diff options
Diffstat (limited to '')
-rw-r--r-- | collectors/python.d.plugin/go_expvar/Makefile.inc | 13 | ||||
-rw-r--r-- | collectors/python.d.plugin/go_expvar/README.md | 319 | ||||
-rw-r--r-- | collectors/python.d.plugin/go_expvar/go_expvar.chart.py | 253 | ||||
-rw-r--r-- | collectors/python.d.plugin/go_expvar/go_expvar.conf | 108 |
4 files changed, 693 insertions, 0 deletions
diff --git a/collectors/python.d.plugin/go_expvar/Makefile.inc b/collectors/python.d.plugin/go_expvar/Makefile.inc new file mode 100644 index 0000000..74f50d7 --- /dev/null +++ b/collectors/python.d.plugin/go_expvar/Makefile.inc @@ -0,0 +1,13 @@ +# SPDX-License-Identifier: GPL-3.0-or-later + +# THIS IS NOT A COMPLETE Makefile +# IT IS INCLUDED BY ITS PARENT'S Makefile.am +# IT IS REQUIRED TO REFERENCE ALL FILES RELATIVE TO THE PARENT + +# install these files +dist_python_DATA += go_expvar/go_expvar.chart.py +dist_pythonconfig_DATA += go_expvar/go_expvar.conf + +# do not install these files, but include them in the distribution +dist_noinst_DATA += go_expvar/README.md go_expvar/Makefile.inc + diff --git a/collectors/python.d.plugin/go_expvar/README.md b/collectors/python.d.plugin/go_expvar/README.md new file mode 100644 index 0000000..a73610e --- /dev/null +++ b/collectors/python.d.plugin/go_expvar/README.md @@ -0,0 +1,319 @@ +<!-- +title: "Go applications monitoring with Netdata" +custom_edit_url: https://github.com/netdata/netdata/edit/master/collectors/python.d.plugin/go_expvar/README.md +sidebar_label: "Go applications" +--> + +# Go applications monitoring with Netdata + +Monitors Go application that exposes its metrics with the use of `expvar` package from the Go standard library. The package produces charts for Go runtime memory statistics and optionally any number of custom charts. + +The `go_expvar` module produces the following charts: + +1. **Heap allocations** in kB + + - alloc: size of objects allocated on the heap + - inuse: size of allocated heap spans + +2. **Stack allocations** in kB + + - inuse: size of allocated stack spans + +3. **MSpan allocations** in kB + + - inuse: size of allocated mspan structures + +4. **MCache allocations** in kB + + - inuse: size of allocated mcache structures + +5. **Virtual memory** in kB + + - sys: size of reserved virtual address space + +6. **Live objects** + + - live: number of live objects in memory + +7. **GC pauses average** in ns + + - avg: average duration of all GC stop-the-world pauses + +## Monitoring Go applications + +Netdata can be used to monitor running Go applications that expose their metrics with +the use of the [expvar package](https://golang.org/pkg/expvar/) included in Go standard library. + +The `expvar` package exposes these metrics over HTTP and is very easy to use. +Consider this minimal sample below: + +```go +package main + +import ( + _ "expvar" + "net/http" +) + +func main() { + http.ListenAndServe("127.0.0.1:8080", nil) +} +``` + +When imported this way, the `expvar` package registers a HTTP handler at `/debug/vars` that +exposes Go runtime's memory statistics in JSON format. You can inspect the output by opening +the URL in your browser (or by using `wget` or `curl`). + +Sample output: + +```json +{ +"cmdline": ["./expvar-demo-binary"], +"memstats": {"Alloc":630856,"TotalAlloc":630856,"Sys":3346432,"Lookups":27, <omitted for brevity>} +} +``` + +You can of course expose and monitor your own variables as well. +Here is a sample Go application that exposes a few custom variables: + +```go +package main + +import ( + "expvar" + "net/http" + "runtime" + "time" +) + +func main() { + + tick := time.NewTicker(1 * time.Second) + num_go := expvar.NewInt("runtime.goroutines") + counters := expvar.NewMap("counters") + counters.Set("cnt1", new(expvar.Int)) + counters.Set("cnt2", new(expvar.Float)) + + go http.ListenAndServe(":8080", nil) + + for { + select { + case <- tick.C: + num_go.Set(int64(runtime.NumGoroutine())) + counters.Add("cnt1", 1) + counters.AddFloat("cnt2", 1.452) + } + } +} +``` + +Apart from the runtime memory stats, this application publishes two counters and the +number of currently running Goroutines and updates these stats every second. + +In the next section, we will cover how to monitor and chart these exposed stats with +the use of `netdata`s `go_expvar` module. + +### Using Netdata go_expvar module + +The `go_expvar` module is disabled by default. To enable it, edit `python.d.conf` (to edit it on your system run +`/etc/netdata/edit-config python.d.conf`), and change the `go_expvar` variable to `yes`: + +``` +# Enable / Disable python.d.plugin modules +#default_run: yes +# +# If "default_run" = "yes" the default for all modules is enabled (yes). +# Setting any of these to "no" will disable it. +# +# If "default_run" = "no" the default for all modules is disabled (no). +# Setting any of these to "yes" will enable it. +... +go_expvar: yes +... +``` + +Next, we need to edit the module configuration file (found at `/etc/netdata/python.d/go_expvar.conf` by default) (to +edit it on your system run `/etc/netdata/edit-config python.d/go_expvar.conf`). The module configuration consists of +jobs, where each job can be used to monitor a separate Go application. Let's see a sample job configuration: + +``` +# /etc/netdata/python.d/go_expvar.conf + +app1: + name : 'app1' + url : 'http://127.0.0.1:8080/debug/vars' + collect_memstats: true + extra_charts: {} +``` + +Let's go over each of the defined options: + +``` +name: 'app1' +``` + +This is the job name that will appear at the Netdata dashboard. +If not defined, the job_name (top level key) will be used. + +``` +url: 'http://127.0.0.1:8080/debug/vars' +``` + +This is the URL of the expvar endpoint. As the expvar handler can be installed +in a custom path, the whole URL has to be specified. This value is mandatory. + +``` +collect_memstats: true +``` + +Whether to enable collecting stats about Go runtime's memory. You can find more +information about the exposed values at the [runtime package docs](https://golang.org/pkg/runtime/#MemStats). + +``` +extra_charts: {} +``` + +Enables the user to specify custom expvars to monitor and chart. +Will be explained in more detail below. + +**Note: if `collect_memstats` is disabled and no `extra_charts` are defined, the plugin will +disable itself, as there will be no data to collect!** + +Apart from these options, each job supports options inherited from Netdata's `python.d.plugin` +and its base `UrlService` class. These are: + +``` +update_every: 1 # the job's data collection frequency +priority: 60000 # the job's order on the dashboard +user: admin # use when the expvar endpoint is protected by HTTP Basic Auth +password: sekret # use when the expvar endpoint is protected by HTTP Basic Auth +``` + +### Monitoring custom vars with go_expvar + +Now, memory stats might be useful, but what if you want Netdata to monitor some custom values +that your Go application exposes? The `go_expvar` module can do that as well with the use of +the `extra_charts` configuration variable. + +The `extra_charts` variable is a YaML list of Netdata chart definitions. +Each chart definition has the following keys: + +``` +id: Netdata chart ID +options: a key-value mapping of chart options +lines: a list of line definitions +``` + +**Note: please do not use dots in the chart or line ID field. +See [this issue](https://github.com/netdata/netdata/pull/1902#issuecomment-284494195) for explanation.