to run when that event is triggered.

However, only one of the jobs runs, and not the other. I then deleted the running job (hoping for some magic), and still one of the jobs is not running. How can I debug this? What might be going on?

I have been following the instructions found here:

https://trigger.dev/docs/integrations/create-triggers

[woeful sundial]

It’s very hard to say without a lot of detail. Building your own integrations is really hard with that system. It’s one of the reasons we’ve decided not to do this at all in v3.

[hollow idol]

So what does V3 integrations look like then?

[woeful sundial]

We don't have integrations in v3, at least not yet. There are no webhook triggers at the moment.

Inside the run function you just use normal Node.js SDKs since there are no timeouts

[hollow idol]

Oh...that is going to be a problem for us then...

[woeful sundial]

Which webhooks are you using?

[hollow idol]

I guess we will have to stick with V2.

Right now we are using a HTTP Endpiont and a generated webhook from an ExternalSource()

[woeful sundial]

Yeah using HTTP endpoint is way easier than building an integration

Our recommendation for v3 users at the moment is to either receive the webhooks to their own servers and then trigger tasks with the payload. Or to use a product like Hookdeck to receive webhooks and then trigger tasks

[hollow idol]

I just liked the idea of automatically registering the webhooks in the source system

I already add trigger as a bolt on, and then now I would also need to add Hookdeck? It's just really a lot at this point

We also have to deploying everything in our own AWS VPC - no other cloud services

FWIW, the fix to the original problem was to create a new job with a new ID. then it just started working all of a sudden

[woeful sundial]

If we do add webhooks to v3 the registration of them will be done differently. That issue you experienced is because they automatically register when the code runs and that means you can't get good error messages, it was a mistake.

We might build webhook triggers for v3, we've not decided for sure yet. They're a ton of work to build and to maintain. It definitely won't be anytime soon because there are so many things to get right with the rest of the product first and that seems to be where more demand is.

[hollow idol]

Understandable. Tradeoffs are hard.

That said I can confidently say I will not move to V3 for the forseeable future then