Increase timeout for metric purging

[jntullo]

BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1445498

Metric purging currently uses the default timeout of 10 minutes which is not long enough for an environment with thousands of VMs. This is causing metrics tables to grow in size so large that they eventually fill the entire VMDB.

This PR adds a queue_timeoutto the settings for metrics_history and performance.history so that it can be altered as needed, and increases it by default to "20.minutes"

Thoughts @yrudman?
@miq-bot bug, core/queue

[miq-bot]

@jntullo unrecognized command 'bug', ignoring...

Accepted commands are: add_label, assign, close_issue, move_issue, remove_label, rm_label, set_milestone

[jntullo]

@miq-bot add_label bug, core/queue

[yrudman]

do we also need to increase performance_realtime_purging_interval to avoid overlapping in scheduler ?

[yrudman]

there is still 5 minutes gap between timeout and next scheduler loop, does it make sense
to make purging interval very close to timeout (may be 21.minutes) ?

[jntullo]

@yrudman yeah, that's a good idea. I wasn't sure if we could keep the current 5 minute gap.

[miq-bot]

Checked commits jntullo/manageiq@da8d96b~...75f603c with ruby 2.2.6, rubocop 0.47.1, and haml-lint 0.20.0
3 files checked, 0 offenses detected
Everything looks fine. 👍

[jntullo]

@miq-bot add_label euwe/yes, fine/yes

[gtanzillo]

👍 LGTM

[simaishi]

Fine backport details:

$ git log -1
commit 6332db37fc931b7c4acc135d0a56dfe96f55d883
Author: Gregg Tanzillo <gtanzill@redhat.com>
Date:   Wed Jun 14 11:12:57 2017 -0400

    Merge pull request #15312 from jntullo/bz/purge_timeout
    
    Increase timeout for metric purging
    (cherry picked from commit 62719214d86bf68b169571af9a761fea6331f5f4)
    
    https://bugzilla.redhat.com/show_bug.cgi?id=1462358

[simaishi]

@jntullo Cherry-pick results in conflicts on all files 😞 Would you mind creating a PR for Euwe?

[jntullo]

@simaishi will do!

[simaishi]

Backported to Euwe via #15702