Add --limit flag to pulumi stack history and consume from automation api

[EvanBoyle]

We had some reports of perf issues related to pulumi stack history. Specifically we use this command to fetch update summary after an update with automation API. As you perform more updates, performance degrades as we fetch more data and decrypt more secrets. This change does two things:

1. add a --limit=N flag to pulumi history and pulumi stack history to allow only fetching the most recent N records.
2. consume this flag in the appropriate places in the Automation API SDKs for node, go, and python.

Note that this is technically a small breaking change for Go (other languages mark this as optional and previous SDK shape is still valid) but this is within the boundaries we've set for the alpha and is an important change to make for performance.

We don't have any infrastructure setup to test the filestate backend changes. There are tests that create stacks, select stacks, but none that go through the lifecycle of create, do an update, etc. Adding this was outside of what I have bandwidth for, but I tested this workflow locally with the filestate backend. See #3760.

Fixes #6253
Fixes #6252

Note this will need to be ported to C# after that lands.

[komalali]

LGTM except for a funky conditional.

[komalali]

I think we're missing an else here.

Suggested change

	}
	limit = len(allFiles) - limit
	} else {
	limit = len(allFiles) - limit
	}

[EvanBoyle]

The condition is a little funky here, but it is written as intended. The first condition is the equivalent of min(limit, len(allFiles) incase the limit is larger than the total number of results:

		if limit > len(allFiles) {
			limit = len(allFiles)
		}

After that, we adjust the loop condition:

limit = len(allFiles) - limit

[komalali]

limit = 10, len(allFiles) = 5

if 10 > 5 {
	limit = 5
}
limit = 5 - 5 = 0

So if the limit is larger than the total number of results we would end up with limit=0? Does limit=0 actually mean "no limit"?

[komalali]

Ah, I get it after looking at the code that follows.

[cnunciato]

Does this also (kinda) fix #6168?

[komalali]

@cnunciato it sure (kinda) does, thanks for bringing it up!