server: implement delta profiling for allocs/heap/block/mutex profiles

[tbg]

Previously, when requesting profiles like allocs, heap, block, or mutex through the pprof UI endpoint (/debug/pprof/ui/allocs/?seconds=10), the seconds parameter was ignored and an immediate snapshot was returned. This differed from the behavior of Go's standard pprof handler at /debug/pprof/allocs?seconds=10, which correctly collected a delta profile over the specified duration.

The issue was in profileLocal() which handled these profile types in the default case without respecting req.Seconds. Now, when Seconds > 0 is specified, we collect two profile snapshots separated by the requested duration and compute their difference (delta profile) using the same algorithm as Go's net/http/pprof: scale the first profile by -1 and merge with the second.

This enables the pprof UI to show "recent allocations" over a specific time window rather than cumulative allocations since process start.

Release note (bug fix): Fixed a bug where the pprof UI endpoints for allocs, heap, block, and mutex profiles ignored the seconds parameter and returned immediate snapshots instead of delta profiles.

TODO: fix lint error: server/status_local_file_retrieval.go:165: ts := time.Now().UnixNano() <- forbidden; use 'timeutil' instead

Epic: 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