シャドーイング練習: 2022-07-25 Analytics Section Meeting - YouTubeで英語スピーキングを学ぶ
C1
Hello everyone and welcome to the July 25th version of the analytics section meeting.
315 文
文が短すぎたり長すぎる場合は、Editをタップして調整してください。
1
Hello everyone and welcome to the July 25th version of the analytics section meeting.
2
Got a number of topics here you can go through.
3
Matt got you from monitor observability.
4
I believe you were having a conversation with Kenny,
5
and you might have a ton of insight that we can benefit from from from your months spent.
6
Would you like to start?
7
And actually, before we start,
8
I did want to give one little bit of preamble.
9
So first of all, Matt's awesome.
10
And I'm excited to have him chat with you all.
11
I do want to say like a principle of mine in thinking about obviously both parts of the application
12
that we're building in observability and in analytics have a lot of potential overlap,
13
including are using likely going to use a common data store in ClickHouse.
14
But I want to try to operate them as independently as possible.
15
So what we should think about when Matt's explaining what they've learned is lessons learned
16
that we can apply going forward as opposed to like trying to couple the groups to say like,
17
oh, one group work on this part and we'll use it
18
and the other group work on another part because these are all.
19
So both early investments for GitLab and so it's important
20
that we are structured to move quickly rather than necessarily be coupled together as groups.
21
So I just want to make sure I give that caveat before we start.
22
Makes sense.
23
Cool.
24
Yeah.
25
Thanks, Kenny.
26
Thanks, Dennis.
27
I'm Matt i'm from the monitor observability group i came over to get lab as part of the obstrace acquisition
28
and in the last six months we've really been building um
29
a lot of what i think you folks are probably going to end up having to build
30
and so that's why i wanted to you know sort of bridge this discussion a little bit
31
because um i you know and i think that uh it's a good principle
32
to operate by you know having kind of you know separate um dependencies
33
and trying not to you know overlap too much um
34
but you know in the last six months we have invested
35
a lot in i think what you guys are gonna have to redo
36
so i you know i think that there is you know we're happy here to
37
if you want to stay you know a little bit more independent we're happy to help in any way we can
38
um
39
but there is a lot of benefit too from you know sharing some of what we've built at the same time
40
we've built a lot of automation around kubernetes um you know
41
infrastructure life cycle of kubernetes clusters with click house on top
42
of it click house scaling limits quota management inside of click
43
house end-to-end testing of the entire system authentication on the apis um you know
44
that connect back with the git lab instance um you know there's there's loads
45
and loads and loads of things
46
that uh you know for the back end a lot of this stuff is not seen by a user
47
but they will be things that an operator of git lab for self-hosted or for our own internal sre team
48
they will have to interact with the system and so i think
49
that from a product standpoint too
50
and user experience you know we don't want to diverge too much i don't think and have you know
51
many different deployment models of you know components that then sres
52
and operators have different ways of managing and different run books
53
and so i think there's a lot here that we you know
54
if we do keep infrastructure separate uh
55
and you guys go down the route of kind of building
56
your own stuff in parallel we should at least make sure
57
that we're building things
58
that align with some principles i guess you know to make it easier for those folks
59
that are going to be you know managing the life cycle of of git lab
60
and its entirety of observability and analytics analytics and whatnot
61
so um we i just wanted to start the conversation um
62
and kind of you know introduce things for a start as a way
63
that I think that if you did want to use what we have,
64
I think it really could accelerate where you guys are knowing the effort
65
and time we've had to put in over the last six months.
66
But, you know, happy to leave that over to you guys to decide,
67
you know, where you'd like to use some things and help and just start the dialogue.
68
Yeah, no, that's great.
69
I mean, I think there's definitely going to be some learnings we can leverage there.
70
And, you know, to your point,
71
if we operate by the same,
72
you know, principles, then at least SREs won't have to worry about managing two very similar architectures,
73
two different ways, right?
74
There are some decisions we are discussions we haven't I had
75
in terms of like how we're going to like organize the tendency of
76
or how we're going to like set up this whole kind
77
of tracking box as we're currently calling it in terms of JetSea,
78
ClickHouse and Cube in terms of like,
79
you know, obviously we're going to operate in multiple regions.
80
So that's going to be useful in terms of how you're managing the provisioning of Kubernetes there.
81
But then, yeah, when it gets down to like per customer,
82
especially on .com, I'm not sure if,
83
well, we haven't gotten there yet,
84
but I'm sure there will be something we can leverage there.
85
I guess in terms of like where we could start to like see where this is happening,
86
is there a repository or does there documentation we can start to like look at,
87
start exploring this? Or?
