The customer environment has VNFs with different SLAs. For instance, vEPC VNFs alarms should be dealt by within 15 minutes, and other VNFs, like speedTest, have not this constraints.
Today, when an alarm is received, the customer cannot identify from which VNF this alarm is coming, so it has to login to the system to check which VNF corresponds to the ID alarming. If the name of the VNF were part of the alarm, the customer would be able to take decisions based purely on alarm, as expected from a central point of view (STF).
Would be of great value to have resource/project names in the virtual layer metrics (ceilometer and collectd both).
Can you provide an example of the alarm configuration that you're using, along with the output of what you're seeing. If I look at the raw ceilometer metric itself, I see both a resource and project label.
ceilometer reports the project and resource in metrics. From my (upstream) perspective, it doesn't make sense to add the names as well, especially since the names are not unique.
The output (in STF) is already:
ceilometer_cpu{container="sg-core", counter="cpu", cpu="864c6990-9823-44fc-a29f-60f4022ff845", endpoint="prom-http", host="f163e5566c31a57dd2b50bc15bef1491ac9248b5ef796d9947bd0a90", project="2a195da23e9b4fa5aefe06a2210dfdbd", publisher="telemetry.publisher.controller-0.redhat.local", resource="864c6990-9823-44fc-a29f-60f4022ff845", service="default-cloud1-ceil-meter-smartgateway", type="cpu", unit="ns"}
@mrunge I see you linked a review on this, but otherwise has been sitting in ASSIGNED for quite a while. What is the status on this issue, and is that review intended to support this functionality?