outliers: collect statement fingerprint id

[matthewtodd]

This change also helps set us up for #79451, where we'll be working with
per-fingerprint statement latencies.

Release note: None

[cockroach-teamcity]

This change is 

[Azhng]

Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @Azhng and @matthewtodd)

---

pkg/sql/crdb_internal.go line 6089 at r1 (raw file):

	transaction_id           UUID NOT NULL,
	statement_id             STRING NOT NULL,
	statement_fingerprint_id STRING NOT NULL

Hmm are we sure we want to have a string representation of the fingerprints ID? At least in SQL CLI, we haven't really committed to a pretty format of the fingerprint. Also encoding it here in the virtual table handler will make joining to the exsting stmt/txn stats table difficult.

---

pkg/sql/sqlstats/outliers/outliers.proto line 30 at r1 (raw file):

    bytes id = 1 [(gogoproto.customname) = "ID"];
    uint64 fingerprint_id = 3 [(gogoproto.customname) = "FingerprintID",
      (gogoproto.casttype) = "github.com/cockroachdb/cockroach/pkg/roachpb.StmtFingerprintID"];

nit: maybe this can be below the latency_in_seconds since it's tag is higher :P

[matthewtodd]

Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @Azhng)

---

pkg/sql/crdb_internal.go line 6089 at r1 (raw file):

Previously, Azhng (Archer Zhang) wrote…

Hmm are we sure we want to have a string representation of the fingerprints ID? At least in SQL CLI, we haven't really committed to a pretty format of the fingerprint. Also encoding it here in the virtual table handler will make joining to the exsting stmt/txn stats table difficult.

Good call, I was wondering about this and saw the base-10 string encoding elsewhere, but now I see we just use bytes, yes? Will change.

---

pkg/sql/sqlstats/outliers/outliers.proto line 30 at r1 (raw file):

Previously, Azhng (Archer Zhang) wrote…

nit: maybe this can be below the latency_in_seconds since it's tag is higher :P

Good question. I was looking to have the ordering here mirror the column ordering in the node_execution_outliers table, assuming there's some logical grouping there. (In particular, having these ID fields at the top feels good.) Maybe I just renumber them since nobody cares yet?

[Azhng]

Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @Azhng)

[matthewtodd]

bors r+

[craig]

Build succeeded:

• GitHub CI (Cockroach)