システムテーブルによるジョブのコストとパフォーマンスの監視

この記事では、システムテーブルを使用してアカウント内のジョブのコストとパフォーマンスをモニタリングする方法の例を示します。

これらのクエリは、ジョブ コンピュートとサーバレス コンピュートでのジョブ実行のコストのみを計算します。 SQLウェアハウスおよび万能コンピュートで実行されるジョブは、ジョブとして請求されないため、コスト属性から除外されます。

注：

これらのクエリーでは、現在のワークスペースが展開されているクラウドリージョン以外のワークスペースのレコードは返されません。現在のリージョン以外のワークスペースのジョブコストを監視するには、そのリージョンに展開されているワークスペースでこれらのクエリーを実行してください。

要件

Jobs モニタリングダッシュボード

次のダッシュボードでは、システムテーブルを使用して、 Databricks ジョブと運用の正常性を包括的にモニタリングします。 これには、ジョブのパフォーマンス追跡、障害モニタリング、リソース使用率などの一般的なユースケースが含まれます。

ジョブコスト監視ダッシュボード

ダッシュボードのインポート

  1. ダッシュボードの JSON ファイルを Databricks GitHub リポジトリからダウンロードします。

  2. ダッシュボードをワークスペースにインポートします。 ダッシュボードのインポート手順については、「 ダッシュボード ファイルのインポート」を参照してください。

コスト監視クエリ

ダッシュボードの次のクエリは、ジョブ コスト モニタリング機能を示しています。

Most expensive ジョブ (last 30 days)

このクエリーは、過去 30 日間で最も支出が多かったジョブを特定します。

with list_cost_per_job as (
  SELECT
    t1.workspace_id,
    t1.usage_metadata.job_id,
    COUNT(DISTINCT t1.usage_metadata.job_run_id) as runs,
    SUM(t1.usage_quantity * list_prices.pricing.default) as list_cost,
    first(identity_metadata.run_as, true) as run_as,
    first(t1.custom_tags, true) as custom_tags,
    MAX(t1.usage_end_time) as last_seen_date
  FROM system.billing.usage t1
  INNER JOIN system.billing.list_prices list_prices on
    t1.cloud = list_prices.cloud and
    t1.sku_name = list_prices.sku_name and
    t1.usage_start_time >= list_prices.price_start_time and
    (t1.usage_end_time <= list_prices.price_end_time or list_prices.price_end_time is null)
  WHERE
    t1.billing_origin_product = "JOBS"
    AND t1.usage_date >= CURRENT_DATE() - INTERVAL 30 DAY
  GROUP BY ALL
),
most_recent_jobs as (
  SELECT
    *,
    ROW_NUMBER() OVER(PARTITION BY workspace_id, job_id ORDER BY change_time DESC) as rn
  FROM
    system.lakeflow.jobs QUALIFY rn=1
)
SELECT
    t2.name,
    t1.job_id,
    t1.workspace_id,
    t1.runs,
    t1.run_as,
    SUM(list_cost) as list_cost,
    t1.last_seen_date
FROM list_cost_per_job t1
  LEFT JOIN most_recent_jobs t2 USING (workspace_id, job_id)
GROUP BY ALL
ORDER BY list_cost DESC

最もコストの高いジョブ実行 (過去 30 日間)

