Calculate queueing delay with milliseconds

What does this MR do and why?

Addressing Wait time to pick a job is always an integer value (#427433 - closed)

Schema changes will be applied manually similar to Add soft delete option to ClickHouse events table (!133579 - merged)

Note that, support for milliseconds was added in https://github.com/ClickHouse/ClickHouse/pull/51291 which landed on 23.7 https://clickhouse.com/docs/en/whats-new/changelog#-clickhouse-release-237-2023-07-27

ClickHouse Cloud instances we use run on 23.8

Screenshots or screen recordings

How to set up and validate locally

1. Update your clickhouse: (for macs just execute this line in the clickhouse directory https://clickhouse.com/docs/en/install#quick-install, it will replace the binary).
2. https://docs.gitlab.com/ee/development/database/clickhouse/clickhouse_within_gitlab.html
3. Execute the following query in CH console:

INSERT INTO ci_finished_builds (id, project_id, pipeline_id, status, finished_at, created_at, started_at, queued_at, runner_id, runner_manager_system_xid, runner_run_untagged, runner_type, runner_manager_version, runner_manager_revision, runner_manager_platform, runner_manager_architecture) SELECT
    number AS id,
    toUInt64(1 * pow(1 - (rand() / 4294967295.), -1 / 1.)) AS project_id,
    toUInt64(1 * pow(1 - (rand() / 4294967295.), -1 / 1.)) AS pipeline_id,
    ['success', 'success', 'success', 'success', 'success', 'failed', 'failed', 'cancelled'][1 + (rand() % 8)] AS status,
    toDateTime64(now(), 3) - toIntervalSecond(rand() % (86400 * 30)) AS finished_at,
    (finished_at - toIntervalSecond(-600 * ln(1 - (rand() / 4294967295.)))) - toIntervalMillisecond(-1000 * ln(1 - (rand() / 4294967295.))) AS started_at,
    (started_at - toIntervalSecond(-10 * ln(1 - (rand() / 4294967295.)))) - toIntervalMillisecond(-1000 * ln(1 - (rand() / 4294967295.))) AS queued_at,
    (queued_at - toIntervalSecond(-10 * ln(1 - (rand() / 4294967295.)))) - toIntervalMillisecond(-1000 * ln(1 - (rand() / 4294967295.))) AS created_at,
    rand() % 10000 AS runner_id,
    toString(rand() % 1000) AS runner_manager_system_xid,
    rand() % 2 AS runner_run_untagged,
    (rand() % 3) + 1 AS runner_type,
    '' AS runner_manager_version,
    '' AS runner_manager_revision,
    '' AS runner_manager_platform,
    '' AS runner_manager_architecture
FROM numbers(30000)

MR acceptance checklist

This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.

• I have evaluated the MR acceptance checklist for this MR.