►
From YouTube: 2020-05-12 .NET SIG
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).
B
A
Man,
it's
getting
I'm
trying
to
get
used
to
all
the
whole
swapping
schedule:
6
o'clock,
10
o'clock,
of
course,
you're
specific,
though
right
here,
yeah
I'm,
a
bus.
If
ik
I
mean
it's
central,
so
6
o'clock,
it's
awkward
hours
for
me,
that's
ok!
It's
alright,
I'm,
just
a
firewall
I'm,
just
trying
to
listen
and
get
as
much
like
you
to
help
eventually
yeah.
B
I
think
I
think
I,
don't
know
if
you
saw
the
channel
the
discussions
about
the
because
this
thing
about
using
the
activity
source
I
think
it's
a
great
idea,
but
I
I
do
think
that
we
need
to
keep
the
open
telemetry
API.
On
top
of
it,
you
know,
I
I
need
to
be
convinced
that
we
should
kind
of
grab
the
API.
B
C
A
B
C
B
C
B
B
C
B
A
B
A
I
understand:
that's
well:
I
try
to
get
I,
try
it's
hard
for
us!
Midwest
earns.
You
know
in
the
middle,
the
United
States
to
make
our
way
that
direction
I
try
to
get
out
working
at
Microsoft
a
few
times,
and
it's
just
like
wow
your
interview.
Questions
are
very
academic
here
in
the
Midwest.
They
focus
more
on
personality,
less
on
academic
I,
see.
C
A
B
And
I
would
say
that
this
thing
about
interview,
at
least
here
on
the
west
coast,
I
say
that
it's
a
specific
skill,
you
know
kind
of
perform
the
interview,
it's
not
a
skill
of
her
job.
It
is
a
skill,
that's
related
to
a
job,
but
it's
a
specific
skill.
So
if
you
want
to
do
great
on
those
who
have
to
practice
for
those
yeah,
I
have
to
train
for
those
which
is
pretty
hard
once
you
are
really
fully
committed
to
the
work,
because
it's
like
hey
I
need
to
stop
what
I'm
doing.
B
B
A
I
get
so
far
along
in
my
career
I'm,
like
man,
I
gotta,
like
change
my
you
know,
mode
of
operation,
especially
if
you're
trying
to
like,
like
for
me,
I've,
been
in
a
business
application
development
for
more
than
there
are
about
15
years
now.
So
it's
like
switching
to
framework,
mind,
that's
what
I've
always
been
wanting
to
do,
but
I've
never
been
in
the
field
to
do
it.
So
then
one
course
open
telemetry
showed
up
like
yeah.
A
This
is
this
is
pretty
close
to
what
I've
been
trying
to
do
at
where
my
work
right
now
cuz.
We
we
have
our
own
in-house
logging
solution
and
and
we're
trying
to
like
leverage
what
open
telemetry
is
doing,
but
they're
not
I,
mean
I'm
trying
to
promote
it,
but
it's
in
the
state
of
that
awkward
state
and
well,
it's
not
quite
GA
kimi.
Can
we
start
building
stuff
on
and
what
vendor
are
we
gonna
use,
I
think
worth
New
Relic
right
now,
but
and
we
have
our
own
in-house
solution
like
yeah.
B
The
point
one
of
the
points
is
kind
of
trying
to
settle
this
question
about
the
vendor
kind
of
okay.
You
instrument
with
your
with
the
open
telemetry,
and
then
you
choose
yeah.
Actually
this
is
a
point
for
us.
That's
why
the
point
that
Splunk
kind
of
tries
to
do
win
even
they
set
up
that
there
was
before
try
to
do,
is
kind
of
a
you
are
not
getting
any
locking
we
found.
There
is
no
such
thing.
If
you
want
to
send
you
yet
you
from
me
choose
just.
C
A
Right
right,
that's
yeah!
That's
what
I
really
like
about
this
movement.
It's
different
different
world
going
to
open
source
from
closed
source
proprietary
world
I've
been
in
all
my
life,
pretty
much!
So
that's
why
I
like
about
the
the
talks?
It's
really
cool
to
see
this
many
vendors
get
together
and
talk
about
what
makes
best.
What's
the
best
choice
for
the
community,
so
yeah
I
see
other
people
are
showing
up
now,
like
we
gotta,
be
quiet
now,
Shh
I,
don't
know
these
other
people
here,
yeah.
D
D
B
I
think
there
is
a
lot
of
kind
of
questions
related
to
some
kind
of
the
the
middle
level.