このクエリーは、過去 30 日間で最も支出が多かったジョブ実行を特定します。

with list_cost_per_job_run as (
  SELECT
    t1.workspace_id,
    t1.usage_metadata.job_id,
    t1.usage_metadata.job_run_id as run_id,
    SUM(t1.usage_quantity * list_prices.pricing.default) as list_cost,
    first(identity_metadata.run_as, true) as run_as,
    first(t1.custom_tags, true) as custom_tags,
    MAX(t1.usage_end_time) as last_seen_date
  FROM system.billing.usage t1
  INNER JOIN system.billing.list_prices list_prices on
    t1.cloud = list_prices.cloud and
    t1.sku_name = list_prices.sku_name and
    t1.usage_start_time >= list_prices.price_start_time and
    (t1.usage_end_time <= list_prices.price_end_time or list_prices.price_end_time is null)
  WHERE
    t1.billing_origin_product = 'JOBS'
    AND t1.usage_date >= CURRENT_DATE() - INTERVAL 30 DAY
  GROUP BY ALL
),
most_recent_jobs as (
  SELECT
    *,
    ROW_NUMBER() OVER(PARTITION BY workspace_id, job_id ORDER BY change_time DESC) as rn
  FROM
    system.lakeflow.jobs QUALIFY rn=1
)
SELECT
    t1.workspace_id,
    t2.name,
    t1.job_id,
    t1.run_id,
     t1.run_as,
    SUM(list_cost) as list_cost,
    t1.last_seen_date
FROM list_cost_per_job_run t1
  LEFT JOIN most_recent_jobs t2 USING (workspace_id, job_id)
GROUP BY ALL
ORDER BY list_cost DESC

支出傾向分析(7-14日)

このクエリーは、過去 2 週間でリストコストの支出の増加が最も大きかったジョブを特定します。

with job_run_timeline_with_cost as (
  SELECT
    t1.*,
    t1.usage_metadata.job_id as job_id,
    t1.identity_metadata.run_as as run_as,
    t1.usage_quantity * list_prices.pricing.default AS list_cost
  FROM system.billing.usage t1
    INNER JOIN system.billing.list_prices list_prices
      ON
        t1.cloud = list_prices.cloud AND
        t1.sku_name = list_prices.sku_name AND
        t1.usage_start_time >= list_prices.price_start_time AND
        (t1.usage_end_time <= list_prices.price_end_time or list_prices.price_end_time is NULL)
  WHERE
    t1.billing_origin_product = 'JOBS' AND
    t1.usage_date >= CURRENT_DATE() - INTERVAL 14 DAY
),
most_recent_jobs as (
  SELECT
    *,
    ROW_NUMBER() OVER(PARTITION BY workspace_id, job_id ORDER BY change_time DESC) as rn
  FROM
    system.lakeflow.jobs QUALIFY rn=1
)
SELECT
    t2.name
    ,t1.workspace_id
    ,t1.job_id
    ,t1.sku_name
    ,t1.run_as
    ,Last7DaySpend
    ,Last14DaySpend
    ,last7DaySpend - last14DaySpend as Last7DayGrowth
    ,try_divide( (last7DaySpend - last14DaySpend) , last14DaySpend) * 100 AS Last7DayGrowthPct
FROM
  (
    SELECT
      workspace_id,
      job_id,
      run_as,
      sku_name,
      SUM(list_cost) AS spend
      ,SUM(CASE WHEN usage_end_time BETWEEN date_add(current_date(), -8) AND date_add(current_date(), -1) THEN list_cost ELSE 0 END) AS Last7DaySpend
      ,SUM(CASE WHEN usage_end_time BETWEEN date_add(current_date(), -15) AND date_add(current_date(), -8) THEN list_cost ELSE 0 END) AS Last14DaySpend
    FROM job_run_timeline_with_cost
    GROUP BY ALL
  ) t1
  LEFT JOIN most_recent_jobs t2 USING (workspace_id, job_id)
ORDER BY
  Last7DayGrowth DESC
LIMIT 100

運用の正常性クエリ

ここでは、このダッシュボードがジョブのパフォーマンスと信頼性の追跡に役立つ方法をいくつか紹介します。

失敗したジョブの分析

このクエリは、過去 30 日間に失敗した実行の数が多いジョブに関する情報を返します。 ジョブの実行回数、失敗回数、成功率、および失敗した実行のコストを表示できます。

