From 5b1ad55184f668223a5f709262fe2c1c52258b21 Mon Sep 17 00:00:00 2001 From: Navid Sassan Date: Wed, 20 Sep 2023 12:21:59 +0200 Subject: [PATCH] add influxdb-remove-old-measurements tool --- .../README.rst | 79 ++++++++++ .../influxdb-remove-old-measurements | 149 ++++++++++++++++++ 2 files changed, 228 insertions(+) create mode 100644 tools/influxdb-remove-old-measurements/README.rst create mode 100755 tools/influxdb-remove-old-measurements/influxdb-remove-old-measurements diff --git a/tools/influxdb-remove-old-measurements/README.rst b/tools/influxdb-remove-old-measurements/README.rst new file mode 100644 index 00000000..8c36c5f1 --- /dev/null +++ b/tools/influxdb-remove-old-measurements/README.rst @@ -0,0 +1,79 @@ +Tool influxdb-remove-old-measurements +===================================== + +Overview +-------- + +This tool removes old measurements in InfluxDB, therefore to reducing InfluxDB disk usage. +Measurements where the latest entry is older than the given threshold will be deleted (per host). + +It is especially useful when monitoring systems do not automatically remove measurements when hosts or services are deleted. + + +Fact Sheet +---------- + +.. csv-table:: + :widths: 30, 70 + + "Tool Download", "https://github.com/Linuxfabrik/monitoring-plugins/tree/main/tools/influxdb-remove-old-measurements" + "Can be called without parameters", "No" + "3rd Party Python modules", "``influxdb``" + + +Help +---- + +.. code-block:: text + + usage: influxdb-remove-old-measurements [-h] [-V] --database DATABASE + [--dry-run] [--hostname HOSTNAME] + [--password PASSWORD] [--port PORT] + [--threshold THRESHOLD] + [--username USERNAME] + + This tool removes old measurements in InfluxDB, therefore to reducing InfluxDB + disk usage. Measurements where the latest entry is older than the given + threshold will be deleted (per host). + + optional arguments: + -h, --help show this help message and exit + -V, --version show program's version number and exit + --database DATABASE InfluxDB Database. + --dry-run Perform a trial run with no changes made. + --hostname HOSTNAME InfluxDB Hostname. Default: localhost. + --password PASSWORD InfluxDB Password. + --port PORT InfluxDB Port. + --threshold THRESHOLD + Threshold in days. + --username USERNAME InfluxDB Username. + + +Usage Examples +-------------- + +.. code-block:: bash + + ./influxdb-remove-old-measurements --database icinga2 --username influxdb-user --password linuxfabrik + +Output: + +.. code-block:: text + + Deleting "cmd-check-ping" for host "mon01" (last entry 3M 1W ago) + Deleting "cmd-check-postfix-version" for host "web01" (last entry 3M 1W ago) + Deleting "cmd-check-procs" for host "web01" (last entry 3M 1W ago) + Deleting "cmd-check-swap-usage" for host "cloud01" (last entry 3M 1W ago) + Deleting "cmd-check-swap-usage" for host "web01" (last entry 3M 1W ago) + Deleting "cmd-check-systemd-units-failed" for host "cloud01" (last entry 3M 1W ago) + Deleting "cmd-check-systemd-units-failed" for host "web01" (last entry 3M 1W ago) + Deleting "cmd-check-uptime" for host "cloud01" (last entry 3M 1W ago) + Deleting "cmd-check-users" for host "cloud01" (last entry 3M 1W ago) + Deleting "cmd-check-users" for host "web01" (last entry 3M 1W ago) + + +Credits, License +---------------- + +* Authors: `Linuxfabrik GmbH, Zurich `_ +* License: The Unlicense, see `LICENSE file `_. diff --git a/tools/influxdb-remove-old-measurements/influxdb-remove-old-measurements b/tools/influxdb-remove-old-measurements/influxdb-remove-old-measurements new file mode 100755 index 00000000..a09e977f --- /dev/null +++ b/tools/influxdb-remove-old-measurements/influxdb-remove-old-measurements @@ -0,0 +1,149 @@ +#!