►
From YouTube: 2021-02-27 meeting
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
B
Good
morning,
I'm
I'm
still
like
fighting
with
my
pm
counterpart,
trying
to
guide
someone.
A
Great,
I
feel
like
stuff
is
starting
to
gel
up
as
far
as
these
initial
projects.
A
B
Yeah
hi
morgan
morning
and
I
I
created
the
metrics
project
tracking
on
the
open
time,
tray
top
level
projects
and
I
also
see
like
they're
they're
more
creative.
So
we
probably
can
start
to
practice
that.
A
B
A
So
we're
here
right
now,
coming
through
march
here
so
march,
we're
going
to
be
doing
the
data
model
and
also
beginning
the
api
design
and
then
we're
expecting
to
hit
the
prototyping
phase,
the
build-out
phase
in
may
and
then
like
a
beta
phase
in
september,
with,
like
everything
finalized
at
the
end
of
november.
A
Something
roughly
like
this.
I
would
love
a
review
of
whether
this
is
the
right
way
to
drive
yeah,
yeah,
okay
and
then,
as
far
as
ordering
the
the
other
initiatives.
A
After
some
discussion,
I
went
with
for
march
working
on
the
convenience
api.
This
is
because
this
requires
very
little
specification
work.
Language
work.
I
think
maintainers
can
just
get
started
on
this
and
we'll
write
some
stuff
in
the
spec
to
help
them.
So
we
can
focus
on
this
right
away.
A
So
right
now,
I'm
going
to
ask
the
maintainers
from
march
to
work
on
this
layer
as
their
primary
primary
activity.
Besides
supporting
supporting
beta
or
sorry
supporting
1.0,
then
we
start
getting
into
some
kind
of
tiered
phases,
so
in.
A
This
should
be
march
over
here,
but
next
step
will
be
like
a
simplified,
install
installation
experience
having
one
month
to
flesh
this
out
as
a
sort
of
like
design
and
discussion
phase
and
then
that
moves
into
the
maintainers
actually
then
going
and
implementing
a
simplified
experience
or
cleaning
up
their
experience.
Obviously
they
can
start
earlier,
but
I
just
want
two
months
for
that
and
trying
to
stack
up
the
as
they
move
on
to
implementing
one
thing.
A
A
A
And
then
so
that
would
be
a
month
of
design
there
kicking
off
in
april
and
then
a
month
of
people
implementing
that
at
the
same
time,
we
start
to
design
our
instrumentation
ecosystem
and
solve
all
the
problems
that
we
need
to
tackle
there
and
then
instrumentation.
Build
out
is
just
going
to
go
on
for
the
rest
of
the
time,
because
we're
going
to
have
to
build
out
tracing.
A
We're
going
to
have
to
build
up
metrics.
The
one
thing
I
realized
we
need
to
add
to
this
chart
is
internships
because
they're
coming
in,
and
that
would
be
a
good
thing
to
know
on
here.
So
like
may,
through
august,
we'll
have
interns
which
makes
me
realize.
Actually
this
instrumentation
thing
we
should
bump
this
sooner,
maybe
one
month
sooner,
so
maybe
switch
that
with
diagnostics.
A
Just
because
I
think
instrumentation
is
a
really
good
thing
to
get
interns
to
work
on,
and
so
we
should
be
prepared
to
catch
them
when
they
arrive
in
may.
We
shouldn't
be
discussing
this
stuff
in
may,
because
they're.
A
B
Yeah,
exactly
and
and
also
like,
I
I
think
we
might
want
to
clarify.
Is
this
just
focusing
on
the
spike
itself,
or
we
expect
certain
like
clients
to
be
ready.
A
Right
so
that's
the
when
I
say
here:
design
versus
beta
there's
for
each
one
of
these
projects.
There's
one
phase
I'll
put
these
on
the
same
line,
but
it's
like
the
lighter
color,
the
darker
color,
lighter
color
is
design
the
darker
color
is
everybody
goes
and
implements
this,
so
this
lighter
color
is
figure
out
what
the
maintainers
should
do
get
agreement
and
then,
let's
all
build
it
out.
A
I
don't
think
there
needs
to
be
a
lot
of
cross
project,
build
out
interaction,
especially
for
the
convenience,
api
and
installation
experience,
so
I
do
think
it
should
be
feasible
to
implement
this
stuff
in
a
month
if
maintainers
are
all
focused
on
doing
just
that,
likewise,
the
design
goals
for
like
a
simplified
installation
design.