with job_run_timeline_with_cost as (
  SELECT
    t1.*,
    t1.identity_metadata.run_as as run_as,
    t2.job_id,
    t2.run_id,
    t2.result_state,
    t1.usage_quantity * list_prices.pricing.default as list_cost
  FROM system.billing.usage t1
    INNER JOIN system.lakeflow.job_run_timeline t2
      ON
        t1.workspace_id=t2.workspace_id
        AND t1.usage_metadata.job_id = t2.job_id
        AND t1.usage_metadata.job_run_id = t2.run_id
        AND t1.usage_start_time >= date_trunc("Hour", t2.period_start_time)
        AND t1.usage_start_time < date_trunc("Hour", t2.period_end_time) + INTERVAL 1 HOUR
    INNER JOIN system.billing.list_prices list_prices on
      t1.cloud = list_prices.cloud and
      t1.sku_name = list_prices.sku_name and
      t1.usage_start_time >= list_prices.price_start_time and
      (t1.usage_end_time <= list_prices.price_end_time or list_prices.price_end_time is null)
  WHERE
    t1.billing_origin_product = 'JOBS' AND
    t1.usage_date >= CURRENT_DATE() - INTERVAL 30 DAYS
),
cumulative_run_status_cost as (
  SELECT
    workspace_id,
    job_id,
    run_id,
    run_as,
    result_state,
    usage_end_time,
    SUM(list_cost) OVER (ORDER BY workspace_id, job_id, run_id, usage_end_time ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) AS cumulative_cost
  FROM job_run_timeline_with_cost
  ORDER BY workspace_id, job_id, run_id, usage_end_time
),
cost_per_status as (
  SELECT
      workspace_id,
      job_id,
      run_id,
      run_as,
      result_state,
      usage_end_time,
      cumulative_cost - COALESCE(LAG(cumulative_cost) OVER (ORDER BY workspace_id, job_id, run_id, usage_end_time), 0) AS result_state_cost
  FROM cumulative_run_status_cost
  WHERE result_state IS NOT NULL
  ORDER BY workspace_id, job_id, run_id, usage_end_time),
cost_per_status_agg as (
  SELECT
    workspace_id,
    job_id,
    FIRST(run_as, TRUE) as run_as,
    SUM(result_state_cost) as list_cost
  FROM cost_per_status
  WHERE
    result_state IN ('ERROR', 'FAILED', 'TIMED_OUT')
  GROUP BY ALL
),
terminal_statues as (
  SELECT
    workspace_id,
    job_id,
    CASE WHEN result_state IN ('ERROR', 'FAILED', 'TIMED_OUT') THEN 1 ELSE 0 END as is_failure,
    period_end_time as last_seen_date
  FROM system.lakeflow.job_run_timeline
  WHERE
    result_state IS NOT NULL AND
    period_end_time >= CURRENT_DATE() - INTERVAL 30 DAYS
),
most_recent_jobs as (
  SELECT
    *,
    ROW_NUMBER() OVER(PARTITION BY workspace_id, job_id ORDER BY change_time DESC) as rn
  FROM
    system.lakeflow.jobs QUALIFY rn=1
)
SELECT
  first(t2.name) as name,
  t1.workspace_id,
  t1.job_id,
  COUNT(*) as runs,
  t3.run_as,
  SUM(is_failure) as failures,
  (1 - COALESCE(try_divide(SUM(is_failure), COUNT(*)), 0)) * 100 as success_ratio,
  first(t3.list_cost) as failure_list_cost,
  MAX(t1.last_seen_date) as last_seen_date
FROM terminal_statues t1
  LEFT JOIN most_recent_jobs t2 USING (workspace_id, job_id)
  LEFT JOIN cost_per_status_agg t3 USING (workspace_id, job_id)
GROUP BY ALL
ORDER BY failures DESC

