Handle all kinds of errors in the workflow

[codeadict]

What

Handles all kinds of errors in the workflow, including "stop loop" (throw, exit) errors from Erlang that could be sent from Finch lower in the stack.

Why

Finch errors not not being caught by the rescue block, DD.

Jira ticket

CARS-13980

Steps to Validate/Verify

1. In cars_platform apps/engine/mix.exs and apps/cars_web/mix.exs update the :pacer dep with {:pacer, git: "https://github.com/codeadict/pacer/", branch: "handle_erlang_stop_loops"}
2. Verify an ISA (the cat picture listing) loads
3. Verify the SRP behaves normally

[zkayser]

Can we add an extra unit test that shows that we return the default value when a batched field throws or exits too? Similar to this test

pacer/test/workflow_test.exs

Lines 985 to 993 in 0fccfcb

	test "all sorts of batch issues, but returns defaults and or resolved results" do
	assert %BatchWorkflowWithMultipleResolverIssues{
	a: 1,
	b: "ddddefault",
	c: "timed out so default here for c",
	d: 24
	} ==
	Pacer.Workflow.execute(%BatchWorkflowWithMultipleResolverIssues{})
	end

In fact, I think we can probably just add a resolver in that batch that throws an error and/or exits and add to the test for it too.

[zkayser]

Just added a comment on adding to an existing test case that shows we can handle multiple different classes of failure on batch resolvers, but this looks good. Verified that we still handle the class of exception we have previously handled with rescue clauses and the unexpected cases where we got throws/exits.

Thank you!