** + +Please see these two links to the official Netdata documentation for more information about the values: + +- [External plugins - charts](/collectors/plugins.d/README.md#chart) +- [Chart variables](/collectors/python.d.plugin/README.md#global-variables-order-and-chart) + +**Line definitions** + +Each chart can define multiple lines (dimensions). +A line definition is a key-value mapping of line options. +Each line can have the following options: + +``` +# mandatory +expvar_key: the name of the expvar as present in the JSON output of /debug/vars endpoint +expvar_type: value type; supported are "float" or "int" +id: the id of this line/dimension in Netdata + +# optional - Netdata defaults are used if these options are not defined +name: '' +algorithm: absolute +multiplier: 1 +divisor: 100 if expvar_type == float, 1 if expvar_type == int +hidden: False +``` + +Please see the following link for more information about the options and their default values: +[External plugins - dimensions](/collectors/plugins.d/README.md#dimension) + +Apart from top-level expvars, this plugin can also parse expvars stored in a multi-level map; +All dicts in the resulting JSON document are then flattened to one level. +Expvar names are joined together with '.' when flattening. + +Example: + +``` +{ + "counters": {"cnt1": 1042, "cnt2": 1512.9839999999983}, + "runtime.goroutines": 5 +} +``` + +In the above case, the exported variables will be available under `runtime.goroutines`, +`counters.cnt1` and `counters.cnt2` expvar_keys. If the flattening results in a key collision, +the first defined key wins and all subsequent keys with the same name are ignored. + +## Enable the collector + +The `go_expvar` collector is disabled by default. To enable it, use `edit-config` from the Netdata [config +directory](/docs/configure/nodes.md), which is typically at `/etc/netdata`, to edit the `python.d.conf` file. + +```bash +cd /etc/netdata # Replace this path with your Netdata config directory, if different +sudo ./edit-config python.d.conf +``` + +Change the value of the `go_expvar` setting to `yes`. Save the file and restart the Netdata Agent with `sudo systemctl +restart netdata`, or the appropriate method for your system, to finish enabling the `go_expvar` collector. + +## Configuration + +Edit the `python.d/go_expvar.conf` configuration file using `edit-config` from the Netdata [config +directory](/docs/configure/nodes.md), which is typically at `/etc/netdata`. + +```bash +cd /etc/netdata # Replace this path with your Netdata config directory, if different +sudo ./edit-config python.d/go_expvar.conf +``` + +The configuration below matches the second Go application described above. +Netdata will monitor and chart memory stats for the application, as well as a custom chart of +running goroutines and two dummy counters. + +``` +app1: + name : 'app1' + url : 'http://127.0.0.1:8080/debug/vars' + collect_memstats: true + extra_charts: + - id: "runtime_goroutines" + options: + name: num_goroutines + title: "runtime: number of goroutines" + units: goroutines + family: runtime + context: expvar.runtime.goroutines + chart_type: line + lines: + - {expvar_key: 'runtime.goroutines', expvar_type: int, id: runtime_goroutines} + - id: "foo_counters" + options: + name: counters + title: "some random counters" + units: awesomeness + family: counters + context: expvar.foo.counters + chart_type: line + lines: + - {expvar_key: 'counters.cnt1', expvar_type: int, id: counters_cnt1} + - {expvar_key: 'counters.cnt2', expvar_type: float, id: counters_cnt2} +``` + +**Netdata charts example** + +The images below show how do the final charts in Netdata look. + +![Memory stats charts](https://cloud.githubusercontent.com/assets/15180106/26762052/62b4af58-493b-11e7-9e69-146705acfc2c.png) + +![Custom charts](https://cloud.githubusercontent.com/assets/15180106/26762051/62ae915e-493b-11e7-8518-bd25a3886650.png) + +[![analytics](https://www.google-analytics.com/collect?v=1&aip=1&t=pageview&_s=1&ds=github&dr=https%3A%2F%2Fgithub.com%2Fnetdata%2Fnetdata&dl=https%3A%2F%2Fmy-netdata.io%2Fgithub%2Fcollectors%2Fpython.d.plugin%2Fgo_expvar%2FREADME&_u=MAC~&cid=5792dfd7-8dc4-476b-af31-da2fdb9f93d2&tid=UA-64295674-3)](<>) diff --git a/collectors/python.d.plugin/go_expvar/go_expvar.chart.py b/collectors/python.d.plugin/go_expvar/go_expvar.chart.py new file mode 100644 index 0000000..f9bbdc1 --- /dev/null +++ b/collectors/python.d.plugin/go_expvar/go_expvar.chart.py @@ -0,0 +1,253 @@ +# -*- coding: utf-8 -*- +# Description: go_expvar netdata python.d module +# Author: Jan Kral (kralewitz) +# SPDX-License-Identifier: GPL-3.0-or-later + +from __future__ import division + +import json +from collections import namedtuple + +from bases.FrameworkServices.UrlService import UrlService + +MEMSTATS_ORDER = [ + 'memstats_heap', + 'memstats_stack', + 'memstats_mspan', + 'memstats_mcache', + 'memstats_sys', + 'memstats_live_objects', + 'memstats_gc_pauses', +] + +MEMSTATS_CHARTS = { + 'memstats_heap': { + 'options': ['heap', 'memory: size of heap memory structures', 'KiB', 'memstats', + 'expvar.memstats.heap', 'line'], + 'lines': [ + ['memstats_heap_alloc', 'alloc', 'absolute', 1, 1024], + ['memstats_heap_inuse', 'inuse', 'absolute', 1, 1024] + ] + }, + 'memstats_stack': { + 'options': ['stack', 'memory: size of stack memory structures', 'KiB', 'memstats', + 'expvar.memstats.stack', 'line'], + 'lines': [ + ['memstats_stack_inuse', 'inuse', 'absolute', 1, 1024] + ] + }, + 'memstats_mspan': { + 'options': ['mspan', 'memory: size of mspan memory structures', 'KiB', 'memstats', + 'expvar.memstats.mspan', 'line'], + 'lines': [ + ['memstats_mspan_inuse', 'inuse', 'absolute', 1, 1024] + ] + }, + 'memstats_mcache': { + 'options': ['mcache', 'memory: size of mcache memory structures', 'KiB', 'memstats', + 'expvar.memstats.mcache', 'line'], + 'lines': [ + ['memstats_mcache_inuse', 'inuse', 'absolute', 1, 1024] + ] + }, + 'memstats_live_objects': { + 'options': ['live_objects', 'memory: number of live objects', 'objects', 'memstats', + 'expvar.memstats.live_objects', 'line'], + 'lines': [ + ['memstats_live_objects', 'live'] + ] + }, + 'memstats_sys': { + 'options': ['sys', 'memory: size of reserved virtual address space', 'KiB', 'memstats', + 'expvar.memstats.sys', 'line'], + 'lines': [ + ['memstats_sys', 'sys', 'absolute', 1, 1024] + ] + }, + 'memstats_gc_pauses': { + 'options': ['gc_pauses', 'memory: average duration of GC pauses', 'ns', 'memstats', + 'expvar.memstats.gc_pauses', 'line'], + 'lines': [ + ['memstats_gc_pauses', 'avg'] + ] + } +} + +EXPVAR = namedtuple( + "EXPVAR", + [ + "key", + "type", + "id", + ] +) + + +def flatten(d, top='', sep='.'): + items = [] + for key, val in d.items(): + nkey = top + sep + key if top else key + if isinstance(val, dict): + items.extend(flatten(val, nkey, sep=sep).items()) + else: + items.append((nkey, val)) + return dict(items) + + +class Service(UrlService): + def __init__(self, configuration=None, name=None): + UrlService.__init__(self, configuration=configuration, name=name) + # if memstats collection is enabled, add the charts and their order + if self.configuration.get('collect_memstats'): + self.definitions = dict(MEMSTATS_CHARTS) + self.order = list(MEMSTATS_ORDER) + else: + self.definitions = dict() + self.order = list() + + # if extra charts are defined, parse their config + extra_charts = self.configuration.get('extra_charts') + if extra_charts: + self._parse_extra_charts_config(extra_charts) + + def check(self): + """ + Check if the module can collect data: + 1) At least one JOB configuration has to be specified + 2) The JOB configuration needs to define the URL and either collect_memstats must be enabled or at least one + extra_chart must be defined. + + The configuration and URL check is provided by the UrlService class. + """ + + if not (self.configuration.get('extra_charts') or