I
don't
know
like
this,
isn't
like
trying
to
design
the
metrics
api
right
like
this
is
more
just
clarifying
with
each
other.
What
what
should
this
be?
Do
we
want
to
have
a
config
file?
Should
that
be
yaml?
A
Should
that
be
like
python,
ini
and
other
native
things
stuff,
like
that?
What
what's
actually
problematic
and
annoying
that
you
know
like
like
doing
an
audit
of
each
system
could
happen
in
this
month,
but
getting
into
those
details
now,
that's
kind
of
so
I
want
to
clean
this
up
a
little
bit
and
bring
it
back
to
the
maintainers
on
monday
and
then
ask
people
to
just
kick
off
with
the
convenience
api
work.
A
So
that's
that's
where
I'm
at
and
then
yeah
I've
started
to
just
add
these,
as
as
projects
to
to
the
board,
since
no
one
seems
to
be
blocking
any
of
these.
These
projects.
B
I
have
a
small
question
about
some
of
the
details.
Probably
we
don't
have
to
dive
into
every
detail,
but
just
quickly
ask
this
question
and
to
get
some
idea,
for
example,
the
convenience
api.
I
expect
once
we
have
like
matrix
api
or
the
other
api,
they
will
need
the
convenience
api
as
well
yeah,
so
this
probably
is
just
scope
to
to
trace
and
later
we
might
need
to
that's
correct.
This
is
for
our
current,
stable.
A
Apis
there's
things
we
can
do
to
make
this
less
code
for
the
application
developer.
That's
like
annotations
and
languages
that
support
annotations,
javascript
ripped
out
anything
that
looks
remotely
convenient.
I
don't
totally
understand
their
motivation
here,
but
you
know
if
you
look
at
like
open
telemetry.js.
B
A
Yeah
yeah
here
it's
like
right,
like
here,
just
to
give
an
example
like
currently
like
they've
removed.
Any
thing
you
can
do
with
like
with
span
like
currently
like
setting
a
span
to
active,
is
like
this.
C
A
Seriously,
someone
defending
this
as
being
easy,
so
so
there's
like
a
desire
to
be
like
why
don't
we
just
add
with
span
back
that
takes
a
callback
and
a
span
like
come
on
yeah,
so
some
stuff
like
that
like
this,
is
good
to
have
under
the
hood
like.
I
think
it's
fine,
if
this
is
how
it
works
under
the
hood
to
interact
with
context
objects
but
like
the
end.
A
Deal
with
that,
like
no
one
needs
to
deal
with
that,
so
so
that
that's.
This
is
an
example
of
stuff
and
they're
kind
of
saying
like
well,
there's
nothing
in
the
spec
that
says
we
have
this,
so
we
removed
it
all
kind
of
like
well
all
right,
I'm
going
to
add
something
to
the
spec
that
says,
you
know,
create
a
single
package
that
has
convenience
methods
for
all
the
all
the
stable
apis
right.
So
just
one
package
it
can
have
separate
objects.
A
You
know
trace
context
whatever,
but
just
one
place
where,
if
your
application
developer
there's
something
simple
and
then
if
you
can
take
those
same
functions
that
are
available
as
functions
and
turn
them
into
like
annotations
that
take
the
same
arguments
or
whatever
the
equivalent
is
in
your
language.
Like
do
that,
and
I
think.
A
Need
to
be
universal
across
every
language,
because
they're
high
level
functions,
you
can't
you
can't
mess
up
the
you
can't
mess
up
the
tracing
model
by
adding
these
as
long
as
they're
they're,
just
high
level
functions
that
that
call
these
lower
level
functions.
They
only
exist
in
api.
In
other
words,
they
don't
have
there's
no
spec
layer
here.
B
Yeah
and
for
diagnostics,
I
have
a
probably
no
need
to
cover
all.
I
got
the
rough
idea
for
diagnostics.
I'll
give
one
example
like
in
openclimate.net,
like
folks,
have
something
like
this:
it's
already
using
the
error.
Logging.
A
Yeah
yeah
and
diagnostics
is
like
it's
one
part:
how
do
we
log
errors
like
and
there's
another
part,
which
is
how
do
we
just
get
a
readout
like
what
is
useful?
If
I'm
gonna
support
someone
they're
gonna
come
into
slack
and
say,
like
I
have
a
open,
telemetry
problem:
is
there
either
on
the
command
line
or
within
the
application,
a
way
to
ask
a
person
to
do
a
setting
where
they
can
get
a
printout?
They
can
copy
paste
exactly
yeah
yeah.
B
This
kind
of
stuff,
so
there's
a
simple
configuration
file
that
you
can
put
in
any
application
using
opentimetrade.net
and,
and
you
specify
the
behavior,
so
I
I
think
probably
I
can
ask
cgo
to
share
some
of
the
requirements
originally
like.
You
can
always
turn
on
diagnostics,
without
having
to
worry
that.
Turning
this
on
will
blow
your
application
away
and
also
the
log
will
have
an
upper
limit
and
the
performance
consideration
and
what
should
go
into
the
log?
What's
the
velocity
level
and
also
how,
if
you
want,
how
would
you
turn
those
into
logs?
B
I
know
this
is
a
hard
topic
in
internet.
We
don't
have
that
struggle.
We
have
to
wait
for
some
logging
api
space,
but
for
c
plus
plus,
because
we
don't
have
a
famous
login
api.
We
have
to
work
with
the
login
seg
to
see
what
the
logging
api
should
be.
Instead
of
inventing
some
random
logging
api
yeah,
it's
kind
of
there's
already,
I
logger
being
there
10
years,
there's
no
point
to
invent
yet
another
login
api.
So
we
use
this
and
we
can
pipe
this
into
the
into
the
I
logger
api
yeah.
A
Very
exciting
either
I
I
think,
there's
like
two
a
couple
things
there
for
the
spec
or
sorry
for
logging.
The
first
part
is
just
getting
the
logging
protocol
stable.
If
we
have
the
logging
protocol
stable,
then
we
can
at
least
start
logging
our
own
information.
This
is
another
part
of
diagnostics.
There's
one
part
is
like.
I
want
local
diagnostics
right,
like
things
are
truly
messed
up,
there
has
to
be
a
way
to
print
logs
out
the
local
file
or
the
screen.
A
Also,
I
want
some
diagnostic
information,
but
then
there's
like
there's
going
to
be
issues
happening
during
runtime,
where,
like
assume
open
telemetry,
is
your
system
for
collecting
information
about
your
system,
and
so
when
there's
problems
with
open
telemetry,
it
needs
to
report
through
its
own
system,
assuming
it's
running
right.
So
these
are
things
like
dropping
spans
users.
Making
api
calls
that
are
causing
exceptions
that
we're
ignoring
like
that
kind
of
stuff,
and
so
we
need
a
design
for
like
what.
A
What
does
that
stuff
look
like
and
is
that
stuff
that
comes
out
of
just
like
our
log?
Are
those
these
just
logs
like
any
other
logs,
and
they
come
through
our
logging
protocol?
That
would
be
a
place
to
start
generating
logs
without
even
needing
a
logging
api
finished.
It's
just.
How
does
open
telemetry
internally
log
its
own
behavior?
A
So
so,
and
then
I
think
related
to
that
is
what
john
brought
up,
which
is
like
we
don't.
Even
we
haven't
really
actually
specified
what
we
do
in
some
of
these
error
cases
and.
C
A
We
don't
want
to
do
is
like
throw
exceptions
that
basically,
to
my
mind,
means
defining
like
how
we
log
what
happened
and
do
we
like
drop
things
on
the
floor
or
like
what
happens.
You
know
in
these
scenarios,
so
so
I
feel
like
that.
Diagnostics
does
have
some
meat
meat
to
it,
which
is
why
it
would
be
great
for
someone
to
to
tackle
tackle
those
different
problems,
but
this
looks
great.
I
I'm,
I
do
think
the
microsoft
team
has
like
a
handle
on
this.
I'm
glad
to
see
cjo
has
stuff
in
there.
B
B
A
B
Yeah
so
probably
like,
like
ping
me
or
like,
I
can
wait
until
next
tuesday,
if
you
believe,
like
all
the
top
level
projects
or
the
timeline,
that's
clear,
we
have
in
our
mind
and
once
we
have
that
I'll
look
at
the
top
level
projects
and
change
all
our
existing
issue.
B
A
A
A
My
proposal
is
just
unassigned
assigned
you
know
pr
ready
or
available
and
like
complete
something
that
that
simple.
But
if
you
have
other
ideas,
you
know
I'd
love
to
hear
them.
A
A
To
be
the
the
steps
we
care
about
like
it's,
and
we
should
like
work
on
things
get
assigned
to
to
people
when,
when
they're
actively
beginning
work
on
it,
so
the
assignee
list
is
like
stuff.
That's
that
someone
says
they're
gonna
do
it,
but
the
work
hasn't
shown
up
yet
so
we
know
who
to
check
in
with.
We
should
avoid,
like
just
pre-assigning
things
to
people,
so
that,
like
like
avoid
this
situation
where,
like
josh,
has
like
20
metric
things
assigned
to
him
he's
not
actually
working
on
them
all
just
you
know.
A
So
we
know
like
what
what's
actually
nobody's
working
on
right
now
and
if
someone
can't
keep
up
with
one,
then
it
moves
back
to
the
prior
thing.
Yeah.
B
If
we
can
make
it
very
explicit,
I
I
I
think
I
I
saw
the
trouble
like
last
year,
I
was
put
as
a
mentor
for
for
a
intern
workstream
for
aws,
where
I
already
have
seven
aws
interns.
So
I
totally.
C
B
Not
have
it
another
five
and
I
haven't
agreed
on
that.
I
like
please,
don't
expect
anything,
and
this
year,
when
I
came
back
from
vacation,
I
noticed
I
was
listed
as
the
approver
for
the
open
time
to
donate
contribute.
Well,
there
are
a
lot
of
things.
I
have
no
idea
like
mass
transit
or
the
components.
Some
people
ask
hey,
please
review,
and
I
told
them
please
remove
my
name.
Yes,
when
I
did
my
name
without
me,
knowing
so
so,
we
have
to
make
make
that
clear.
First
and
b,
yeah.
B
A
Yeah-
and
we
just
had
a
pattern
before
I
think
of
feeling
like
it-
was
important
to
find
assignees
early
and
quickly
to
issues
like
especially
in
the
specs,
so
it's
like
spec
comes
in,
it
gets
labels
tagged
on
it
and
then
who
should
be
assigned.
Someone
should
be
assigned
at
all
times,
and
I
think
we
should
move
just
move,
move
away
from
that
and
only
assign
people
when
they're
gonna
work
on
it.
Yeah
and
it'll
be
the
tc's
job
to
like
keep
track
of
all
of
this
stuff.
A
I
feel
like
we're
doing
this
job,
but
I
want
the
tc
to
do
more
of
it,
but
yeah
we,
it
should
just
be.
I
think
that's
that
step
is
like
muddled
in
the
past
and
yeah,
if
someone's
assigned
to
it,
they've
been
asked
and
said
yes
and
ideally
even
we.
I
don't
know
how
we
put
like
due
dates
or
something,
but
it's
nice
when
someone
gets
assigned
to
it
to
at
least
like
post
in
the
issue
when
they
expect.
C
A
A
C
A
B
A
Like
well,
bob
said
he
would
do
it.
He
said
he'd
do
a
thing,
but
he
said
he
wouldn't
deliver
it
till
next
week.
So,
okay,
we
don't
need
to
poke
him.
You
know
about
it
whatever
it
is,
and
then
last
but
not
least,
if
we
can
get
a
bot
that
can
actually
move
issues
through
this
through
these
project
boards,
so
that
we
don't
have
to
manually
because
that's,
like
the
other
part,
that's
a
pain
in
the
ass.
It's
like
I
don't
want
to
have
to
like
when
a
pr
shows
up.
A
A
So
the
project
board,
you
don't.
Actually
you
don't
actually
move
things
manually
to
the
project
board.
It's
just
a
state
of
truth
about
the
source
of
truth
about
the
project
project
boards.
Don't
do
that
automatically
right
now,
which
is
really
dumb.
B
I
know,
even
in
microsoft,
a
lot
of
people
they're
struggling
internally,
there's
a
viral
studio
online
system,
which
is
very
comprehensive
and
can
manage
like
10
millions
of
work
items.
I
have
no
idea
like
why
why
we
should
have
10
millions
of
anything
like
it's
just
impossible
to
manage
on
the
other
side.
B
Github
in
microsoft,
for
a
lot
of
products
is
prohibited
because
compliance
you
know
like
like
last
year,
there
was
a
source
called
league
for
microsoft,
like
500
people
was
privately
called
leaked
on
github,
it's
like
sliding
on
its
own
face,
so
so
yeah.
We
have
this
weird
process
of
some
compliant
thing
sitting
in
university
online
and
some
of
the
things
that
leading
towards
open
source
in
github
and
some
speaking
in
between
I
get
a
private
report
from
microsoft
and
people.
People
did
crazy.
B
Like
sync,
they
did
some
bot
like
try
to
do
the
nightly
like
merge
from
one
ripple
to
another
and
it's
all
crazy
yeah.
That's.
A
This
is
actually
my
hope
I
mean
I
do
feel
like
microsoft
is
being
a
good
steward
to
a
lot
of
this
stuff,
and
I
I
actually
hope
that
yeah
microsoft,
trying
to
dog
food
github
will
lead
to
like
the
the
project
management
tool
side
of
it
and
like
how
do
you
do
a
big
project
on
on
github
side
of
it?
They.
A
A
Yeah
yeah
anyways
is
there
other
other
business
to
attend
to
as
far
as
triage
again,
just
to
remind
us,
we
do
need
to
triage
these
nearly
250
open,
spec
issues.
I
think
we
have
to
get
our
new
stuff
sorted.
I
just
want
to
remind
us
that
we
need.
We
need
to
figure
out
how
we're
going
to
cut
some
scope
there
or
something
is.
A
A
Least,
like
audit,
what
the
hell's
in
there
and
try
to
figure
out
like
like
how
do
we
get
to
under
100
issues
is
like
a
goal
and
always
ideally
stay
under
100
issues.
It's
there's
something
about
these
projects
where
they,
once
that
starts
to
spin
out
of
control.
It
leads
to,
like
I
said
things,
just
get
lost.
Yeah.
B
I
will
try
to
enforce
the
ownership,
for
example
like
if
we
have
a
clear
understanding
of
the
key
milestones,
the
projects
and.
A
B
We
should
change
the
issue
template
and
I
want
to
make
sure
that
the
new
issues
come.
They
will
go
to
a
specific
bucket
and,
for
example,
if
this
matrix
api
sdk,
I
should
go
and
take
take
care
of
that
if
it's
like
convenience
api
someone
else
who
owns
the
convenience
asia.
So
we
stop
having
these
new
issues
coming
sitting
nowhere
and
then
like
whether
it's
200
300,
it's
just
existing
stuff
and
it's
just
a
matter
of
time
and
effort.
So
at
least
we
stop
bleeding
and
we
come
back
and
solve
the
existing
issues
right.
A.
A
A
Do
we
stop
making
those
issues
like?
Do
we
just
say,
look
anything.
That's
like
a
would
be
nice
like
we
don't
actually
record.
We
only
record
issues
that
we
think
we're
gonna
actually
get
through,
because
I
it's
just
not
it's
not
clear
that
it's
helpful
for
us
to
like
write
down
and
I
think.
C
B
This
is
what
I
observed
in
microsoft,
like
when
a
lot
of
folks.
They
don't
have
experience,
they
open
some
github
real
power
like
they
got.
They
got
a
lot
of
issues
that
they
cannot
afford
to
respond.
So
later
they
decided
to
close
the
issues
entirely
and
only
open
that
for
internal
folks
and
what
they
did
instead
is
tell
people
hey.
If
this
is
a
support,
ask
you
use
microsoft
product.
B
You
should
go
to
the
support
center
and
like
make
a
call
or
something-
and
if
this
is
a
feature,
ask
here
goes
the
place,
so
they
have
a
user
voice
and
they
have
the
vote,
and
then
it
makes
the
pm
job
easier,
because
people
can
search
if
they
have
some
ask
they
can
search.
Oh
that
I'll,
add
one
vote
to
it
and
we'll
take
the
top
10
votes
right.
Do
the
job.
A
A
So
if
you
have
feedback
around
these
issues
like
come,
talk
to
us
like
come
into
slack
and
talk
to
us
or
come
like
open
an
issue,
but
for
things
that
are
like
outside
the
scope
of
the
stuff
that
we're
talking
about
right
now,
we're
actually
just
going
to
say
no
thank
you
and
close
respectfully
close
those
issues
for
the
time
being,
and
maybe
just
track
high
level
stuff
somewhere
else,
but
just
try
to
it's
just
not
not
not
feel
like.
A
Oh,
we
want
to
keep
track
of
this
idea
because
it
was
a
good
idea,
but
we're
not
going
to
do
it
right
now
instead
be
like
it's
a
good
idea.
So
when
we
get
to
this
phase
we'll
you
know
like
you
can
raise
it
again
or
someone's
gonna
raise
it
again,
but
we
don't.