Project Tracking: System Semantic Conventions stability

[frzifus]

Description

The OpenTelemetry Collector community would like to stabilize system semantic conventions including system metrics in order to help the adoption of the OTel hostmetricsreceiver.

Project Board

See: https://github.com/orgs/open-telemetry/projects/55/views/3

Roadmap

• Add system metrics reporting total memory capacity or clarify how to recover existing ones semantic-conventions#127
• (@ChrsMark) Define attributes to identify a CPU semantic-conventions#130
• (@mx-psi) Resource attributes for network addresses of a host semantic-conventions#131
• (@frzifus) Move metric definitions to YAML - System Metrics semantic-conventions#73
• (@dineshg13) Clarify semantics of system.cpu.time and system.cpu.utilization semantic-conventions#647
• The protocol attribute should be removed for system.network.connections metrics semantic-conventions#649
• (@braydonk) Add available memory state for hostmetrics receiver opentelemetry-collector-contrib#7417
• Document the difference between a host and a system metric semantic-conventions#226

done

• (@frzifus) add system.cpu.count metric semantic-conventions#98

Optional

• Add system uptime metric semantic-conventions#648
• Proposal to add system.memory.slab  semantic-conventions#531

Finally

• Stabilize existing semantic conventions (system metrics)

Deliverables

• Mark the system semantic conventions as stable.
• Update existing OpenTelemetry Collector hostmetricsreceiver to conform with the stable conventions.

Staffing / Help Wanted

The goal is to follow @tedsuo's proposed Semantic Convention Process.

• Stage 1: Working Group Preparation
• Stage 2: Stabilizing the Specification
• Stage 3: Implementation

Required staffing

• @mx-psi, @frzifus project leads
• @jsuereth technical committee sponsor
• domain experts:
  • @MovieStoreGuy
  • @dmitryax
  • @gbbr
  • @sallyom
  • @ChrsMark
  • @kaiyan-sheng
  • @braydonk
  • @bertysentry
  • @dineshg13

Meeting Times

July 10 - August 21 (6 weeks):

• Wednesday at 8:30-09:00 Pacific Time

cncf-slack #otel-system-metrics-wg
meeting-notes google-doc

Timeline

Stage 1 (Working Group Preparation) is happening now.

Stage 2 (Stabilizing the Specification) will begin as soon as we have adequate staffing for this project, and we coordinate a weekly meeting times (currently targeting mid-july).

Stage 3 (Implementation) will begin as soon as the system metrics are marked stable, and it should be relatively short we only need to update conformance to the specification for a single collector package.

Labels

The tracking issue should be properly labeled to indicate what parts of the specification it is focused on.

Linked Issues and PRs

All PRs, Issues, and OTEPs related to the project should link back to the tracking issue, so that they can be easily found.

[AlexanderWert]

Can we add @ChrsMark and @kaiyan-sheng as collaborators who will help with ECS alignment? Thanks!

[frzifus]

Sure, done. :)

[braydonk]

I work on the Google Cloud Ops Agent which packages the OpenTelemetry Collector for metrics collection, and the hostmetrics specification is very important to us. I also help support other teams at Google that are leveraging the collector and collecting hostmetrics.

I would love to join this working group and contribute to the specification, providing the perspective of both a vendor and user.

[frzifus]

Thats great, added you 🙂

[jsuereth]

As mentioned in the Specification meeting, do we have a back-log of issues to triage or work through as part of this working group? Could we add that in this proposal?

[frzifus]

We will try to deliver by the end of the week. We will compile the scope in a google doc and move it here.

cc @mx-psi

[bertysentry]

I'm definitely interested in helping in this wg

[dineshg13]

i am interested in this work.

[frzifus]

Hi @bertysentry, @dineshg13 - added you both :)

@jsuereth synced with @mx-psi and we linked the roadmap.

[jsuereth]

This is approved to run by TC.

@mx-psi @frzifus Are you able to set up the calendar invites and meeting notes?

I'll get you a project status board and move the issues listed to it to start.

[frzifus]

Sure, we will figure out how that works. 👍

[dmitryax]

We also need to find a way to capture "available" memory reported by Linux kernel (>3.16) open-telemetry/opentelemetry-collector-contrib#7417

[braydonk]

I'd be interested in helping with open-telemetry/opentelemetry-collector-contrib#7417

[ChrsMark]

I can look into the open-telemetry/semantic-conventions#130.

[tigrannajaryan]

Project is setup, closing the issue. If anything else is needed please reopen or create a new issue.