Get Redshift metrics from Cloudwatch

cloudwatch_redshift.py

import bisect
import datetime
import operator

import boto3

client = boto3.client('cloudwatch')

def get_metric(offset, metricname, threshold, oper='ge', stat='Average'):
    if offset >= 64:
        period, scalar = 3600, 60
    elif offset >= 15:
        period, scalar = 300, 5
    else:
        period, scalar = 60, 1

    dt = (datetime.datetime.utcnow().replace(hour=0, minute=0, second=0, microsecond=0)
              - datetime.timedelta(offset))
    response = client.get_metric_statistics(
        Namespace='AWS/Redshift',
        MetricName=metricname,
        Dimensions=[
            {
                'Name': 'ClusterIdentifier',
                'Value': 'sg-data-warehouse-2'
            },
        ],
        StartTime=dt,
        EndTime=dt + datetime.timedelta(1),
        Period=period,
        Statistics=[stat],
    )

    yes, no = 0, 0
    for datapoint in response['Datapoints']:
        if eval("operator.{}(datapoint[stat], threshold)".format(oper)):
            yes += 1
        else:
            no += 1

    return [x * scalar for x in (yes, no)]

def get_quarter_begin():
    today = datetime.datetime.utcnow()

    qbegins = [datetime.datetime(today.year, month, 1,) for month in (1,4,7,10)]

    idx = bisect.bisect(qbegins, today)
    return qbegins[idx-1]

def main():
    trail_days = max(7, (datetime.datetime.utcnow() - get_quarter_begin()).days)

    cpu = []
    disk = []
    for offset in range(1, trail_days + 1):
        cpu.append(get_metric(offset, 'CPUUtilization', 95, 'gt', 'Maximum'))
        disk.append(get_metric(offset, 'PercentageDiskSpaceUsed', 80, 'gt', 'Minimum'))

    msg = """
CPU: {} of {} events were over the 95% threshold.
{:.2f}%
Disk Space: {} of {} events were over the 80% threshold.
{:.2f}%
""".format(
        sum(zip(*cpu)[0]),
        sum([y for x in zip(*cpu) for y in x]),
        100. * float(sum(zip(*cpu)[0])) / sum([y for x in zip(*cpu) for y in x]),
        sum(zip(*disk)[0]),
        sum([y for x in zip(*disk) for y in x]),
        100. * sum(zip(*disk)[0]) / sum([y for x in zip(*disk) for y in x])
    )

    print msg

if __name__ == '__main__':
    main()