kvserver: do lazy map allocations in replicaFlowControl

[sumeerbhola]

These methods will typically return empty maps, so we can avoid the allocations.

Informs #111561

Epic: none

Release note: None

[blathers-crl]

It looks like your PR touches production code but doesn't add or edit any test code. Did you consider adding tests to your PR?

_{🦉 Hoot! I am a Blathers, a bot for CockroachDB. My owner is dev-inf.}

[cockroach-teamcity]

This change is 

[pav-kv]

Nice simple fix, thanks.

Could also store the maps in replicaFlowControl, and reuse them across the runs. Or, to avoid the sticky overhead, draw them from a sync.Pool and release after use. These strategies are less simple and possibly unnecessary. Feel free to leave a TODO though.

[sumeerbhola]

TFTR!

Could also store the maps in replicaFlowControl, and reuse them across the runs. Or, to avoid the sticky overhead, draw them from a sync.Pool and release after use. These strategies are less simple and possibly unnecessary. Feel free to leave a TODO though.

I've left a comment in replicaFlowControlIntegrationImpl.notActivelyReplicatingTo for one of these possible optimizations. I didn't frame it as a TODO since I suspect we may never need to do it.

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

---

pkg/kv/kvserver/flow_control_replica.go line 63 at r1 (raw file):

Previously, pavelkalinnikov (Pavel Kalinnikov) wrote…

Could do with a comment why initialization is lazy here. Same below.

Done

[sumeerbhola]

test failure is unrelated #113112

[aadityasondhi]

thanks for the quick fix!

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

[sumeerbhola]

bors r=pavelkalinnikov,aadityasondhi

[craig]

Build succeeded:

• Bazel Essential CI (Cockroach)