Add Global Error Handler

[MitchellDumovic]

Related issue: #696

Make a customizable error handler for non-recoverable errors a "MUST" requirement for all languages. Includes examples from the existing Java and Go implementations of the error handler.

[carlosalberto]

@Oberon00 Your feedback has been addressed. Looks good to go? ;)

[carlosalberto]

@open-telemetry/specs-approvers Please review this PR - it's straightforward, and it's essentially a recommendation, rather a strict item, and it's definitely worth having IMHO ;)

[andrewhsu]

From the bug triage meeting today, looks like this change is good to be merged.

[bogdandrutu]

@andrewhsu this is a violation of our public rules, this change was not approved by 2 global approvers. It is a limitation of GitHub that anyone who is approver of one subdirectory counts as approver overall for the 2 required approvals, but our rules is that 2 approvals with ownership which cannot be guaranteed by the current GitHub settings.

Please make sure you enforce our public rules.

[andrewhsu]

@bogdandrutu apologies for the overstep in the public rules. in the spirit of increasing velocity, this one was targeted for merge at our otel spec issue scrub 🧽 mtg today (i called it bug triage above) given the limited scope and existing 2 approvals.

i admit it is not immediately clear for me to grasp the required minimum approvals for a PR, but now after some digging i found them in another repo here: https://github.com/open-telemetry/community/blob/9b841a653997ec3d708af22006f54b8600a26807/community-members.md#specifications-and-proto

i'll help stick to the public rules in the future.

[carlosalberto]

@bogdandrutu The fault is on me - let's discuss this on the Maintainers meeting next Monday how to proceed here (I'm OOO but happy to jump in for one hour).

For context: we all know approvals are not coming at a fast-enough speed, and that either we have the TC members jump in or we won't have all the important items ready for GA (I remember PRs being stuck because they only got 1 approval and were in the limbo for weeks, for example) - specially for items that are not polemic nor complex.

Also, for reference: this issue was discussed 4 or 5 weeks ago, there was initial overall agreement, then there was some push back by you, then we talked about this 2 weeks ago, and this PR is the result - which, by the way, is more of a suggestion than an enforcement. I'd imagine that by this time it would be clear this is not a polemic issue anymore.

Anyway, happy to discuss this on next Monday in the maintainers meeting.