Since cockroachdb#106750, ProposalData is being copied on superseding reproposals. The
caller only knows about the original proposal, and only detaches its context /
tracing span when abandoning the request (e.g. on timeout). By the time of
abandoning the request, there might have been a few superseding reproposals,
and the context / tracing span is being used by multiple ProposalData objects.

In addition to rewriting the original proposal.ctx, we should do the same for
all the ProposalData objects corresponding to the same request. This commit
adds links between consecutive reproposals, and walks the resulting linked list
of proposals when cleaning up / abandoning the request.

Epic: CRDB-25287
Release note: none

Since cockroachdb#106750, ProposalData is being copied on superseding reproposals. The
caller only knows about the original proposal, and only detaches its context /
tracing span when abandoning the request (e.g. on timeout). By the time of
abandoning the request, there might have been a few superseding reproposals,
and the context / tracing span is being used by multiple ProposalData objects.

In addition to rewriting the original proposal.ctx, we should do the same for
all the ProposalData objects corresponding to the same request. This commit
unbinds the context for older proposals when reproposing them, and unbinds the
latest reproposal context when cleaning up / abandoning the request.

Epic: CRDB-25287
Release note: none

Since cockroachdb#106750, ProposalData is being copied on superseding reproposals. The
caller only knows about the original proposal, and only detaches its context /
tracing span when abandoning the request (e.g. on timeout). By the time of
abandoning the request, there might have been a few superseding reproposals,
and the context / tracing span is being used by multiple ProposalData objects.

In addition to rewriting the original proposal.ctx, we should do the same for
all the ProposalData objects corresponding to the same request. This commit
unbinds the context for older proposals when reproposing them, and unbinds the
latest reproposal context when cleaning up / abandoning the request.

Epic: CRDB-25287
Release note: none

108483: roachtest: wait smaller duration before timining out lease prefs r=erikgrinaker a=kvoli

Previously, the `lease-preferences` roachtests would fail on the test
tiemout, despite the test actually failing much earlier. 5 minutes
stopping a node, the node is considered dead, and up-replication will
begin. If the cluster doesn't satisfy lease preferences within that 5
minute window, the test will reliably fail -- making the remainder of
the test timeout (30 or 60 minutes), wasted time.

Fail earlier, by adhering to a post event timeout. This timeout set to
10 minutes.

Informs: #108425
Epic: none
Release note: None

108775: kvserver: fix double span Finish on reproposals r=erikgrinaker a=pavelkalinnikov

Since #106750, `ProposalData` is being copied on superseding reproposals. The caller only knows about the original proposal, and only detaches its context / tracing span when abandoning the request (e.g. on timeout). By the time of abandoning the request, there might have been a few superseding reproposals, and the context / tracing span is being used by multiple `ProposalData` objects.

In addition to rewriting the original `proposal.ctx`, we should do the same for all the `ProposalData` objects corresponding to the same request. This commit unbinds the context for older proposals when reproposing them, and unbinds the latest reproposal context when cleaning up / abandoning the request.

Fixes #107521
Fixes #108534
Fixes #108241
Fixes #108663
Fixes #108696
Fixes #108837

Epic: CRDB-25287
Release note: none

108942: backupccl: deflake memory monitor restore test r=rhu713 a=rhu713

Deflake the TestRestoreMemoryMonitoring tests by only asserting the lower bound in the number of files produced in the backup to account for elastic CPU limiter.

Fixes: #108239

Release note: None

109040: server: add more info to the combined statement endpoint r=maryliag a=maryliag

Previously, it was hard to know which tables were
used to populate the SQL Activity, making debug for it complicated.
This commit adds extra information to the
`combinedstmts` to help:
- oldestAggregatedTsReturned
- stmtsSourceTable
- txnsSourceTable

Returning a value for `olderAggregatedTsReturned` will also allow us to show proper warning when the older timestamp doesn't match the start date of the selected time period on the Search Criteria.

Part Of #108989

Release note: None

Co-authored-by: Austen McClernon <austen@cockroachlabs.com>
Co-authored-by: Pavel Kalinnikov <pavel@cockroachlabs.com>
Co-authored-by: Rui Hu <rui@cockroachlabs.com>
Co-authored-by: maryliag <marylia@cockroachlabs.com>