Read exclude_events_log value from policy when generating a diagnostic

[jlind23]

According to the tests @lucabelluccini performed in this pull request it seems that the exclude_events_log value sent down from the policy is not being looked at by the Elastic Agent internal.

[elasticmachine]

Pinging @elastic/elastic-agent-control-plane (Team:Elastic-Agent-Control-Plane)

[cmacknz]

I see it passed into ZipArchive

https://github.com/elastic/elastic-agent/blob/434986ed9fa6fbed6257557a4c6773d519742f66/internal/pkg/agent/application/actions/handlers/handler_action_diagnostics.go#L154
https://github.com/elastic/elastic-agent/blob/434986ed9fa6fbed6257557a4c6773d519742f66/internal/pkg/agent/application/actions/handlers/handler_action_diagnostics.go#L353

This is the exact same function call used by the elastic-agent diagnostics command: https://github.com/elastic/elastic-agent/blob/434986ed9fa6fbed6257557a4c6773d519742f66/internal/pkg/agent/cmd/diagnostics.go#L74

I think the actual problem is it isn't in the Fleet server OpenAPI spec and gets stripped when Fleet server serializes the action for agent.

fleet-server/model/openapi.yml

Lines 555 to 565 in 4f9fdc0

	actionRequestDiagnostics:
	description: The REQUEST_DIAGNOSTICS action data.
	properties:
	additional_metrics:
	description: list optional additional metrics.
	type: array
	items:
	type: string
	enum:
	- CPU
	- CONN