admission: enable disk bandwidth bottleneck resource

[sumeerbhola]

Followup to #82898 which created the basic infrastructure, configuration scheme, and did experimentation with regular and elastic kv0 traffic.

• [x] Eliminate the need to provide the disk-name and bandwidth fields for common environments e.g. linux with store on EBS or GCP PD. This will make it possible to enable this by default for v23.1.
• [x] We need some more experiments, with user-facing regular SQL traffic and lower-priority (and elastic) background work like index backfills. This overlaps with #85641 and #85469. Prior experimentation was done over at https://github.com/cockroachdb/cockroach/pull/82813#issuecomment-1183477491.
• [x] https://github.com/cockroachdb/cockroach/issues/121576
• [x] Enable for v24.3

cc: @irfansharif @andrewbaptist

Jira issue: CRDB-18968

Epic: CRDB-37479

[sumeerbhola]

We should remember to also extend the disk bandwidth control to encompass disk writes of sstables due to incoming range snapshots.

[jbowens]

As a part of online restore, we'll need admission control to control downloading of external sstables so that we can download as fast as we can without affecting foreground workload latencies.

[dshjoshi]

@aadityasondhi @nicktrav Is this in progress or done or still in backlog? If it is in progress or done, can we update the status?

[aadityasondhi]

The functionality should be complete as of https://github.com/cockroachdb/cockroach/pull/129005.

The remaining work would be to:

• Enable on internal test clusters (DRT)
• Enable on cloud clusters
• Write public docs for users

The enable part means setting the cluster setting or store config flags for the provisioned bandwidth.