再試行パターン

このクエリは、過去 30 日間に頻繁に修理が行われたジョブに関する情報 (修理の数、修理実行のコスト、修復実行の累積期間など) を返します。

with job_run_timeline_with_cost as (
 SELECT
   t1.*,
   t2.job_id,
   t2.run_id,
   t1.identity_metadata.run_as as run_as,
   t2.result_state,
   t1.usage_quantity * list_prices.pricing.default as list_cost
 FROM system.billing.usage t1
   INNER JOIN system.lakeflow.job_run_timeline t2
     ON
       t1.workspace_id=t2.workspace_id
       AND t1.usage_metadata.job_id = t2.job_id
       AND t1.usage_metadata.job_run_id = t2.run_id
       AND t1.usage_start_time >= date_trunc("Hour", t2.period_start_time)
       AND t1.usage_start_time < date_trunc("Hour", t2.period_end_time) + INTERVAL 1 HOUR
   INNER JOIN system.billing.list_prices list_prices on
     t1.cloud = list_prices.cloud and
     t1.sku_name = list_prices.sku_name and
     t1.usage_start_time >= list_prices.price_start_time and
     (t1.usage_end_time <= list_prices.price_end_time or list_prices.price_end_time is null)
 WHERE
   t1.billing_origin_product = 'JOBS' AND
   t1.usage_date >= CURRENT_DATE() - INTERVAL 30 DAYS
),
cumulative_run_status_cost as (
 SELECT
   workspace_id,
   job_id,
   run_id,
   run_as,
   result_state,
   usage_end_time,
   SUM(list_cost) OVER (ORDER BY workspace_id, job_id, run_id, usage_end_time ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) AS cumulative_cost
 FROM job_run_timeline_with_cost
 ORDER BY workspace_id, job_id, run_id, usage_end_time
),
cost_per_status as (
 SELECT
     workspace_id,
     job_id,
     run_id,
     run_as,
     result_state,
     usage_end_time,
     cumulative_cost - COALESCE(LAG(cumulative_cost) OVER (ORDER BY workspace_id, job_id, run_id, usage_end_time), 0) AS result_state_cost
 FROM cumulative_run_status_cost
 WHERE result_state IS NOT NULL
 ORDER BY workspace_id, job_id, run_id, usage_end_time),
cost_per_unsuccesful_status_agg as (
 SELECT
   workspace_id,
   job_id,
   run_id,
   first(run_as, TRUE) as run_as,
   SUM(result_state_cost) as list_cost
 FROM cost_per_status
 WHERE
   result_state != "SUCCEEDED"
 GROUP BY ALL
),
repaired_runs as (
 SELECT
   workspace_id, job_id, run_id, COUNT(*) as cnt
 FROM system.lakeflow.job_run_timeline
 WHERE result_state IS NOT NULL
 GROUP BY ALL
 HAVING cnt > 1
),
successful_repairs as (
 SELECT t1.workspace_id, t1.job_id, t1.run_id, MAX(t1.period_end_time) as period_end_time
 FROM system.lakeflow.job_run_timeline t1
 JOIN repaired_runs t2
 ON t1.workspace_id=t2.workspace_id AND t1.job_id=t2.job_id AND t1.run_id=t2.run_id
 WHERE t1.result_state="SUCCEEDED"
 GROUP BY ALL
),
combined_repairs as (
 SELECT
   t1.*,
   t2.period_end_time,
   t1.cnt as repairs
 FROM repaired_runs t1
   LEFT JOIN successful_repairs t2 USING (workspace_id, job_id, run_id)
),
most_recent_jobs as (
 SELECT
   *,
   ROW_NUMBER() OVER(PARTITION BY workspace_id, job_id ORDER BY change_time DESC) as rn
 FROM
   system.lakeflow.jobs QUALIFY rn=1
)
SELECT
 last(t3.name) as name,
 t1.workspace_id,
 t1.job_id,
 t1.run_id,
 first(t4.run_as, TRUE) as run_as,
 first(t1.repairs) - 1 as repairs,
 first(t4.list_cost) as repair_list_cost,
 CASE WHEN t1.period_end_time IS NOT NULL THEN CAST(t1.period_end_time - MIN(t2.period_end_time) as LONG) ELSE NULL END AS repair_time_seconds
FROM combined_repairs t1
 JOIN system.lakeflow.job_run_timeline t2 USING (workspace_id, job_id, run_id)
 LEFT JOIN most_recent_jobs t3 USING (workspace_id, job_id)
 LEFT JOIN cost_per_unsuccesful_status_agg t4 USING (workspace_id, job_id, run_id)
WHERE
 t2.result_state IS NOT NULL
GROUP BY t1.workspace_id, t1.job_id, t1.run_id, t1.period_end_time
ORDER BY repairs DESC