self.configuration.get('collect_memstats')): + self.error('Memstats collection is disabled and no extra_charts are defined, disabling module.') + return False + + return UrlService.check(self) + + def _parse_extra_charts_config(self, extra_charts_config): + + # a place to store the expvar keys and their types + self.expvars = list() + + for chart in extra_charts_config: + + chart_dict = dict() + chart_id = chart.get('id') + chart_lines = chart.get('lines') + chart_opts = chart.get('options', dict()) + + if not all([chart_id, chart_lines]): + self.info('Chart {0} has no ID or no lines defined, skipping'.format(chart)) + continue + + chart_dict['options'] = [ + chart_opts.get('name', ''), + chart_opts.get('title', ''), + chart_opts.get('units', ''), + chart_opts.get('family', ''), + chart_opts.get('context', ''), + chart_opts.get('chart_type', 'line') + ] + chart_dict['lines'] = list() + + # add the lines to the chart + for line in chart_lines: + + ev_key = line.get('expvar_key') + ev_type = line.get('expvar_type') + line_id = line.get('id') + + if not all([ev_key, ev_type, line_id]): + self.info('Line missing expvar_key, expvar_type, or line_id, skipping: {0}'.format(line)) + continue + + if ev_type not in ['int', 'float']: + self.info('Unsupported expvar_type "{0}". Must be "int" or "float"'.format(ev_type)) + continue + + # self.expvars[ev_key] = (ev_type, line_id) + self.expvars.append(EXPVAR(ev_key, ev_type, line_id)) + + chart_dict['lines'].append( + [ + line.get('id', ''), + line.get('name', ''), + line.get('algorithm', ''), + line.get('multiplier', 1), + line.get('divisor', 100 if ev_type == 'float' else 1), + line.get('hidden', False) + ] + ) + + self.order.append(chart_id) + self.definitions[chart_id] = chart_dict + + def _get_data(self): + """ + Format data received from http request + :return: dict + """ + + raw_data = self._get_raw_data() + if not raw_data: + return None + + data = json.loads(raw_data) + + expvars = dict() + if self.configuration.get('collect_memstats'): + expvars.update(self._parse_memstats(data)) + + if self.configuration.get('extra_charts'): + # the memstats part of the data has been already parsed, so we remove it before flattening and checking + # the rest of the data, thus avoiding needless iterating over the multiply nested memstats dict. + del (data['memstats']) + flattened = flatten(data) + + for ev in self.expvars: + v = flattened.get(ev.key) + + if v is None: + continue + + try: + if ev.type == 'int': + expvars[ev.id] = int(v) + elif ev.type == 'float': + expvars[ev.id] = float(v) * 100 + except ValueError: + self.info('Failed to parse value for key {0} as {1}, ignoring key.'.format(ev.key, ev.type)) + return None + + return expvars + + @staticmethod + def _parse_memstats(data): + + memstats = data['memstats'] + + # calculate the number of live objects in memory + live_objs = int(memstats['Mallocs']) - int(memstats['Frees']) + + # calculate GC pause times average + # the Go runtime keeps the last 256 GC pause durations in a circular buffer, + # so we need to filter out the 0 values before the buffer is filled + gc_pauses = memstats['PauseNs'] + try: + gc_pause_avg = sum(gc_pauses) / len([x for x in gc_pauses if x > 0]) + # no GC cycles have occured yet + except ZeroDivisionError: + gc_pause_avg = 0 + + return { + 'memstats_heap_alloc': memstats['HeapAlloc'], + 'memstats_heap_inuse': memstats['HeapInuse'], + 'memstats_stack_inuse': memstats['StackInuse'], + 'memstats_mspan_inuse': memstats['MSpanInuse'], + 'memstats_mcache_inuse': memstats['MCacheInuse'], + 'memstats_sys': memstats['Sys'], + 'memstats_live_objects': live_objs, + 'memstats_gc_pauses': gc_pause_avg, + } diff --git a/collectors/python.d.plugin/go_expvar/go_expvar.conf b/collectors/python.d.plugin/go_expvar/go_expvar.conf new file mode 100644 index 0000000..4b821cd --- /dev/null +++ b/collectors/python.d.plugin/go_expvar/go_expvar.conf @@ -0,0 +1,108 @@ +# netdata python.d.plugin configuration for go_expvar +# +# This file is in YaML format. Generally the format is: +# +# name: value +# +# There are 2 sections: +# - global variables +# - one or more JOBS +# +# JOBS allow you to collect values from multiple sources. +# Each source will have its own set of charts. +# +# JOB parameters have to be indented (using spaces only, example below). + +# ---------------------------------------------------------------------- +# Global Variables +# These variables set the defaults for all JOBs, however each JOB +# may define its own, overriding the defaults. + +# update_every sets the default data collection frequency. +# If unset, the python.d.plugin default is used. +# update_every: 1 + +# priority controls the order of charts at the netdata dashboard. +# Lower numbers move the charts towards the top of the page. +# If unset, the default for python.d.plugin is used. +# priority: 60000 + +# penalty indicates whether to apply penalty to update_every in case of failures. +# Penalty will increase every 5 failed updates in a row. Maximum penalty is 10 minutes. +# penalty: yes + +# autodetection_retry sets the job re-check interval in seconds. +# The job is not deleted if check fails. +# Attempts to start the job are made once every autodetection_retry. +# This feature is disabled by default. +# autodetection_retry: 0 + +# ---------------------------------------------------------------------- +# JOBS (data collection sources) +# +# Any number of jobs is supported. +# +# All python.d.plugin JOBS (for all its modules) support a set of +# predefined parameters. These are: +# +# job_name: +# name: myname # the JOB's name as it will appear at the +# # dashboard (by default is the job_name) +# # JOBs sharing a name are mutually exclusive +# update_every: 1 # the JOB's data collection frequency +# priority: 60000 # the JOB's order on the dashboard +# penalty: yes # the JOB's penalty +# autodetection_retry: 0 # the JOB's re-check interval in seconds +# +# Additionally to the above, this plugin also supports the following: +# +# url: 'http://127.0.0.1/debug/vars' # the URL of the expvar endpoint +# +# As the plugin cannot possibly know the port your application listens on, there is no default value. Please include +# the whole path of the endpoint, as the expvar handler can be installed in a non-standard location. +# +# if the URL is password protected, the following are supported: +# +# user: 'username' +# pass: 'password' +# +# collect_memstats: true # enables charts for Go runtime's memory statistics +# extra_charts: {} # defines extra data/charts to monitor, please see the example below +# +# If collect_memstats is disabled and no extra charts are defined, this module will disable itself, as it has no data to +# collect. +# +# Please visit the module wiki page for more information on how to use the extra_charts variable: +# +# https://github.com/netdata/netdata/tree/master/collectors/python.d.plugin/go_expvar +# +# Configuration example +# --------------------- + +#app1: +# name : 'app1' +# url : 'http://127.0.0.1:8080/debug/vars' +# collect_memstats: true +# extra_charts: +# - id: "runtime_goroutines" +# options: +# name: num_goroutines +# title: "runtime: number of goroutines" +# units: goroutines +# family: runtime +# context: expvar.runtime.goroutines +# chart_type: line +# lines: +# - {expvar_key: 'runtime.goroutines', expvar_type: int, id: runtime_goroutines} +# - id: "foo_counters" +# options: +# name: counters +# title: "some random counters" +# units: awesomeness +# family: counters +# context: expvar.foo.counters +# chart_type: line +# lines: +# - {expvar_key: 'counters.cnt1', expvar_type: int, id: counters_cnt1} +# - {expvar_key: 'counters.cnt2', expvar_type: float, id: counters_cnt2} + |