Gluster cluster wide profiling and analysis web app
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


A cluster wide profiling and analysis web app for GlusterFS


This is in very early stages of development, but making good progress already


Usage Graph Most Used Files Most Common Operations


  • Glupy (part of Gluster 3.5 and above, but can be installed manually on Gluster 3.4)
  • Python 2.7.x
  • ElasticSearch 0.90.x
  • Kibana 3
  • LogStash 1.3.3 or above

Supported Platforms

Linux for the GlusterFS nodes:

  • RHEL/CentOS 6.x is known to work
  • Fedora 19 is known to work
  • Debian/Ubuntu should work (but untested so far)

For the LogStash, ElasticSearch, and Kibana nodes, anything supported by those platforms should be good enough.


1. Ensure you have a working LogStash, ElasticSearch, & Kibana 3 cluster

The instructions on the LogStash website should get you started.

2. Install the Logstash Formatter for Python

At present, we pass information to LogStash using a file based approach, so need to install the "logstash_formatter" Python module.

On a new minimal install of CentOS 6.5, the normal pip installation command doesn't work out of the box so you'll need to run this first:

$ sudo yum install python-pip
$ sudo pip install --upgrade pip setuptools
$ sudo pip install Distribute

After that, the normal pip installation commands should work:

$ sudo pip install logstash_formatter

3. Ensure you have Glupy on your GlusterFS nodes

GlusterFlow depends on Glupy being installed on your GlusterFS nodes first:

  • If you're using a recent version of GlusterFS (3.5 and above), then Glupy is already a part of it.
  • If you're using GlusterFS 3.3 or 3.4, then you'll need to install Glupy from the external project repository:

4. Install the GlusterFlow filter

Follow the README instructions in the filter/ directory

5. Install the GlusterFlow Translator

Instructions still To Be Written.

Running it

With the GlusterFlow filter active, GlusterFlow messages should automatically be sent to ElasticSearch on the local host, for viewing in Kibana.

To Do

  • Update for multi-node operation (hopefully not hard)
  • Write instructions for using the Kibana Dashboard
  • Ensure the "Most Used Files" panel only shows files (not dir names)
  • Add a "Most Used Directories" panel
  • Add authentication
  • Consider adding thresholds and alerting
  • Investigate replacing the filter script with the Python code jdarcy mentioned