88
Absolutely.
89
You start.
90
Yeah, I'll drop some links in to the notes here with the repositories that are relevant,
91
but you can go through it.
92
You can see all the code there.
93
You know, your model around having these,
94
you know, boxes that you're deploying into different regions
95
and different clouds all fits our exact same model too um
96
the tendency model is exactly the same um you know hence
97
kind of why i think personally i think it doesn't make a lot of sense to go
98
and rebuild all that um you know
99
because it's a i think it's a it's a massive undertaking um
100
and you need you know to have a team that has you know a lot of experience i think in the
101
realm of kubernetes and you know i think
102
that we can probably help you get a lot further there
103
so you can focus on more of the user facing pieces
104
and get that you know demo up and running for
105
or not the demo but you know the early alpha up
106
and running much faster right um
107
so there is um you know i sort of was seeing it as a way
108
and i talked about it with sebastian a little bit too uh he's out of office here today
109
but you know i think there is a model where we
110
could work on collaborate on the same code um without it slowing us down
111
and getting it our way because we have all the
112
pieces in place that you guys could just fill in the bits
113
that you need and deploy without you know needing us to do anything for you
114
so i think that you know might just get you a whole lot further Rahid.
115
No, great.
116
Yeah, that was the principle I don't think I very artfully said earlier.
117
I, like, GitLab, everyone can contribute.
118
Typically, we operate in one code base.
119
So I was thinking about it in those terms.
120
What I didn't want to create is,
121
like, if analytics need something,
122
they need observability to ship something first.
123
But if we're all operating from the same code base,
124
I think that's a great mental model.
125
Yeah, a shared framework
126
that also we can hopefully contribute back to to help you
127
know you know also just bring some light to y'all
128
but um yeah if we can if we can operate it in such a way
129
that we both benefit from it but we can contribute from shared work
130
or benefit from shared work but then both contribute to it hopefully have
131
that backwards then that'd be great um
132
because there's there's a bunch of different topics
133
that i have running in my mind in terms of how
134
is this all going to be maintained i've you know heard
135
from multiple people like click house requires to some degree of like you know management especially and
136
if we're going to have to have you know n number
137
of click house instances i don't even know the first thing
138
about how the things we're going to be looking at there
139
so if you've already done some work around
140
that that's already a huge boost but then yeah everything from testing run books all
141
that kind of stuff like you know product analytics currently has two engineers
142
so there's a lot of work to be done even just from a feature perspective standpoint um
143
so yeah even getting into the infrastructure part of it any
144
anything you've got will at least you know be a huge starting point um Yeah, for us.
145
If you're saying like where our tendency miles the same,
146
then there's probably a lot of you can just leverage and bootstrap off of.
147
So, yeah, it's appreciate y'all bringing it up
148
and have coming to talk about it
149
and we'll be happy to explore all the as many links you can throw our way. Of course.
150
Yeah, I think it's, you know,
151
everybody comes out stronger and better off if we can all collaborate on the same code base.
152
And we have, Kenny mentioned in here,
153
I think he's got to take it down a little further.
154
We do have a contract with Clickhouse Inc.
155
And we've been engaged with Clickhouse Inc now for a number of months,
156
working with them on a number of issues,
157
getting reviews of how we manage Clickhouse, things like this.
158
And so, you know, we're building up that muscle in the team.
159
And we've also hired some ClickHouse experts in the team.
160
So, you know, we're kind of getting more acquainted with running ClickHouse and,
161
you know, it's quirks.
162
So, you know, we can certainly help there too, I think.
163
That's amazing.
164
Cool.
165
I'll drop these links in here for the main repo.
166
We're still under heavy development, of course.
167
So not everything's there.
168
Runbooks, you won't see those in the repo yet.
169
There's things, but you'll find issues too where we can follow up on things that are yet to be done.
170
The core tenancy model deployment,
171
lifecycle, and authentication, all that stuff is all in place.
172
Yeah, authentication will be huge.
173
Right now, everything's just out,
174
as Sam saw, it's just very development purposes,
175
very out in the open.
176
So lock that out, lock that up before we start deploying it.
177
Yeah.
178
Yeah.
179
Cool.
180
Awesome.
181
All right.
182
Kenny.
183
Cool.
184
Yeah.
185
Yeah.
186
It's a private channel, but it's a direct connection between us and ClickHouse that we pay for support from them for.
187
So at any point, Dennis,
188
you and team want access to that channel.
189
Let me know.
190
I mean, if you want Dennis,
191
I can just invite you to that channel right now so you can follow along.
192
Then feel free to invite others as needed. Will do.
193
Thanks for sharing.
194
Awesome.
195
Have a couple of short points and then we'll open it up for anyone else.
196
Sam is joining us as interim product manager for product analytics.
197
I worked with him.
198
I still work with him rather in compliance.
199
um yeah sam you want to do a little quick intro
200
yeah hey folks really excited to be here um i think
201
product analytics is going to be a really exciting space to to jump in
202
and i'm excited to go on
203
that journey with you all together um like dennis said i'm
204
on the compliance team the compliance group right now um started on
205
that group in 2021 i've been with gitlab since 2019 i'm
206
based in cincinnati ohio in the united states um over the
207
next couple of weeks i'd like to to put a coffee
208
chat together with each of you just to get to know you a little bit more personally,
209
let you get to know me a little more personally.
210
And we can also talk about some work stuff along the way.
211
So I'm sure I will have lots of questions and things to discuss with each of you in the days to come.
212
And I'm excited to get things started off here today.
213
Awesome, Sam.
214
Thank you.
215
Very cool.
216
And then my last point was just kind of just getting a little bit more,
217
well, not just Sorry, just one of the highlight kind of next steps.
218
So we're headed for product analytics in that group.
219
So I'm in writing mode,
220
just trying to organize a bunch of different epics for all the different categories and features,
221
but basically just kind of taking the notes from our proof of concept demo,
222
what we've been having discussions,
223
you know, online and offline.
224
and I'm just trying to get an idea of all the
225
things we've been thinking about what we want to do for things like product analytics,
226
instrumentation and whatnot, and then starting to nail
227
that down into more of an MVC in terms of what we really want to pursue for those.
228
But in terms of a more immediate focus,
229
I think discussions between Kenny,
230
Tim and I, we're looking at product analytics or more specifically,
231
you know, web and mobile analytics in terms of how just your general setting up instrumentation,
232
trying to see what default dashboard you want to kind of present.
233
And also just getting a little bit more productionized.
234
And so getting some instrumentation set up to be able to track maybe the handbook,
235
maybe our design library, maybe even .com,
236
just with a targeted audience of just,
237
you know, GitLab.org, just to be able to get more information there
238
and just start dogfooding and seeing what we want to really provide there in terms of your general web analytics.
239
So that's what's coming up for the product analytics group um yeah
240
and dennis i know uh i don't know
241
if you want to discuss here the like tactical process of taking all of the vision poc
242
and breaking that down into epics and issues
243
and kind of targeting what our initial first mvcs versus subsequent ones is a big task
244
so how can sam still ramping up so this should be on me this week but how can I help this week?
245
I know we have some time,
246
but is there something that I can start contributing to or reviewing or putting that on paper?
247
So, I mean, the main point
248
that I'm trying to push out is just to organize all the different areas so
249
that we at least have like it organized within our on GitLab.com in terms of like everyone can start editing
250
and having discussions there.
251
So I will drop a link in the section channel probably end of day today
252
or tomorrow to just have the general structure of all the different epics in terms of the features
253
and then start building out that list.
254
So if anyone wants to jump in and just start,
255
you know, even, you know,
256
all the wild ideas that we were discussing in terms of like what we could provide to the features,
257
like feel free to add that to the epics.
258
um there's no strict policy on who's uh who can do what
259
so any as
260
if anyone has anything i'd like to add in terms of
261
like things we should be thinking about uh for any of
262
the features we want to build out then um please do um that's all i'd ask
263
but yeah right now i'm just putting the general structure together just
264
so that um we have somewhere to record all this stuff so okay
265
and this is we can start to to to filter out
266
from there i think what i can do this week is
267
make sure I'm contributing to you know like our what's next section we've been updating
268
but it's still pretty vague I want to maybe try to get something very crisp
269
that will help if we can use
270
that discussion in an MR to like really target what is
271
you know milestones one through three specific focus going to be yeah I'll work on
272
that for sure yeah that's that's the general goal of um
273
I think what we'll be both be working on then is just like
274
or in Sam as well as he ramps up there's just
275
to give a little bit more of a clearer direction than yeah we want to build an analytics thing so yeah
276
yeah and i'm happy to help with either of those initiatives i'll be ramping up
277
but i'm still willing to contribute things this week
278
so don't uh feel free feel free to pull me into
279
those things i'm looking forward to it i want you to
280
tell me all of the the the tell me how poorly
281
i've organized all of it it's just gonna be perfect this
282
never never never is um cool well that's all I've got
283
as far as agenda anyone anyone else would like to share
284
anything um um Amanda I don't want to put you on the spot
285
but I um your comment about the visibility of the section thing I think seems material
286
but I didn't see anyone of a concern I didn't maybe completely understand it
287
but is is everyone aware of the change in visibility for the section group and comfortable with that change Yeah,
288
so TLDR, late Friday my time,
289
we had a report that the metrics dictionary wasn't accessible externally.