Let's
say
of
the
use
of
actives
for
sentencing
I
as
I
mentioned,
guitar
I
think
we,
the
question
that
I
really
want
to
concentrate.
First
is
the
question
about
the
day
open,
telemetry
API
itself?
What
we
wanna
do
in
that
regard,.
F
F
Yeah,
so
let
me
just
like,
if
there
is
any
up
I'm
generally
asking
like
if
anyone
has
any
topics
to
discuss
about
instrumentation
or
anything
else,
please
speak
up
right
now.
If
there
are
no
topic,
we
will
move
into
P
source
and
activity.
So
that's
going
to
take
a
lot
of
time.
So,
let's
finish
any
other
topics.
First
I.
B
C
C
What
should
we
recommend
what
we
want
to
see
actually
in
open
telemetry
as
a
base
instrumentation
my
code
engine,
so
it's
ever
be
dated
of
orientation
engine
we
haven't
decided
yet
still
investigating
it,
and
in
fact
we
have
a
couple
new
folks
on
the
call
for
interested
in
the
same
topic
as
well
is
know
that
he's
from
give
deep.
He
is
a
guy
who
is
responsible
for
core
file
a
decade,
and
so
he
is
a
p.m.
where
Jessica
p.m.
so
in
this
past
accept
him
as
well.
I
I.
G
I
think
we
did
I
think
a
number
of
people
in
this
forum,
iÃve
seen
names
around
on
email,
threads
and
and
others
so
yeah
I've
well
I'm
on
the
team
that
has
owned
I
core
profiler
since
forever
and
I
was
sort
of
its
direct
owner
for
years
and
now
I'm,
sort
of
architect
II
role
in
the
space
and
I.
You
know
so
my
team
still
owns
it
so
anyway.
I
still
have
lots
of
knowledge
about
I,
core,
profiler
and
sort
of
figured
I
should
be
a
little
more
involved
in
this
discussion.
F
F
Mike
I
have
a
hard
time
figuring
out.
If
there
was
anyone
else
new
we
have
shown,
and
probably
Alan
was
here
last
name:
Alexis
I
turn
on
the
name,
how
to
pronounce
aid.
So
if
you
want
to
say
hi,
if
you
want
to
say
about
anything
which
you
are
putting
in
particular
looking
for
looking
for
guidance,
so
looking
for
areas
to
contribute
just
use
this
time,
the.
H
Fact
that,
like
I
mentioned
earlier
on,
they
get
a
chat.
This
is
chunky
hard
I'm,
really
coming
from
an
angle,
looks
like
in
a
consumer
for
you.open
telemetry,
APs
and
I
had
some
thoughts
around
activity
stuff,
so
I'm
kind
of
interested
in
participating
that
sort
of
discussion.
Okay,
it's
really
yeah.
It's
gonna
be
quiet.
Most
of
the
time
now,
yeah.
I
F
Yeah
welcome
nice
to
have
you
but
I.
Think
Ireland
wanted
to
ask
something,
but
did
you
get
like
Alan
like?
Are
you
like
having
any
further
questions
over
your
food
too,
because
you
are
asking
about
adding
instrumentation
for
so
the
adapters
for
G,
RPC,
client
and
I
hope
I
was
able
to
answer.
Do
you
have
anything
which
require
clarification
with
a
broader
group,
I.
J
F
Before
we
go
into
the
night
activity,
I
have
one
update
for
my
trig
so,
like
last
week,
polo
try
to
make
a
fix
to
Prometheus
exporter,
because
we
we
have
an
implementation
issue
right
now,
where
we
are
exporting
the
same
metric
with
a
different
value
but
overall,
like
in
the
same
scraping
this
behave.
The
behavior
for
Prometheus
is
undefined
in
the
scenario,
but
unfortunately,
we
didn't
really
close
on
that.
I
have
a
peer
which
actually
introduces
memory
and
stage
in
do
batcher.
I
will
be
submitting
it
shortly.
F
I
would
like
for
quite
some
time,
but
I
will
be
submitting
it
shortly,
so
that
should
also
fix.
I
issue
where
we
are
submitting,
or
we
are
exposing
like
the
same
metric
twice
instant,
will
be
merging
them
and
will
be
just
exporting
a
single
record.
This
would
also
match
what
Python
and
go
is
doing.
I
haven't
like
really
finished
testing
it,
but
I
expect
to
send
that
sometime
this
week
so
follow
thanks,
I
introduced
already.
F
If
we're
not
able
to
continue
with
your
peer,
but
hopefully
the
new
one,
you
can
really
contribute
the
your
expertise
there
as
well
Thanks,
okay,
okay,
so
we
can
now
discuss
about
this
change
till
I.
