Python library for querying metrics from several time-series databases into Pandas DataFrames.
It support two types of metric queries, the first is instant metric,
returning the value in precise moment in time. The second is the range
metric, giving you the series of values for given time range and step.
Install the required dependencies on Debian based systems.
apt-get -y install librrd-dev libpython-devInstall library from pip package.
pip install libmetricInstall library from source.
git clone https://github.com/cznewt/python-libmetric.git
cd python-libmetric
python setup.py installParameters can be either set by environmental parameters or passed as command arguments.
For example passing the parameters as environmental parameters.
export LIBMETRIC_ENGINE='prometheus'
export LIBMETRIC_URL='https://metric01:9090'
export LIBMETRIC_QUERY='alertmanager_notifications_total'
export LIBMETRIC_START='2017-11-12T00:00:00Z'
export LIBMETRIC_END='2017-11-16T00:00:00Z'
export LIBMETRIC_STEP='3600s'
range_metricAnd the example of passing parameters as command arguments.
range_metric --engine prometheus --url 'https://metric01:9090' --query '...'- LIBMETRIC_ENGINE
- Type of the endpoint to make query.
- LIBMETRIC_URL
- URL of the endpoint service.
- LIBMETRIC_PARTITION
- Data partition on target service endopoint.
- LIBMETRIC_QUERY
- Query to get the metric time-series or value.
Parameters that apply only for the range meters.
- LIBMETRIC_START
- Time range start.
- LIBMETRIC_END
- Time range end.
- LIBMETRIC_STEP
- Query resolution step width.
Parameters that apply only for the intant meters.
- LIBMETRIC_MOMENT
- Single moment in time.
Parameters that apply only for the all meters/alarms. Except the
LIBMETRIC_AGGREGATION is applicable only for range meters.
- LIBMETRIC_ALARM_THRESHOLD
- Threshold for the alarms.
- LIBMETRIC_ALARM_OPERATOR
- Arithmetic operator for alarm evaluation. [gt, lt, gte, lte, eq]
- LIBMETRIC_AGGREGATION
- Aggregation function for the given time-series [min, max, sum, avg]
The libmetric supports several major time-series databases to get the
results in normalised way. The endpoints are queried thru HTTP API calls.
Example configuration to query the Graphite server.
export LIBMETRIC_ENGINE='graphite'
export LIBMETRIC_URL='http://graphite.host:80'
export LIBMETRIC_QUERY='averageSeries(server.web*.load)'
...Example configuration to query the InfluxDb server.
export LIBMETRIC_ENGINE='influxdb'
export LIBMETRIC_URL='http://influxdb.host:8086'
export LIBMETRIC_USER='user'
export LIBMETRIC_PASSWORD='password'
export LIBMETRIC_PARTITION='prometheus'
export LIBMETRIC_QUERY='SELECT mean("value") FROM "alertmanager_notifications_total"'
...Example configuration to query the Prometheus server.
export LIBMETRIC_ENGINE='prometheus'
export LIBMETRIC_URL='https://prometheus.host:9090'
export LIBMETRIC_QUERY='alertmanager_notifications_total'
...Example configuration to query the RRD file. The query is the consolidation
function and the partition is the data set.
pip install python-rrdtool
export LIBMETRIC_ENGINE='rrd'
export LIBMETRIC_URL='file:///tmp/port.rrd'
export LIBMETRIC_PARTITION='INOCTETS'
export LIBMETRIC_QUERY='AVERAGE'
...from libmetric.query_range import PrometheusRangeQuery
query = 'cpu{metric="load"}'
url = 'http://localhost:9090'
data = {
'queries': [query],
'url': url,
'step': step,
'start': start,
'end': end
}
query = PrometheusRangeQuery(**data)
print query.get()Following lists show allowed values for alarm functions, the alarm arithmetic
operators and aggregation function for range meters.
- gt
- Greater than
>. - gte
- Greater or equal than
>=. - lt
- Lower than
<. - lte
- Lower or equal than
<=. - eq
- Equals to
==.
- avg
- Arithmetic average of the series values.
- min
- Use the minimal value from series.
- max
- Use the maximal value from series.
- sum
- Sum the values together.