db: refactor disk usage estimate using range annotations

[anish-shanbhag]

This change updates db.EstimateDiskUsage to use range annotations to
estimate the disk usage of a key range. This should improve the
performance of repeated disk usage estimates for similar or identical
key ranges.

At the Cockroach layer we use db.EstimateDiskUsage in a few places,
most notably when computing MVCC span stats.

Informs: #3793

[cockroach-teamcity]

This change is 

[RaduBerinde]

Nice!

Reviewable status: 0 of 4 files reviewed, 2 unresolved discussions (waiting on @anish-shanbhag, @itsbilal, and @jbowens)

---

db.go line 508 at r1 (raw file):

		// should be protected from concurrent access.
		annotators struct {
			totalSizeAnnotator    *manifest.Annotator[uint64]

[nit] Annotator is redundant given that these are inside annotators.

---

internal/manifest/annotator.go line 281 at r1 (raw file):

// for all files within the given Version which are within the range
// defined by bounds.
func (a *Annotator[T]) VersionRangeAnnotation(v *Version, bounds base.UserKeyBounds) *T {

[nit] Why not return T?

[anish-shanbhag]

Reviewable status: 0 of 4 files reviewed, 2 unresolved discussions (waiting on @itsbilal, @jbowens, and @RaduBerinde)

---

db.go line 508 at r1 (raw file):

Previously, RaduBerinde wrote…

[nit] Annotator is redundant given that these are inside annotators.

Thanks, fixed this.

---

internal/manifest/annotator.go line 281 at r1 (raw file):

Previously, RaduBerinde wrote…

[nit] Why not return T?

This is because of the convention I went with in #3760 where annotators always return a pointer type since a pointer is necessary for the internal caching behavior. Some annotators have a T = fileMetadata and in this case we would want to return a *fileMetadata instead of fileMetadata.

[anish-shanbhag]

TFTR!

Reviewable status: 0 of 4 files reviewed, 2 unresolved discussions (waiting on @itsbilal, @jbowens, and @RaduBerinde)