Think
most
people
here
are
familiar
with
what
we
are
trying
to
achieve
so
I,
don't
know
how
do
I
start
like
Rayleigh?
K
F
K
K
K
K
Instrument
for
their
traces
logs
measures,
they
don't
have
to
take
dependency
on
itself,
will
be
compatible
with
time,
and
that
might
be
actually
a
good
thing,
because
that
a
lot
of
solid
components,
I,
guess
he'll-
donate
those
core
things
shipped
as
part
of
the
permits
platform
or
the
runtime
can
take
dependency
on
this
idea.
Instead
of
taking
external
dependency
on
country
and
also,
we
believe
this
is
the
thing,
because
a
lot
of
customers
they
might
get
confused
if
you
see
separate
scenes
that
we
have
volley
open
time
should
have
spent.
K
G
K
So
so
far,
what
what
we're
trying
to
focus
on
is
to
have
the
tracing
API
implemented
in
da
net,
so
trying
to
evolve
the
ability
to
embrace
open
climate
ray
tracing,
API
stack
and
our
goal
is
to
eventually
close
the
gap
and
be
as
close
as
possible.
But
we
realized
there
they're,
like
existing
history
internet
so
probably
like
week,
not
rename
activity.
K
Will
be
able
to
work
with
any
like
were
like
version
of
da
met
as
long
as
it's
officially
supported
by
Microsoft,
so
you
can
use
that
for
done.
That
course
repent,
a
row
or
four
point,
six
one
two
and
seems
like
both
teams
are
moving
in
their
own
way.
The
way
I
think
the
best
model
for
us
to
collaborate,
hist
ship
things
as
fast
as
possible
and
try
to
integrate
and
lower
from
that.
So
I
think
the
initial
thinking
here
is.
We
want
to
have
the
first.
K
K
So
we
can
compare
that
side-by-side
and
once
we
learn
that
feedback
we
take
that
back
to
down.
That
would
make
some
adjustment
to
try
to
call
the
guy
and
eventually
I,
think
probably
will
hundred
90
percent
or
95
percent
of
the
compliance
and
for
things
that
we
do
see
subtle
difference.
We
need
to
discuss
whether
the
difference
will
be
a
good
thing
for
us
or
meantime
a
walk
round.
This
is
my
general
thinking
relieve.
K
The
the
overall
goal,
as
I
mentioned
earlier,
the
goal
is
we
want
people
to
use
donut
as
their
instrumentation
layer
that
covers
the
three
pillars:
logs
traces
and
metrics,
but
for
the
the
current
effort,
I
think
worth
focusing
on
tracing
AGI
and
if
tracing
like
I
proved
to
be
successful,
will
have
energy
to
come
back
on
the
folks
on
the
other
80s
and
regarding
the
power
and
the
order
I
think
for
tracing.
This
is
definitely
something
we
committed
to
commit
to
ship
this
as
part
of
the
country
from
back
system.
K
D
K
Just
yeah
so
so
trace
tracing
ETA
will
be
the
first
and
what
we'll
use
this
model
to
see
how
we
can
collaborate
and,
yes,
the
right
balance
lee
eventually
we
can
align
with
open
Kalama
tree
in
the
in
the
like
instrumentation
layer
in
donut
and
after
we
take
a
learning
at
the
next
step,
we'll
be
focusing
on
matrix.
Given
we
already
have
a
good
understanding
of
MAGIX
and
the
spec
is
getting
like
finalized
for
locks.
The
situation
today
is
time.
K
She
were
still
in
the
discussion
about
what
was
the
future
for
knox
and
what's
the
scope,
so
that
part
will
come
later
and
also
from
Donna's
perspective.
Ilogger
API
so
far
is
like
it's.
You
know
pretty
good
shape,
comparing
to
like
tracing
or
matrix.
We
don't
have
a
very,
very
nice
matrix
H,
so
logs
will
come
after
facing
and
metrics
API
and
for
the
particular
like
activity
that
we
want
to
take
recently
about
the
PRS
discussions.
I
think
that
were
worth
scoping
to
tracing
API.
First.
B
I
think
the
direction
is
really
is
really
great.
I'm
I'm
I'm
have
the
same
question
about
kind
of
I.
Understand
them
the
thing
about
confusing
the
api's,
but
I
still
have
a
sensation
that
we
we
should
have
the
the
trace
API
exposed,
even
if
it's
likely
implemented
even
on
top
of
activity,
an
activity
source
and
I'll
say
why
I
think
that
perhaps
is
valuable
is
because
I
think.
