Using Sentry to track and manage "background job failures". Good idea?


I’m interested in using Sentry to track failures of different kinds of background jobs (eg. Resque, cron jobs etc.) These jobs are not caused by a particular visitor on my site (no user context), meaning they couldn’t be tracked using some client-side JS code. Rather, they occur in backend servers.

Nevertheless, they are failures of an application and need to be tracked by developers. They also have a context: which server ran this job, what type of job is it, how long did it take to fail etc. They also should have some kind of custom actions so that a developer can “replay/requeue” the failed jobs (more on this in my other question).

Is Sentry a good tool for this job?

Thanks in advance.

Sentry could definitely track this and it sounds like you’d want to use our Ruby SDK.

As far as getting Sentry to requeue the job… that’s not something we can support right now. In theory, a plugin could be written to provide behavior here, similar to creating a ticket for tracking the issue in another system. You could hook in an action to replay, which maybe just fires off all the data captured to some endpoint, which could requeue. I’m not entirely sure how you’d implement that, but a plugin could support it.

Thanks, that’s what I wanted to know.

As far as the “replay” option, yes, it’ll have to be done via an HTTP endpoint exposed by the queueing application (also described in my linked question).

Thanks again!