... because it is of no use since neutron-lbaas was retired and was
replaced by Octavia. This feature was officially deprecated during
Yoga cycle[1] so we are ready to remove it.
[1] 8917c73964
Change-Id: Ic145c23cc0b0372ef78f4a45ffb084bec24936c3
Gnocchi start returning regular 404 when requesting resource that do not
exist [1]. That means, that in case of 404, which is also excpected,
we just fail resource creation.
[1] http://paste.openstack.org/show/796938/
Change-Id: Ie5f8caba3111db7e7e8e50b9bc58c6e8de606338
Gnocchi start returning regular 404 when requesting resource that do not
exist [1]. That means, that in case of 404, which is also excpected,
we just fail resource creation.
[1] http://paste.openstack.org/show/796884/
Change-Id: I7061a31405a421da3ab2f192259fb23c5aa54ce6
Adds the loadbalancer resource type for usage
with the network.services.lb.* metrics.
Change-Id: Ib6e78438c3da0e22d93f720f00cdeadf0ed7a91f
Closes-Bug: #1848286
Signed-off-by: Jorge Niedbalski <jnr@metaklass.org>
Since moving to libvirt nova-compute poller we no longer have
the availability_zone attribute of an instance available.
We can capture instance.create.end events and create gnocchi resources
based on these which do have the availability_zone availabile.
Currently we are only handling delete events for resources, this
patch adds the handling of create events for this purpose.
Change-Id: Ieddbddd4ddb6af11d5158f5c90c87ae4f847bb96
In default, since the url in gnocchi has no netloc, it will trigger the
keystoneauth.serssion to get the gnocchi endpoint again. It will
consumer lots of CPU time and make lots of pressure on keystone service.
This patch try to pass the gnocchi endpoint into adapter through
endpoint_override parameter. It saves around almost 20% CPU time.
Change-Id: Ia0c825a2c81298c3a10fedc1ad3861f1d0b2f3e8
This change adds launched_at/started_at/deleted_at attribute to instance.
Rely on started_at/ended_at of Gnocchi to known when the operation have
done is not precise. Nova records the exact time of this event, just
copy them to Gnocchi.
Change-Id: If7870fd525d1ef58e4c2dbbbf93525bab3623b0f
support for explicitly setting interface/regoin_name was changed in
gnocchiclient 3.1 and remove in gnocchiclient 4.
temporarily cap gnocchiclient to allow gate to pass
Closes-Bug: #1714483
Change-Id: I851e16f537f547fb3a31fb22535dd4dae62fb235
Availability zone and name attributes were marked
as required while they are not, hence causing errors
when Gnocchi tried to create the resource.
This patch set removes the requirement over those attributes.
Availability zone has been removed, it will be shown only
when its present. Name has been replaced with share_id, which
is always present.
Other attributes has been added as well to keep track of other
relevant data.
Closes-Bug: #1706440
Change-Id: I191bcfae155278f7688c1148ad7e6994a0322fa6
By default, there is no supporting configuration in Ceilometer
to handle CPU monitors provided by the Nova compute_monitors
driver (cpu.virt_driver in this case).
Added metric defitions provided by cpu.virt_driver to
ceilometer/dispatcher/data/gnocchi_resources.yaml under a new
resource type, nova_compute. This enables these metrics to be
passed by the dispatcher to Gnocchi.
Added create_resource_type operation to gnocchi_client
upgrade_resource_types operations. This way the resources_initial
dict can stay static, but new resource types can be added as
needed by the upgrade resource types function.
Change-Id: I7698ba3168f79d05943042971039aea9a4c4e1a9
Closes-Bug: #1693846
This overrides the number of $executor_thread_pool_size with a global option
that is also used to set the number of parallel requests to Gnocchi that can be
done.
Change-Id: Iaa7e3d0739a63d571dd2afc262d191dffe5a0eef
set gnocchi timeout to 6.05 because requests recommends multiple
of 3[1]. also, it seems safe to assume gnocchi is not functioning well
and will not scale if we can't post in 6s. in my benchmark, it took
under 0.25s to post 5000 points. the only thing that took >3s was
getting >80K datapoints and this workflow is only pushing, and only
getting small chunks occasionally. making it 6s because it seems
keystone adds decent overhead.
do not set timeout for upgrade because it might take longer than normal
workflow
[1] http://docs.python-requests.org/en/master/user/advanced/#timeouts
Change-Id: I41b5e4f0c0f54b7f4b627d9116e63096ca2decf9
Closes-Bug: #1666072
We currently track only the flavor id, but this one can disapear over
the time. When the name is always present.
Change-Id: Ib1da71a0fad06c844b75683193a10e774a1c94e3
This change will allow to remove keystoneauth in Gnocchi
dispatcher for people where authentication of Gnocchi
request doesn't make sense.
Change-Id: If3a47f091a80333a0e7e0246e28040254c4f423b
This remove pollsters option from configfile sample
due to a duplicate option registration. This will
be fixed later.
The exchange_control group in the config sample doesn't exit
in reality exchange are in DEFAULT group.
This removes usage of cfg.CONF everywhere left.
This adds all missing OPTS in sample file.
Change-Id: I48c11ee7e1aae65847958b98532b3bdb48a3ceb5
This adapter is created by keystoneauth1 when no request session
object is pass to keystoneauth1. So remove it from Ceilometer.
Change-Id: I71f16ab0830cb4f3931be8615ffeb7fecf6f6af3
When we use ceilometer-upgrade, we get the same gnocchi resources as
gnocchi-upgrade produces. To avoid confusing people with the duplicate
resources, we will drop the gnocchi side. So the comment has no sense.
Change-Id: Id5e31e3c2015a65d63a7ad787a93036c6214c279
Takashi Kajinami