B
First,
there
is
the
thing
about
people
that
work
in
environments
with
multiple
languages
and
multiple
platforms.
They
you'll
naturally
look
for
the
open,
telemetry
implementation
and
that
you
be
a
nice
convenience
to
have
that,
instead
of
depending
on
what
we
are
gonna
have
to
implement
at
some
level
some
kind
of
conventions
about
what's
transformed
from
activity
to
what
it
represents.
B
All
open,
telemetry
and
I
have
the
sensation
that,
if
we
just
have
activity
in
the
open
telemetry
for
dotnet
is
just
enabling
specific
exporters,
I
kind
of
think
that
people
are
going
to
rely
basically
on
implementing
straighter
listeners
for
activity
and
have
whatever
exporter
has
so
any
conventions
that
we
perhaps
have
open.
Telemetry
that
are
trying
to
unify
the
space
regarding
tracing
is
kind
of
gonna
be
lost
in
the
picture.
B
You
know,
I
understand,
for
one
side
for
people
right
frameworks,
SDK
is
big
ones,
I'll
read,
establish
and
more
specifically,
dotnet
core
doesn't
make
any
sense
for
them
to
take
dependence
on
open
telemetry,
but
perhaps
for
people
writing
applications
that,
in
these
polyglot
environments,
I'm
going
to
make
a
lot
of
sense
for
them
to
look
for
implementation
of
open,
telemetry
API
for
tracing
matrix.
You
know,
so
it
seems
you
mentioned
that
there
was
some
previous
mention
of
doing
this
kind
of
wrapper
and
there
was
a
decision
to
not
do
it.
B
B
I
I'm
think
about
the
stuff.
That's
in
the
data.
Let's
say
somebody
really
wants
to
use
status.
How
am
I
gonna
do
that
in
some
level
we
are
gonna,
have
some
convention
right.
So
what
I'm
saying
is
that
okay,
we
implement
the
API
as
open
telemetry
and
we
do
the
conversion
the
conventions
to
represent
status
that
are
done
in
activity
inside
our
implementation.
B
So
to
care
this
kind
of
information
we
we
are
gonna
need
some
convention
in
some
level.
You
know
if
people
want
to
represent
that
we
are
prepared
via
egg.
We
are
gonna
need
some
convention.
What
I'm
trying
to
say
is
kind
of,
since
you
are
going
to
need
these
conventions,
why
we
don't
implement
the
API,
the
tracing
PI,
showing
how
to
best
should
be
disconnections.
You
know
there
are
other
things,
I'm,
really
I,
don't
think.
B
None
of
this
is
a
deal-breaker
I,
don't
think
at
all
I'm,
just
trying
to
enumerate
kind
of
the
cases
that
come
into
my
mind,
kind
of
they
start
us.
We
are
going
to
need
some
convention
to
handle
that
the
attributes
there
are
other
things
like
we
support,
Long's
doubles.
This
can
be
arranged
and
I
think
we
can
live
in
even
if
there
are
just
strings,
but
preferably
we
can
arrange
that,
but
furnaces
there
is
is
more
different
semantics
that
it
breaks.
B
B
If
you
pass
a
new
value
for
attribute
in
open
telemetry,
you
were
supposed
to
remove
the
key
I.
Don't
think
activity
does
that
as
far
as
I
remember,
maybe
I'm
a
muse
remember
activity,
but
there
are
this
kind
of
difference
that
we
can
enforce
the
open,
telemetry
semantics
on
top
of
that
with
a
thin
layer
and
is
to
use
activity.
So
the
export
as
the
pipeline,
everything
down
that
just
use
activity,
but
we
do
have
the
proper
layer
API
that
offers
the
behavior
of
OPA
telemetry.
K
B
Was
thinking
in
really
trying
to
initially
I
the
first
four
that
occur
to
me
is
kind
of
to
have
the
open
element
or
API
in
its
own
namespace,
and
it
uses
under
the
hood
as
its
representation
of
span
and
everything
activity,
and
it
can
expose
the
activity
and
have
the
expose
access
to
it.
But
in
principle
it
has
a
API
that
looks
like
the
API
saw,
how
other
languages
type
I
think
the
reference,
because
of
the
nature
of
similarity,
will
be
Java.
You
know.
G
So
I
think
a
few
of
the
things
you
mentioned,
which
may
be
current
disparities,
though
I
would
aim
not
I'm,
not
sure
they're
going
to
be
disparities
for
too
long
I
would
like
to
resolve
the
status.
The
only
reason
that
we
don't
yet
have
the
status
is
because
there
were
some
concerns
about
the
way
the
specification
defines
status.
