openstack/puppet-openstack-integration - puppet-openstack-integration - OpenDev: Free Software Needs Free Tools

These are removed in commit [1]
[1] https://review.openstack.org/#/c/569532/
Change-Id: If458c9cfbcdb07dd9da9debefe7af316426203c8

puppet-gnocchi switched to www_authenticate_uri in
[1], update puppet-openstack-integration for the same.
[1] https://review.openstack.org/#/c/558445/
Change-Id: I91f217b6d20f3a20346a6d281414fc5838a12113

Change-Id: Ifcac59aebe6e20044d88e5bf6246b355cb426d00

Change-Id: I3680dbad11435ac82ad03d0195d8e348d21b1433

Change-Id: I88145f1b20165f4a7581767c6406c5af98e8f32c

Improving scenario001 use-case where we use Telemetry services for
Auto-Scaling. It's a end-to-end scenario, so it will really test all the
stack.
- It adds Heat to scenario001
- It runs TestAutoscalingGabbi test
- It increases token expiration to 2400
Change-Id: Ibaa44cf53b8c8fb59b765ce4f9b219dd65f38a42

Because we configure Keystone to expire tokens after 600s, we don't
want to rely on default value in Gnocchi which is 300s to cleanup old data.
Indeed, Gnocchi might use an old token that expired to clean up and then it would
fail. It happens when running Tempest tests in the gate with low resources.
Production value (300) shouldn't be changed by default.
Depends-On: I8d2fccb0cef49246b3c15b0942ba6067519cee1f
Change-Id: If4d4c57698048ea89b9b271e66d6a7a031d21b33

Ubuntu and Redhat/Centos doesn't have all telemetry packages.
Instead of checking everywhere if we are on Ubuntu or Redhat/Centos,
this change uses a boolean 'enable_telemetry_legacy'.
If true, ::openstack_integration::ceilometer will configure ceilometer
with legacy services/configuration (without panko/gnocchi). If false,
Gnocchi and Panko backend is used.
scenario001 set it to true for Ubuntu and false for Redhat/Centos.
This centralize the kind of ceilometer deployement in one place.
This change also drops the now useless collector configuration and
configure the notification agent correctly instead.
Change-Id: If229829975d4f25ca0a2a3c6c93e674204d71dc2

Currently gnocchi metricd is really slow making some tests timeout,
because the tooz backend is mysql.
And the futur autoscaling will reach the max connections of mysql
by creating many tooz locks.
This change switches to redis, this is one of the recommended backend to
use (mysql is not).
Change-Id: Iae81848e883f74b7f8f8f4cad58c2168e01b7321

Change-Id: Iac8fa0035934c06850833368a13625e797f11899

Change-Id: I1c6f3fad8ea1896b409f57b0268883bb99a2decd

Change-Id: Ia4ba77e4ea6e2afe591772cb33cd4ca0080ef6a6

This patch enables SSL by default on scenario001 for both Ubuntu &
CentOS7, like it's done for scenario002.
Change-Id: If7a6ae3825052420ac76b5c8899ce68f003cb903

Reducing workers will help us to consume less memory reduce the number
of timeouts we have in upstream CI.
* Neutron & Swift: reduce from 4 to 2 workers
* When possible, set workers to 2: aodh, ceilometer, gnocchi.
Change-Id: Ie064a9ad0389f112bbb69f4253f7f6dbc26deb0a

After Glance, Nova, Cinder, now we want to deploy Gnocchi with Ceph
backend instead of file.
Change-Id: I1f8c83da9d3dfcee597cec7643ebc0e09ecea30d

* Deploy Gnocchi services
* Create default archive policy and archive policy rule
* Configure Aodh with Gnocchi which is used to evaluate the alarms for threshold
Note: Tempest does not fully support Gnocchi yet.
* test_alarming_api will work (Aodh) with Gnocchi
* test_telemetry_notification_api will not work (Ceilometer) with
 Gnocchi. That's why we don't configure Ceilometer to use Gnocchi
 dispatcher.
Depends-On: I4ed26d3e294e82c383afc443914b0d7bb1906e46
Depends-On: Ib818b5ae4fa301c6c496fe800ae45c9b0a73222c
Change-Id: I32e5f80e4eb8a5cd8fc07798aaf6d710332fcdda