path: root/collectors/python.d.plugin/changefinder/integrations/python.d_changefinder.md

<!--startmeta
custom_edit_url: "https://github.com/netdata/netdata/edit/master/collectors/python.d.plugin/changefinder/README.md"
meta_yaml: "https://github.com/netdata/netdata/edit/master/collectors/python.d.plugin/changefinder/metadata.yaml"
sidebar_label: "python.d changefinder"
learn_status: "Published"
learn_rel_path: "Data Collection/Other"
message: "DO NOT EDIT THIS FILE DIRECTLY, IT IS GENERATED BY THE COLLECTOR'S metadata.yaml FILE"
endmeta-->

# python.d changefinder

Plugin: python.d.plugin
Module: changefinder

<img src="https://img.shields.io/badge/maintained%20by-Netdata-%2300ab44" />

## Overview

This collector uses the Python [changefinder](https://github.com/shunsukeaihara/changefinder) library to
perform [online](https://en.wikipedia.org/wiki/Online_machine_learning) [changepoint detection](https://en.wikipedia.org/wiki/Change_detection)
on your Netdata charts and/or dimensions.


Instead of this collector just _collecting_ data, it also does some computation on the data it collects to return a changepoint score for each chart or dimension you configure it to work on. This is an [online](https://en.wikipedia.org/wiki/Online_machine_learning) machine learning algorithm so there is no batch step to train the model, instead it evolves over time as more data arrives. That makes this particular algorithm quite cheap to compute at each step of data collection (see the notes section below for more details) and it should scale fairly well to work on lots of charts or hosts (if running on a parent node for example).
### Notes - It may take an hour or two (depending on your choice of `n_score_samples`) for the collector to 'settle' into it's
  typical behaviour in terms of the trained models and scores you will see in the normal running of your node. Mainly
  this is because it can take a while to build up a proper distribution of previous scores in over to convert the raw
  score returned by the ChangeFinder algorithm into a percentile based on the most recent `n_score_samples` that have
  already been produced. So when you first turn the collector on, it will have a lot of flags in the beginning and then
  should 'settle down' once it has built up enough history. This is a typical characteristic of online machine learning
  approaches which need some initial window of time before they can be useful.
- As this collector does most of the work in Python itself, you may want to try it out first on a test or development
  system to get a sense of its performance characteristics on a node similar to where you would like to use it.
- On a development n1-standard-2 (2 vCPUs, 7.5 GB memory) vm running Ubuntu 18.04 LTS and not doing any work some of the
  typical performance characteristics we saw from running this collector (with defaults) were:
    - A runtime (`netdata.runtime_changefinder`) of ~30ms.
    - Typically ~1% additional cpu usage.
    - About ~85mb of ram (`apps.mem`) being continually used by the `python.d.plugin` under default configuration.


This collector is supported on all platforms.

This collector supports collecting metrics from multiple instances of this integration, including remote instances.


### Default Behavior

#### Auto-Detection

By default this collector will work over all `system.*` charts.

#### Limits

The default configuration for this integration does not impose any limits on data collection.

#### Performance Impact

The default configuration for this integration is not expected to impose a significant performance impact on the system.


## Metrics

Metrics grouped by *scope*.

The scope defines the instance that the metric belongs to. An instance is uniquely identified by a set of labels.



### Per python.d changefinder instance



This scope has no labels.

Metrics:

| Metric | Dimensions | Unit |
|:------|:----------|:----|
| changefinder.scores | a dimension per chart | score |
| changefinder.flags | a dimension per chart | flag |



## Alerts

There are no alerts configured by default for this integration.


## Setup

### Prerequisites

#### Python Requirements

This collector will only work with Python 3 and requires the packages below be installed.

```bash
# become netdata user
sudo su -s /bin/bash netdata
# install required packages for the netdata user
pip3 install --user numpy==1.19.5 changefinder==0.03 scipy==1.5.4
```

**Note**: if you need to tell Netdata to use Python 3 then you can pass the below command in the python plugin section
of your `netdata.conf` file.

```yaml
[ plugin:python.d ]
  # update every = 1
  command options = -ppython3
```



### Configuration

#### File

The configuration file name for this integration is `python.d/changefinder.conf`.


You can edit the configuration file using the `edit-config` script from the
Netdata [config directory](https://github.com/netdata/netdata/blob/master/docs/configure/nodes.md#the-netdata-config-directory).

```bash
cd /etc/netdata 2>/dev/null || cd /opt/netdata/etc/netdata
sudo ./edit-config python.d/changefinder.conf
```
#### Options

There are 2 sections:

* Global variables
* One or more JOBS that can define multiple different instances to monitor.

The following options can be defined globally: priority, penalty, autodetection_retry, update_every, but can also be defined per JOB to override the global values.

Additionally, the following collapsed table contains all the options that can be configured inside a JOB definition.

Every configuration JOB starts with a `job_name` value which will appear in the dashboard, unless a `name` parameter is specified.


<details><summary>Config options</summary>

| Name | Description | Default | Required |
|:----|:-----------|:-------|:--------:|
| charts_regex | what charts to pull data for - A regex like `system\..*/` or `system\..*/apps.cpu/apps.mem` etc. | system\..* | True |
| charts_to_exclude | charts to exclude, useful if you would like to exclude some specific charts. note: should be a ',' separated string like 'chart.name,chart.name'. |  | False |
| mode | get ChangeFinder scores 'per_dim' or 'per_chart'. | per_chart | True |
| cf_r | default parameters that can be passed to the changefinder library. | 0.5 | False |
| cf_order | default parameters that can be passed to the changefinder library. | 1 | False |
| cf_smooth | default parameters that can be passed to the changefinder library. | 15 | False |
| cf_threshold | the percentile above which scores will be flagged. | 99 | False |
| n_score_samples | the number of recent scores to use when calculating the percentile of the changefinder score. | 14400 | False |
| show_scores | set to true if you also want to chart the percentile scores in addition to the flags. (mainly useful for debugging or if you want to dive deeper on how the scores are evolving over time) | False | False |

</details>

#### Examples

##### Default

Default configuration.

```yaml
local:
  name: 'local'
  host: '127.0.0.1:19999'
  charts_regex: 'system\..*'
  charts_to_exclude: ''
  mode: 'per_chart'
  cf_r: 0.5
  cf_order: 1
  cf_smooth: 15
  cf_threshold: 99
  n_score_samples: 14400
  show_scores: false

```


## Troubleshooting

### Debug Mode

To troubleshoot issues with the `changefinder` collector, run the `python.d.plugin` with the debug option enabled. The output
should give you clues as to why the collector isn't working.

- Navigate to the `plugins.d` directory, usually at `/usr/libexec/netdata/plugins.d/`. If that's not the case on
  your system, open `netdata.conf` and look for the `plugins` setting under `[directories]`.

  ```bash
  cd /usr/libexec/netdata/plugins.d/
  ```

- Switch to the `netdata` user.

  ```bash
  sudo -u netdata -s
  ```

- Run the `python.d.plugin` to debug the collector:

  ```bash
  ./python.d.plugin changefinder debug trace
  ```

### Debug Mode



### Log Messages