doc: add minutes for 08 april 2020

PR-URL: nodejs#377
Reviewed-By: Michael Dawson <Michael_Dawson@ca.ibm.com>
Reviewed-By: Chengzhong Wu <legendecas@gmail.com>

wg-meetings/2020-04-08.md

@@ -0,0 +1,95 @@
+# Node.js  Diagnostics WorkGroup Meeting 2020-04-08
+
+## Links
+
+* **Recording**: Unavailable
+* **GitHub Issue**: https://github.com/nodejs/diagnostics/issues/374
+* **Minutes Google Doc**: https://docs.google.com/document/d/1hD30I0WsXxOrDDp-2RUk1Bb4T5lsMArwYoAwA_021lc/edit
+
+## Present
+
+* Harshitha K P (@harshithakp)
+* Peter Marton (@hekike)
+* Matheus Marchini (@mmarchini)
+* Chengzhong Wu (@legendecas)
+* Stephen Belanger (@Qard)
+* Gireesh Punathil (@gireeshpunathil)
+
+
+
+
+## Agenda
+
+## Announcements
+
+*Extracted from **diag-agenda** labelled issues and pull requests from the **nodejs org** prior to the meeting.
+
+### nodejs/diagnostics
+
+* COVID-19 and WG sittings [#370](https://github.com/nodejs/diagnostics/issues/370)
+  * Should we leave open or close
+  * Leave open during the pandemic or least for couple of more weeks until people are adjusting
+* Action Items in meeting notes (hard to track)
+   * Stephen: +1
+   * Stephen: action items are tied to people but they can be busy and it can feel bad
+   * Stephen: would defer more to champion structure, make sure to document person with time the person could dedicate to it: like in the next month I’ll work on this
+  * Peter: how could we track champions and related tasks
+  * Stephen: we could carry over action items to next doc and go through them
+  * Gireesh: I think it’s also part of the action items are in the meeting notes
+  * Stephen: should we track them in GH issues?
+  * Peter: is there a concept of sub tasks
+  * Stephen: there are checklists, could work
+  * Peter: i’ll attempt to collect old assigned from meeting notes and convert them GH check lists to corresponding issues
+* proposal: elevate diagnostic report to tier1 [#369](https://github.com/nodejs/diagnostics/issues/369)
+  * Last time we agreed we are doing it
+  * Gireesh: we agreed to keep issue open for a week, one week is passed now, I’ll open a PR to raise the tier
+  * Gireesh: failing tier1 tests should block release, we should discuss tier with release group (Matheus is taking point on this)
+  * Gireesh: I don’t think the two tasks are tied together
+  * Peter: agree
+  * Gireesh: I’ll follow with report tier 1
+  * Gireesh: what’s the relationship between node-report and diagnostics-report for specific versions. Node-report is still used in 10 which has no diagnostics-report.
+  * Peter: i guess when 10 goes out of LTS we can make this change
+  * Gireesh: I think so
+* reportVersion semantics are not defined [#349](https://github.com/nodejs/diagnostics/issues/349)
+  * No GH issue activity
+  * Gireesh: conversation stuck on defining when to bump version, like what is a breaking change
+  * Gireesh: last time Matheus suggested to use schema change for breaking, but it’s not clear how to map schema to versioning yet, Matheus will propose how to map schema to versions
+* Proposal to drive Diagnostics WG initiatives through user journeys [#295](https://github.com/nodejs/diagnostics/issues/295)
+* Diagnostics "Best Practices" Guide? [#211](https://github.com/nodejs/diagnostics/issues/211)
+  * No progress since last meeting
+* \[async_hooks\] stable API - tracking issue [#124](https://github.com/nodejs/diagnostics/issues/124)
+  * Stephen: i can work full time in next three months due sponsorship with Datatog
+  * I’m planning to do couple of experiments to improve things:
+  * After callback native-js makes three travels, will try to optimize similar to timers
+  * Async resource is represented on JS side (async pooling) it has to currently communicate back to native side even if nothing needs it. Moving after to JS side would help with it and step over all the extra cycles.
+  * I’m also looking to move as much as possible to JS side from native
+  * Because of async resources we can’t do everything fully on native side and today we have expensive communication between js and native sides.
+  * Things that doesn’t uses libuv handle like outgoing request which is initiated from JS, with a JS side wrapper we could avoid native side overhead
+  * Some other V8 related things, I would need to dive into V8 more, I haven’t done before
+  * (async_hook receives events from promise_hooks in v8)
+  * Thenables doesn’t work on await, but works on Promise.all
+  * Currently if you try to use async_hook hooks it doesn’t bind all of them on native side.
+  * Matheus: in 2018 we discussed to change Promise hooks to do it on JS side instead of native promises, do you plan to take a look?
+  * Stephen: potentially, depends on time
+  * Stephen: async hooks was designed as this event bus, but some events require more complicated things
+* Zones proposal
+  * legendecas: I have a proposal around zones which can be related to async_hooks
+  * Stephen: I was part of original zone proposal, but it’s a controversial topic
+  * Stephen: due the history of features I think it’s going to be slow but I think we should push on it
+  * Matheus: any proposal should take account ongoing proposals
+  * legendecas: I could jump on this
+  * Matheus: there are multiple related proposals floating around which should be considered
+  * Stephen: when original zone proposal come out task scheduling was opinionated, now we have microtask queue which can make proposal more welcomed
+  * Matheus: current champion of agents is willing to “monitor” task scheduling
+  * Matheus: I’m on TC39, I’m willing to help to evaluate it
+  * Matheus: should we discuss on Zone repo or in diagnostics repo
+  * legendecas: old Zone repo is closed, maybe I can use diagnostics repo
+  * Stephen: diagnostics repo could raise awareness that someone is working on it
+  * ---> running out of time
+* Node CPU Profiling Roadmap[#148](https://github.com/nodejs/diagnostics/issues/148)
+  * Postponed
+
+
+## Q&A, Other
+
+## Upcoming Meetings