Facette 0.2.0 released

Here it is: the first big Facette milestone since its first release. We're happy to announce the version 0.2.0, that comes with a nice load of bug fixes, new features and various improvements.

New in 0.2.0:

InfluxDB connector

Due to popular demand, we've implemented a connector to query metrics stored in a InfluxDB database. This connector is experimental at the moment, and supports InfluxDB up to version 0.7.x. Check out the Facette documentation to see how to configure a provider using the influxdb connector.

UNIX socket support for backend

When using Facette through a reverse proxy on the same host, it is now possible to access the back-end server using a UNIX socket instead of a TCP connection to both reduce network overhead and improve security. Thanks to Skrupellos for this contribution.

"sieve" provider filter

The new sieve filter does the opposite of the discard filter: it only lets records whose target match the rule pattern pass through.

Series operations

Series operations — sum, average etc. — can now be performed on series coming from different origins, as the caculations are now handled internally by Facette.

This new version also introduces a new series operation: normalize. It aligns several series with different precisions on common time step (this is also done implicitely for sum and average operations).

Read-only mode

A read-only mode has been implemented to expose a Facette instance with all data-altering API calls disabled. It can be useful when you want to provide a "looking glass" graph front-end to you users and you don't want them to modify defined graphs and collections.

Embeddable graphs

Graphs toolbar now feature a "Embeddable graph" button that links the graph to a standalone view that can for instance be embedded in a page through a HTML iframe from another dashboard:

Multiple fixed graph plot lines

It is now possible to display multiple fixed plot lines on a graph to highlight significative values from the legend:

Series auto-naming

Series added to graphs at creation/modification stage are now automatically pre-named according to their origin/source/metric names instead of the previous serie# name.

Foldable graph legends

Graph legends are now foldable, and are folded by default to save space at collections display: click on the arrow a the bottom of a graph to expand the legend.

Series per stack sums

Tooltips on graphs now display the sum of the series values for each point. When using stacks, sums are displayed per stack.

Toggle graphs visibility in collections

If you created a collection, it is now possible to "hide" a graph included in that collection without having to remove it and add it later if need be. In a collection edition page, just click on the toggler near the graph name and it won't be displayed in the collection page.

Graphs collections auto-refresh

You can now set an automatic refresh delay at graph-level or collection-level.

Docker image support

A Dockerfile is now provided with the sources to build a Docker image straight from the source tree. Thanks to Michael Crosby from Docker for this contribution.

OpenSearch support

The landing page search bar is now available through the OpenSearch standard.

You can now search directly for Facette's items in your favorite browser.

HiDPI displays support

Facette now looks better when displayed on HiDPI-compatible (a.k.a. "Retina") screens.

Shortcut links are now available to help switching between edition and display modes when dealing with graphs and collections.

Changes

Base server settings are now optional

Prior to version 0.2.0 some base server settings were mandatory: default values are now set for all settings, refer to the documentation for the details.

Server catalog refresh API call removed

For consistency purpose we decided to remove the API call and front-end controls allowing to refresh server catalog from the front-end in favor of the facettectl utility.

Provider filter rules format

The provider filtering rules format has changed and is incompatible with the previous versions: please refer to the filters documentation to adapt your filtering rules. We're sorry for the inconvenience, but this was a necessary change to support matched pattern suppression (see this Github issue for further explanations).

Dude, where are my graphs?

As a side-effect of the new feature enabling to toggle graph visibility in a collection, all graphs listed in a collection will be hidden by default when you upgrade from 0.1.x to 0.2.0. Here is a little Python code snippet that you can drop in a shell to "convert" your existing collection definitions to the new format (don't forget to adapt the path to your actual Facette collections directory in the $FACETTE_COLLECTIONS variable):

export FACETTE_COLLECTIONS="/var/lib/facette/collections/"
cat <<EOF | python - $FACETTE_COLLECTIONS
import json
import os
import sys

for root, _, files in os.walk(sys.argv[1]):
  for file in files:
    collectionFile = os.path.join(root, file)

    print "* Converting data file", os.path.join(collectionFile)

    cfp = open(collectionFile, 'rw')
    collection = json.load(cfp)
    cfp.close()

    for entry in collection["entries"]:
      entry["options"]["enabled"] = True

    cfp = open(collectionFile, 'w')
    json.dump(collection, cfp, indent=4)
    cfp.close()
EOF

Thank you!

Thank you to our first users and especially early contributors: it's nice to see that Facette is getting some traction and visibility that soon, and your feedbacks help us improving this software.