可以通过使用Bigquery API发出警报来监视Bigquery查询槽数量和查询并发量。

以下是如何设置监视查询槽数量的示例代码：

import google.auth
from google.cloud import bigquery

def check_slots():
    _, proj_id = google.auth.default()
    client = bigquery.Client(project=proj_id)
    
    query_job_config = bigquery.QueryJobConfig(dry_run=True, use_query_cache=False)
    query_job_config.dry_run = True
    query_job_config.use_query_cache = False
    query = (
        "SELECT CAST(IFNULL(sum(total_slot_ms), 0) / 1000 / 60 / 60 as INT64) AS used_hours "
        "FROM `region-us`.INFORMATION_SCHEMA.JOBS_BY_PROJECT "
        "WHERE REGEXP_CONTAINS(project_id, r'^" + proj_id + "$') "
        "AND creation_time BETWEEN "
        "   TIMESTAMP_SUB(CURRENT_TIMESTAMP(), INTERVAL 3 DAY) "
        "   AND CURRENT_TIMESTAMP() "
        "   AND state = 'DONE';"
    )

    query_job = client.query(query, job_config=query_job_config)
    results = query_job.result()

    for row in results:
        used_hours = row['used_hours']

    if used_hours > 100:
        send_alert('Slots usage is high. Used Hours: {}'.format(used_hours))

def send_alert(msg):
    # send an alert
    pass

以下是如何设置监视查询并发量的示例代码：

import google.auth
from google.cloud import monitoring_v3

def check_concurrency():
    _, proj_id = google.auth.default()
    client = monitoring_v3.MetricServiceClient()

    end_time = time.time()
    start_time = end_time - 3 * 60 * 60 # the past 3 hours
    project_resource = client.project_path(proj_id)

    duration = monitoring_v3.types.TimeInterval()
    duration.end_time.FromSeconds(int(end_time))
    duration.start_time.FromSeconds(int(start_time))

    aggregation = monitoring_v3.types.Aggregation()
    aggregation.alignment_period.seconds = 600 # 10 minutes
    aggregation.per_series_aligner = (monitoring_v3.enums.Aggregation.Aligner.ALIGN_MEAN)

    metric_type = 'bigquery.googleapis.com/query/count'
    filter = (
        'metric.type="{}" AND '
        'resource.type="global" AND '
        'resource.labels.project_id="{}"').format(
        metric_type, proj_id)

    response = client.list_time_series(
        project_resource,
        filter,
        aggregation,
        interval,
    )

    num_queries = 0
    for ts in response:
        for point in ts.points: