grafilter

Dashboards like Grafana are awesome, but sometimes they make it hard to look at all the metrics, without first spending time to configure a dashboard. Grafilter is meant to provide lightweight and easy access to your metrics. You can still customize visuals of course, but in a way that let's you store these customizations in a versioned config management system such as Ansible, BundleWrap or Chef as opposed to a database like Grafana 2 does.

Grafilter supports InfluxDB 0.9.0 and onwards.

     

Install

apt-get install python3-pip
pip3 install grafilter

Configure

Create a file with these contents anywhere:

CACHE_TIMEOUT = 300
CONFIG_DIR = "/var/lib/grafilter"
DEBUG = False
IGNORED_TAGS = ["_key"]
INFLUXDB_DB = "metrics"
INFLUXDB_URL = "http://user:pass@influxdb.example.com:8086"

Note that INFLUXDB_URL points to the HTTP API port of InfluxDB.

Run

GRAFILTER_SETTINGS=/path/to/grafilter.cfg grafilter

Customize

You can customize the appearance of your individual metrics by placing files in the metrics subdirectory of your CONFIG_DIR. These files must have a .json extension and look like this:

{
	"pattern": "^load/",
	"transform": "lambda x: math.floor(x)",
	"type": "area"
}

Note that pattern is a regular expression that has to match the name of the metrics you want to customize. The name is formatted like this: measurement/tag1:value1/tag2:value2. Each metric will only be styled according to the first file with a matching pattern. All other options listed here are optional.

Option	Explanation
max	force highest value on y axis (int or float)
merge	a Python expression describing a function that takes two numeric values and the time interval between them and returns one numeric value. It can be used to calculate the difference between two sequential datapoints and use the result in your graph (useful for ever increasing counters, e.g. lambda x, y, t: (y - x) / t). You have access to the math module from the Python standard library. Applied before transform.
min	force lowest value on y axis (int or float)
short_name	a cosmetic name for this metric, e.g. "Load Average"
stack	set this to true if you want to stack values when selecting multiple metrics
transform	a Python expression describing a function that takes a numeric value and returns another. It can be used to perform conversions and other math on your metrics. You have access to the math module from the Python standard library. Applied after merge.
type	chart type: area, area-spline, area-step, line, spline or step (defaults to line)
unit_format	a d3 format specifier for your y axis (e.g. "s" to turn "1000" into "1K" and "1000000" into "1M")
unit	any string you want to be displayed next to your y axis