DX: Bump stack trace limit to infinity

This is a Feature Proposal

Description

Our stack traces in our error reports are often partial traces because the default in node is a measly 10 lines. It seems acceptable to incur the performance cost of bumping our stack trace limit to Infinity so that our stack traces are complete when an error occurs. This will give us much better error reporting with the recent addition of error tracking in PR #3564 and help users deliver better traces when reporting errors/bugs.

Error.stackTraceLimit = Infinity;

[HyperBrain]

The current sentry configuration only sets up the default global exception handler anyway (with raven.initialize()). This captures unhandled exceptions by default, but not unhandled rejections. If you want to propagate unhandled rejections to Sentry too, you’d have to explicitly enable the unhandled rejection handler in raven: https://docs.sentry.io/clients/javascript/usage/#promises

You have to define/overwrite the Bluebird unhandled rejection handler explicitly and call Raven.captureException() in there.

[pmuens]

🤔 @brianneisler how would this work in a Promise context. We already have the longStackTraces option enabled.

I tried to place Error.stackTraceLimit = Infinity; in bin/serverless but this won’t change anything about the stack traces since Bluebird seems to handle the rejections.

This is already included in the sentry package (as @DavidWells mentioned).

Should this only apply for our error tracking via sentry or should users be able to see the long stack traces as well?

If it’s only for error tracking, then I agree that it’s sufficient enough to let sentry take care of that. If it’s for users then I’d vote to add it somewhere in our code again so that we don’t have a dependency on the sentry package (and this functionality won’t disappear if we decide to remove sentry later down the road).