Generate Kibana index pattern

[monicasarbu]

This PR creates a python script to generate the index pattern in json format from the fields.yml. The generated index pattern is needed by Kibana to know how to format the fields, think of a percentage value.

As Kibana supports number, string and boolean types, all the variations of half_float, float, integer, long become number in the index-pattern.

For special fields that require a special format (like percent, bytes), usually you need to define the format manually for each field in Kibana in order for Kibana to know how to display a number when it's a percentage (using %) or in MB/GB for bytes values. To automate the process, this PR lets you define the format of the field in the format option.

- name: used.bytes
   type: long
   format: bytes
   description: >
        Used memory.

- name: used.pct
   type: half_float
   format: percent
   description: >
        The percentage of used memory.

This PR is another way to solve elastic/beats-dashboards#117

[tsg]

Is "integer" correct here? Shouldn't it be "number"?

[monicasarbu]

Oh, good spot, @tsg. Thanks!

[tsg]

LGTM, great work! I see only the Packetbeat index was re-generated, should we do them all in this PR?

Also, needs a rebase.

[monicasarbu]

Yes, I will do for all the Beats in the same PR.

[monicasarbu]

With this change the import-dashboards.sh doesn't work anymore to load the dashboards and it needs to be changed in a different PR.

[ruflin]

It's nice to have this also in the docs 👍

[ruflin]

LGTM. Great work. Could you rebase?

[monicasarbu]

Rebased.

[ruflin]

Where are these changelog entries coming from?

[monicasarbu]

Yes, it looks like "git rebase master" didn't do a good job. They were part of alpha5. Thanks for noticing it. I am fixing it manually.