cli: debug zip truncates rangelog and event data at 1000 entries

[erikgrinaker]

When cockroach debug zip retrieves data via the admin APIs, it does not specify a per-request limit:

cockroach/pkg/cli/zip_cluster_wide.go

Lines 38 to 69 in 49b57e3

	func makeClusterWideZipRequests(
	admin serverpb.AdminClient, status serverpb.StatusClient,
	) []zipRequest {
	return []zipRequest{
	// NB: we intentionally omit liveness since it's already pulled manually (we
	// act on the output to special case decommissioned nodes).
	{
	fn: func(ctx context.Context) (interface{}, error) {
	return admin.Events(ctx, &serverpb.EventsRequest{})
	},
	pathName: eventsName,
	},
	{
	fn: func(ctx context.Context) (interface{}, error) {
	return admin.RangeLog(ctx, &serverpb.RangeLogRequest{})
	},
	pathName: rangelogName,
	},
	{
	fn: func(ctx context.Context) (interface{}, error) {
	return admin.Settings(ctx, &serverpb.SettingsRequest{})
	},
	pathName: settingsName,
	},
	{
	fn: func(ctx context.Context) (interface{}, error) {
	return status.ProblemRanges(ctx, &serverpb.ProblemRangesRequest{})
	},
	pathName: problemRangesName,
	},
	}
	}

This causes it to apply a default limit of 1000 entries for the Events and RangeLog endpoints (which ends up in rangelog.json and events.json).

cockroach/pkg/server/admin.go

Line 75 in f085778

defaultAPIEventLimit = 1000

This is problematic, since e.g. the rangelog may often contain older data of interest. debug zip should paginate in order to include older data.

Jira issue: CRDB-11387

[github-actions]

We have marked this issue as stale because it has been inactive for
18 months. If this issue is still relevant, removing the stale label
or adding a comment will keep it active. Otherwise, we'll close it in
10 days to keep the issue queue tidy. Thank you for your contribution
to CockroachDB!