CM API Time-series

Provide the aggregated timeseries for a given query in CSV format

Requirements

Python 2.7+
CM API Python Client

Install

The CM API Python Client is a dependency and must be installed via:

git clone git://github.com/cloudera/cm_api.git
cd cm_api/python
python setup.py install

The CM API Time-series script can then be installed via:

git clone git@github.com:ggear/cm-api-timeseries.git
cd cm-api-timeseries

Running

Example usages:

./timeseries.py
./timeseries.py --help
./timeseries.py "select cpu_percent"
./timeseries.py "select total_cpu_user where roleType=DATANODE"
./timeseries.py "select (waiting_maps + waiting_reduces) where roleType=JOBTRACKER"
./timeseries.py --from_time 2013-08-29T10:00 --to_time 2013-08-29T16:00 "select (waiting_maps + waiting_reduces) where roleType=JOBTRACKER"
./timeseries.py --host my.cm.server.com --port 80 --version 4 --user my-user --password my-user-password "select (waiting_maps + waiting_reduces) where roleType=JOBTRACKER"

For details on how to define valid queries, see the CM tsquery language reference.

The console output is structed as a CSV and can therefore be piped to a file for looking at using a CSV reader (eg Excel, OpenOffice etc) and further analysis.