 Hello and welcome to the OpenStack Telemetry project overview and update for the Wallaby cycle. My name is Matthias Runge. If you don't know me, I'm a principal software engineer at Red Hat and I have been around with OpenStack since the Grizzly cycle. I've been recently elected as PTL for telemetry and I am the successor of Rong who is currently serving as PTL for the project Murano and for Solom. So what does telemetry do? Telemetry is used for gathering metrics and events. It is listening on the message bus and captures events like a VM was spawned, a network was created or a volume has been deleted. It is also using the OpenStack API to pull data out of the services to collect information about usage, memory consumption on a tenant basis, combined with the service AODH or A, and together with HEAT it provides an autoscaling service, for example to scale up or scale down resources. OpenStack Telemetry was probably founded during the Grizzly cycle and it started with only CELOMETER, which was later split into separate components during the Metacar cycle. Actually, I couldn't find accurate data about this and please correct me if I'm wrong here. Over the past cycle we had about 20 different contributors. While this number sounds pretty great and sane, I would love to encourage you all to contribute more because telemetry is an important part of OpenStack as we see from the latest survey numbers. Telemetry is used in 45% of production environments, is being tested in 8% of deployments and another 14% are considering to use it. Since there are a lot of different names around and under the telemetry umbrella, I have an architecture overview here where you can see CELOMETER is collecting data. Metrics are being sent to PANCO, metrics to NOKI and A is the alarming component. It is repeatedly pulling metrics from the NOKI API, compares the results with its rules and would issue a call to HEAT if an alarm is triggered. The most notable feature over the past cycle or past few cycles is the dynamic pulse subsystem. The idea here is to create or update pulse tests on the fly, which wasn't possible before. Pulse tests are being used to pull metrics out via the OpenStack API. For the future, we have two major changes in planning. You may have seen discussions around NOKI and NOKI being supported or not. NOKI being scalable or not, etc., etc. Personally, I would like to solve this rather sooner than later. There has been discussion around getting NOKI back under the OpenStack umbrella. To be honest, I'm not sure what that would solve compared to NOKI being independent. At least the discussion about that was uncontroversial without a clear outcome. Unfortunately, HEAT also stopped testing against CELOMETER and NOKI because it caused them to many issues. For the future, I would like to encourage you to contribute to this Leverie project. With that, I'm turning to my last slide. And I would like to hear your feedback, hear your pain points or see use cases. And finally, if you didn't participate in the survey yet, please consider to do so. Thank you.