storagebase: add CheckIngestedStats helper, call after all bulk-ingest ops

[dt]

This helper is intended to be called on the span(s) in which a caller
has been doing bulk-ingestion, to trigger consistency checks on the
range(s) in that span. This might be a good idea on its own but also has
the side effect of fixing up any MVCC stats inaccuracies that might
have been introduced by estimates during ingestion.

It is probably not critical to actually call this: eventually the
consistency queue will call anyway and fixup those stats if needed.
However, if we know we'll call it right away -- specifically before we
mark the operation that was ingesting as completed and turn those ranges
over to real traffic that might expect correct-ish stats -- we might be
able to make coarser, cheaper estimates, for instance during SST
ingestion where the current, accurate stats recomputation is expensive.

See #35231.

Release note: None

[cockroach-teamcity]

This change is 

[nvb]

Reviewed 7 of 7 files at r1.
Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @danhhz and @dt)

---

pkg/ccl/importccl/read_import_proc.go, line 710 at r1 (raw file):

	// If somehow we only ingested one key...
	if ingestSpan.Key.Equal(ingestSpan.EndKey) {

How will this happen? We should never have an ingestSpan where Key == EndKey.

---

pkg/ccl/importccl/sst_writer_proc.go, line 195 at r1 (raw file):

							log.Errorf(ctx, "failed to scatter span %s: %s", roachpb.PrettyPrintKey(nil, end), pErr)
						}
						if k, cur := sst.span.Key, ingestSpan.Key; cur == nil || k.Compare(cur) < 0 {

Here and above: can we use Span.Combine?

---

pkg/storage/storagebase/bulk_adder.go, line 50 at r1 (raw file):

		return nil
	}
	if ingested.Key.Equal(ingested.EndKey) {

Again, is this possible?

[dt]

Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @danhhz, @dt, and @nvanbenschoten)

---

pkg/ccl/importccl/read_import_proc.go, line 710 at r1 (raw file):

Previously, nvanbenschoten (Nathan VanBenschoten) wrote…

How will this happen? We should never have an ingestSpan where Key == EndKey.

I thought I answered that in the comment on the check, but if the ingestion job ingested exactly one key (e.g. IMPORT of a single row)?

[nvb]

Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @danhhz and @dt)

---

pkg/ccl/importccl/read_import_proc.go, line 710 at r1 (raw file):

Previously, dt (David Taylor) wrote…

I thought I answered that in the comment on the check, but if the ingestion job ingested exactly one key (e.g. IMPORT of a single row)?

Then shouldn't EndKey be either nil or Key.Next already? Span.EndKey should always be exclusive.

[dt]

Closing this in favor of a core-driven change here, since it sounds like there're plans for how to actually clear ContainsEstimates which this doesn't do.