290
And it turned out it was because we moved the metrics dictionary project over to the analytics group,
291
parent group, and that parent group was private,
292
which caused only members of that domain to be able to see it.
293
So I made the parent group,
294
the analytics group, as well as the product intelligence subgroup and the metrics dictionary project,
295
all public visibility.
296
So visibility settings are least restrictive to most.
297
So you can still make things more restrictive within um but
298
if that's a problem then we can revert everything
299
and we can move metrics dictionary out to a separate patient cool good call okay
300
so the action for anyone who is like hey i wanted
301
something in there to be private is now you have to explicitly make it private
302
because it's not cool yeah no thanks for the call out
303
because i remember the visibility options being restricted and i i hadn't realized
304
that the well i i didn't create oh yeah i had the section group have had to be created
305
so it was just created as private
306
so we can be public by default there's no problem at all
307
but i think that's a good thing to call out cool thanks for the the quick action on
308
that amanda all right um anyone else would like to share
309
anything well i'm going to start uh advertising the job i
310
got the job descriptions done i have job racks open so we're hiring engineers
311
uh for for back in front of product analytics so
312
if you know anyone um i'll i'll drop the links in the agenda after this call
313
and then uh feel free to get some referrals i'd really much appreciate
314
that if there's nothing else everyone gets 10 minutes back enjoy your mondays enjoy your weeks
315
take care thanks everyone thanks guys take care
アプリをダウンロード
話したすべての文をAIが採点
TRENDING
人気動画
文脈と背景
この動画は、2022年7月25日に開催された分析セクションミーティングの様子を記録したものです。参加者たちは、モニターの可観測性や分析の分野での最近の進展について話し合っています。特に、マットが持っている知見がどのように日本の技術トレンドに役立つかを探求しています。彼の経験は、効率的なシステム構築やプロジェクト進行にとって重要なデータストレージ管理に関するもので、他のメンバーにも多くの示唆を与えています。
日常会話で使えるトップ5フレーズ
- 「運営を別々に行う」 - 異なるチーム間の業務を独立させる重要性を強調するフレーズ。
- 「レッスンを適用する」 - 過去の教訓を未来に活かすことの重要性。
- 「自動化を進める」 - システムの効率化と作業負担軽減を指示する言葉。
- 「相互作用を理解する」 - システムのインタフェースについての理解を深めるための表現。
- 「できるだけ早く進める」 - プロジェクトのスピードを上げるための目的意識を表す言葉。
ステップバイステップ シャドウイングガイド
この動画の内容は、技術的な議論が中心ですが、英語学習においても非常に有益です。以下の手順を参考にしながら、動画の内容をシャドースピーク及びシャドースピーチの練習に活用しましょう。
- 動画を視聴する: 最初に全体を通して見て、全体の文脈を掴みましょう。
- フレーズを拾う: 上記のような重要なフレーズを意識してリスニングし、メモしましょう。
- 部分的に再生する: 難しい部分を何度も再生し、耳で聞き取れるように練習します。
- 声に出して繰り返す: リスニングしたフレーズを自分の声で再現してみて、発音を改善しましょう。特に「英語の発音を良くする」ことを意識してください。
- YouTubeで英語学習を続ける: 他の動画も活用し、さまざまなアクセントに触れながら実践を重ねていきましょう。
このように、シャドウイングサイトを使用することで、技術的な会話に限らず、豊かなボキャブラリーと特定の業界用語が身につきます。英語学習者にとって、深い理解が求められる内容を扱うことで、語彙力や発音の向上が期待できます。
シャドーイングとは？英語上達に効果的な理由
シャドーイング（Shadowing）は、もともとプロの通訳者養成プログラムで開発された言語学習法で、多言語習得者として知られるDr. Alexander Arguelles によって広く普及されました。方法はシンプルですが非常に効果的：ネイティブスピーカーの英語を聞きながら、1〜2秒の遅延で声に出してすぐに繰り返す——まるで「影（shadow）」のように話者を追いかけます。文法ドリルや受動的なリスニングと異なり、シャドーイングは脳と口の筋肉が同時にリアルタイムで英語を処理・再現することを強制します。研究により、発音精度、抑揚、リズム、連音、リスニング力、そして会話の流暢さが大幅に向上することが確認されています。IELTSスピーキング対策や自然な英語コミュニケーションを目指す方に特におすすめです。