Scrappy visualizes resource usage in your Mesos cluster. Currently it only outputs text version, but a nice web interface can eventually be implemented on top of it.

It'd be nice to see that information exposed in Mesos UI nicely too.

Scrappy comes in a Docker container. Just point it to your Mesos master URL to get resource usage breakdown. There are two flavors available.

slaves flavor reports stats on slave level:

docker run --rm bobrik/scrappy slaves -u http://mesos-master:port

Sample report for you to know what to expect:

dev: 0.80 CPUs, 0.50GB RAM / 2.00 CPUs, 0.98GB RAM
  roles:
    - *: 0.80 CPUs, 0.50GB RAM / 2.00 CPUs, 0.98GB RAM
      tasks: 2
        - scrappy-example: 0.40 CPUs, 0.25GB RAM
        - scrappy-example: 0.40 CPUs, 0.25GB RAM

dev: 0.40 CPUs, 0.25GB RAM / 2.00 CPUs, 0.98GB RAM
  roles:
    - *: 0.40 CPUs, 0.25GB RAM / 2.00 CPUs, 0.98GB RAM
      tasks: 1
        - scrappy-example: 0.40 CPUs, 0.25GB RAM

The following options are available:

• -u Mesos URL to fetch data from.
• -s sort method, one of host, cpu, mem.
• -r reverse sort order.
• -f role name to filter on.

roles flavor reports stats on role level, useful if you have multiple of them:

docker run --rm bobrik/scrappy roles -u http://mesos-master:port

Sample report for just one default role for you to know what to expect:

role CPUs used CPUs total     CPU %  RAM used RAM total     RAM %
   *      0.30       2.00    15.00%    0.12GB    0.98GB    12.79%

The following options are available:

• -u Mesos URL to fetch data from.
• -f role name to filter on.