Closed Bug 1570386 Opened 6 years ago Closed 5 years ago

Discuss how to "suggest new or change existing metric types" for the Glean SDK

Categories

(Data Platform and Tools :: Glean: SDK, task, P1)

task

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: Dexter, Assigned: Dexter)

References

Details

(Whiteboard: [telemetry:glean-rs:m11])

We need a clear and open process for suggesting new metric types for the Glean SDK (e.g. Urls?). The Glean team should not be deciding alone what's potentially useful or not, other stakeholders might have opinions: data science, data engineering.

We need to create a process for requesting new metric type and make it available in the glean documentation.

Priority: -- → P3
Whiteboard: [telemetry:glean-rs:m?]
Whiteboard: [telemetry:glean-rs:m?] → [telemetry:glean-rs:m9]
See Also: → 1573191
Whiteboard: [telemetry:glean-rs:m9] → [telemetry:glean-rs:backlog]
Blocks: 1595723
Blocks: 1557299
Blocks: 1557358
Blocks: 1557353
Summary: Discuss how to "suggest new metric types" for the Glean SDK → Discuss how to "suggest new or change existing metric types" for the Glean SDK
Blocks: 1560068
Assignee: nobody → alessio.placitelli
Priority: P3 → P1
Whiteboard: [telemetry:glean-rs:backlog] → [telemetry:glean-rs:m11]

The final proposal document is available here: https://docs.google.com/document/d/1Iaol-MfpQxgTeE9FIyptBx5LfNhMUwDaT84J6XLvhwM/edit?ts=5ddbdf6e#

The first round of feedback is for the Glean SDK team only. Please kindly review before Monday the 9th of December, as I'm planning to expand the audience at the Monday meeting.

Please do sign-off at top of the document, if you want (no need to do that).

Flags: needinfo?(tlong)
Flags: needinfo?(mdroettboom)
Flags: needinfo?(jrediger)
Flags: needinfo?(chutten)
Flags: needinfo?(beatriz.rizental)

General comments:

  • It might be nice to have a clean statement of what the expected resource cost is for being a steering committee member. "How many hours a week will I spend doing what, exactly?" sort of thing
  • I'm not sure where it might fit, but I think a question of "How many metrics of this type are expected?" might be worth considering as part of the decision. If there's only one distribution of samples of unit "pixels^2", then we can foist it off on custom_distribution. But if they are going to be the latest fad of measurement, then maybe we should have a screen_rect_distribution after all.
Flags: needinfo?(chutten)

Read and added a couple of minor comments and I generally agree with the other comments to the doc.

Flags: needinfo?(tlong)

Raised a couple of small questions, but nothing blocking the overall proposal.

Flags: needinfo?(jrediger)

My comments are in the doc.

Flags: needinfo?(mdroettboom)
Flags: needinfo?(beatriz.rizental)
Depends on: 1603639
Blocks: 1608137

The proposal document from comment 2 was sent around for feedback.

Status: NEW → RESOLVED
Closed: 5 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.