server: return elapsed time for active executions

[xinhaoz]

Previously, we calculated the time elapsed for an active stmt or txn based on the start time returned from the server and the time the response was last received. Calculating this value on the client is not reliable and can lead to negative values when the server time is slightly ahead. This commit fixes this issue by including the time elapsed as part of the active txns and stmts response.

Release note (bug fix): time elapsed for active txns and stmts is never negative.

[cockroach-teamcity]

This change is 

[maryliag]

you added the label to backport to 22.1, but this feature didn't exist on 22.1

Reviewed 12 of 12 files at r1, all commit messages.
Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @xinhaoz)

[matthewtodd]

Reviewable status: complete! 0 of 0 LGTMs obtained (and 1 stale) (waiting on @maryliag and @xinhaoz)

---

pkg/ui/workspaces/cluster-ui/src/activeExecutions/execTableCommon.tsx line 191 at r2 (raw file):

      title: executionsTableTitles.elapsedTime(execType),
      cell: (item: ActiveExecution) =>
        Duration(item.elapsedTime.asMilliseconds() * 1e6),

Hmm, I see it was like this before, but is it a bug that we multiply milliseconds by 1,000,000? I'd expect to see 1,000.

[xinhaoz]

pkg/ui/workspaces/cluster-ui/src/activeExecutions/execTableCommon.tsx line 191 at r2 (raw file):

Previously, matthewtodd (Matthew Todd) wrote…

Hmm, I see it was like this before, but is it a bug that we multiply milliseconds by 1,000,000? I'd expect to see 1,000.

Yep, I want nanoseconds here since that is what the Duration function expects so that's why I used 1e6. There's no asNanoseconds function for durations unfortunately :(

[xinhaoz]

Test failure unrelated.
bors r+

[craig]

Build succeeded:

• Bazel Essential CI (Cockroach)

[matthewtodd]

pkg/ui/workspaces/cluster-ui/src/activeExecutions/execTableCommon.tsx line 191 at r2 (raw file):

Previously, xinhaoz (Xin Hao Zhang) wrote…

Yep, I want nanoseconds here since that is what the Duration function expects so that's why I used 1e6. There's no asNanoseconds function for durations unfortunately :(

Oh, right, I was thinking backwards. Thank you!