We previously used a less-than-optimal way of storing the configuration
values for Graphite/Carbon, which meant you had to override the entire
configuration (including restating default values, etc) to change one
thing.

This now mirrors how we approach things in the Grafana role, where each
distinct value is a distinct variable that can be overridden separately,
and is pulled together at a higher level into the configuration.

It's not a requirement and has the potentialy to appear like an actual
string, and not an empty string, which breaks things like Grafana.

fix alton bucket variable

Add new Alton variables 

Set a cluster name on new rabbit clusters

add monitoring to antivirus scans

This is also click-changeable in the UI, and seems to not impact running
clusters from some testing.
This gives you a better name than i-asdfasdfasdfasdf.rabbit.lol which is
what you get by default on AWS.
This is mostly seen in logs and rabbitmqctl output.

Revert "chmod and chown the edx_ansible_venv_dir to make vagrant provision more reliable"

Install jenkins nginx config to nginx_sites_available_dir

Add check to see if hive metastore is initialized.

Make it possible to run the newrelic play on a box without ec2 access

chmod and chown the edx_ansible_venv_dir to make vagrant provision more reliable

Jibsheet/multiple worker cluster detection

Many of our machines aren't granted DescribeInstances because there's no
real need.  This let's you -e in those vars instead.

This should have been set prior to https://github.com/edx/auth-backends/pull/9 being merged. We did not catch this problem until now because the library currently used for JWT validation, pyjwt, simply skips issuer validation if the issuer is set to None.

LEARNER-693

ECOM-3629

Robustify TLS nginx settings

Added the option to split env_extras by app

We can detect the SuspendedProcess properties that Asgard puts on the
ASG when it's being deleted.

Add HELP_TOKENS_BOOKS setting

Add a configuration setting for the edx language cookie

Upgrade Hadoop and Hive

This header will provide more information to downstream caches so they know which headers to consider when determining a match. This is especially important to ensure browsers do not cache incorrect CORS headers when switching between sites/tenants.

DEVOPS-5807

bake pip requirements into the loadtest worker AMI

This follows the same pattern as with the platform workers: (1) clone
the repo, (2) install requirements, (3) package the virtualenv into a
tarball which becomes part of the AMI.

PERF-393

* Updates paths for Hadoop and Hive installs, including Hadoop native library compiling.
* Initialize Hive metadata store schema

Always reboot when taking an AMI

Make changes for running analyticstack on 16.04.

* Replace llvm-gcc with gcc.
* Define new Hadoop services.
* Add retry when first using Hadoop.
* Add procps to debian packages to address bug.
* Increase memory allocation.

Not doing this means we can't guarantee filesystem integrity
and we've seen this bear out recently when several 16.04 instances had 0
length minos installs (the last task to run before the AMI is cut)

Add Enterprise Service worker username and API URL