* All the telemetry related projects intree tempest plugins
are now moved to telemetry-tempest-plugin
Change-Id: I57738964f65495364dcc74821634d54aea1769fd
This conf.py is redundant, all project guides only use the
"doc/source/conf.py" file. We should remove this one.
Change-Id: I820688d3ffbf01cc057c1a44bfd69f5291841159
do not build metrics off compute.instance.create.start as it
doesn't have enough information in it to push to Gnocchi. there is
an .end event that happens seconds later so .start event has little
value outside context of event.
we change fnmatch to regex for more useful regex comparison and it's
faster
not a complete fix as compute.instance.update also is missing host
value occasionally
Partial-Bug: #1718290
Change-Id: I2aa7045873fd740255a3be5f64b9e9f5fd94e2d4
i'm not sure what this does but it is: not used in
gnocchi_resources.yaml, not tested, and not ever explained in
original patch[1]
[1] https://review.openstack.org/#/c/289807
Change-Id: I20c158747c79a05778f27bf4b29319a6262db640
we run through samples and format samples and resources to what
gnocchi expects. this simplifies that process a little:
- don't needlessly sort and group by metric name
- build a full resource model iff resource needs to be created
Change-Id: I2ac25b3b0978eed664c500e645bae2d1b4ae6781
we shouldn't waste our time looping through resources and searching
metrics to find a match. just build hashmap and find the resource
based on metric.
this removes fnmatch functionality because we don't match on
wildcards and the code itself actually requires metric names to be
explicit to create_metrics
Change-Id: I2398247270217759c876ab5a9b60038dad79a9d3
- remove note about previously storing events as meters.
- docs are not branchless anymore, remove anything that has been
removed
- minor grammar changes
Change-Id: If8486f5edb3d83c2d520345c028bc3f23609b376
ujson has not had any active maintenance for the last 12 months;
switch to using json module instead.
Change-Id: I39027b534e94b3f877d881647a7c843183f60f92
Closes-Bug: 1737989
- remove 'prior to ocata' conditions.
- remove shuffle_time_before_polling_task as it doesn't work in
reality and is being removed.
- add a note when to enable workload_partitioning of notification
agent
Change-Id: I44c030835de1517a3c067ab0632c09f4a5fe2f15
- add a note mentioning transformers are not durable and you
might be able to use gnocchi
- explain how to enabled/disable pipelines.
Change-Id: I3f50d5b5a1e22a114ef2c8e17c1e659f1c0eb761
- remove list of notifications used for meters, it's not accurate
and is arguably of no value to an administrator even if it's
accurate. just list where the endpoints are and how to configure.
- remove comments about database connection, this means nothing
- remove libvirt comment, listing all services polled as these are
stated elsewhere and are just repeated.
Change-Id: If67ecf7e507da70abae3734618358c2458f49037
- remove reference to api
- remove dup details on compute, central, ipmi already in data
collection
- remove hbase from panko support
- remove listing libvirt hypervisors, we only test kvm/qemu.
- remove generic listing of 'OpenStack Networking' meters. we already
have a measurements page for this
- remove user, roles, projects notes. we don't have api and it
means very little as is
Change-Id: I292cdd9297549ab8cb78caf6b15941b3986ca898