/usr/bin/env python3 +# -*- coding: utf-8; py-indent-offset: 4 -*- +# +# Author: Linuxfabrik GmbH, Zurich, Switzerland +# Contact: info (at) linuxfabrik (dot) ch +# https://www.linuxfabrik.ch/ +# License: The Unlicense, see LICENSE file. + +# https://github.com/Linuxfabrik/monitoring-plugins/blob/main/CONTRIBUTING.rst + +import argparse # pylint: disable=C0413 +import sys +import datetime + +import influxdb + +import lib.args +import lib.base +import lib.human +from lib.globals import STATE_UNKNOWN + +__author__ = 'Linuxfabrik GmbH, Zurich/Switzerland' +__version__ = '2023092001' + +DESCRIPTION = """This tool removes old measurements in InfluxDB, therefore to reducing InfluxDB disk usage. + + Measurements where the latest entry is older than the given threshold will be deleted (per host).""" + +DEFAULT_HOSTNAME = 'localhost' +DEFAULT_PORT = 8086 +DEFAULT_THRESHOLD = 90 # days + +def parse_args(): + """Parse command line arguments using argparse. + """ + parser = argparse.ArgumentParser(description=DESCRIPTION) + + parser.add_argument( + '-V', '--version', + action='version', + version='%(prog)s: v{} by {}'.format(__version__, __author__) + ) + + parser.add_argument( + '--database', + help='InfluxDB Database.', + dest='DATABASE', + required=True, + ) + + parser.add_argument( + '--dry-run', + help='Perform a trial run with no changes made.', + dest='DRY_RUN', + action='store_true', + default=False, + ) + + parser.add_argument( + '--hostname', + help='InfluxDB Hostname. Default: %(default)s.', + dest='HOSTNAME', + default=DEFAULT_HOSTNAME, + ) + + parser.add_argument( + '--password', + help='InfluxDB Password.', + dest='PASSWORD', + ) + + parser.add_argument( + '--port', + help='InfluxDB Port.', + dest='PORT', + default=DEFAULT_PORT, + ) + + parser.add_argument( + '--threshold', + help='Threshold in days.', + dest='THRESHOLD', + default=DEFAULT_THRESHOLD, + ) + + parser.add_argument( + '--username', + help='InfluxDB Username.', + dest='USERNAME', + ) + + return parser.parse_args() + + +def main(): + """The main function. Hier spielt die Musik. + """ + + # parse the command line, exit with UNKNOWN if it fails + try: + args = parse_args() + except SystemExit: + sys.exit(STATE_UNKNOWN) + + client = influxdb.InfluxDBClient( + host=args.HOSTNAME, + port=args.PORT, + username=args.USERNAME, + password=args.PASSWORD, + database=args.DATABASE, + ) + + try: + client.ping() + except: + lib.base.cu('Failed to connect to InfluxDB.') + + threshold = datetime.timedelta(days=args.THRESHOLD) + + for measurement in client.get_list_measurements(): + measurement = measurement["name"] + + hosts = client.query('SELECT * FROM "{}" GROUP BY "hostname" ORDER BY time DESC LIMIT 1'.format(measurement)) + + for key, generator in hosts.items(): + hostname = key[1]['hostname'] + # remove trailing Z and floating points if present + timestamp = next(generator)['time'][:-1].split('.')[0] + time = datetime.datetime.strptime(timestamp, '%Y-%m-%dT%H:%M:%S') + + delta = datetime.datetime.now() - time + + if delta > threshold: + print('Deleting "{}" for host "{}" (last entry {} ago)'.format( + measurement, + hostname, + lib.human.seconds2human(delta.total_seconds()), + )) + + if not args.DRY_RUN: + client.delete_series( + measurement=measurement, + tags={ + "hostname": hostname, + }, + ) + +if __name__ == '__main__': + main()