G
So
we
wanted
to
first
just
chat
at
the
specification
level
to
see
if
it
could
be
resolved,
but
kind
of,
regardless
of
which
way
it
went
whether
we
were
able
to
get
some
changes
or
if
people
said
no,
we
love
it
the
way
it
is
now.
Please
just
make
it
exactly
what
it
is
now
we
could
do
both
of
those
things,
but
we
didn't
want
to
put
the
API
in
prior
to
you
know.
G
Having
that
discussion
I
think
we
can
resolve
the
attribute,
like
the
the
typing
on
the
attributes,
I
mean
we
did
just
start
with
string
string,
but
it
doesn't
seem
like
a
fundamental
problem
to
have
other
types
there
I
think
we
may
actually
have
the
behavior
that
if
you
set
null
as
the
value
we
clear
it
but
I'll
need
to
go
check
so
I
mean
this.
Isn't
too
like
fundamentally
undercut
your
point,
but
it
is
to
say
some
of
the
examples
that
we're
using.
They
may
not
be
good
examples
in
the
long
run.
G
B
Really
don't
think
that
these
things
about
the
translations
are
blockers.
You
know,
I
really,
don't
think
I
I
just
thinking,
because
even
for
the
things
that
we
can't
have
a
direct
mapping,
we
are
gonna
have
some
convention
to
say
hey
in
this
context.
This
is
gonna
mean
this.
You
know
so
I,
don't
think
their
blockers
I'm.
Just
really
thinking
I
understand
the
question
about
I.
Don't
want
to
have
chewy
pies
that
people
want
to
come
from.
B
Choo-Choos
I
really
understand
that
so
I
think
in
a
sense,
it's
natural
that
people
will
gravitate
toward
achieved
source
and
activity
more
most
of
the
time,
but
I
do
if
you
feel
that,
especially
because
backends
do
special
treatment
for
some
of
these
information
that
some
people
may
want
kind
of
Oh
in
our
services
in
our
environment
that
is
using
open,
telemetry
I
won't
Express.
They
start
with.
Let's
say:
let's
say
they
start
in
having
the
API
makes
very
easy
for
them
to
do
that
in
dotnet,
because
they
can
just
go
in
kind
of
oh
I.
B
Did
this
in
Java
I?
Do
this
in
dotnet
and
I
want
to
get
the
same
effect
in
my
back-end.
You
know.
If
we
do
this
kind
of
thing,
relying
on
some
kind
of
convention,
then
they
have
to
kind
of
do
one
small
extra
work
that
is
searched
the
documentation
in
the
proper
way
in
what
I'm
a
bit
afraid
is
in
that
sense,
it's
true
that
we
orbit
a
limit.
We
provide
the
open
source
exporters
for
this
and
we
follow
the
conventions,
but
I
have
a
little
bit.
B
I
think
there
is
a
likelihood
that
perhaps
for
vendors,
they
not
have
a
lot
of
incentives
to
use
in
open
telemetry
it.
If
you
make
kind
of
hey,
let
put
my
listener
to
the
activities
and
I
already
have
my
export
and
use
that
you
know
in
innocences
is
like
what's
the
motivation
for
people
to
use
open
the
limiters
here,
if
everything
is
more
activity
source
and
putting
this
just
two
sides
together,
I
think
I'm,
I'm,
really
kind
of
think
that
we
should
preserve
the
API.
You
know.
G
I
guess
those
seem
like
two
different,
so
so
the
API
well,
maybe
I
should
ask
when
I
think
of
the
API
like
wrapping
the
activity
API
with
a
you
know,
a
tracers
band,
something
like
that.
I
think
that
is
fight
primarily
let's
say
customer-facing
like
the
end,
application,
developer
or
library
developer.
The
guy
who's
writing
his
instrumentation
that
they
are
the
primary
consumer
of
that
wrapper.
B
B
D
For
interruption,
you
just
interest
of
time,
I
think
what
like
there
may
be.
Two
requests,
like
one
request,
is:
please
take
API
and
implement
in
the
net,
but
this
may
not
be
realistic
because
we
really
want
to
one:
have
activity
continue:
API
continuation,
so
there's
like
this
is
what
drives
the
shape
of
a
pair.
Is
this
today,
so
we
cannot
have
like
both.
D
B
And
so
it
sounds
reasonable.
That's
what
was
one
of
the
things
that
I
was
gonna
say.
Is
that
perhaps
for
now
we
keep?
The
interface
is
about
the
export
and
we
can
export
both
for
the
time
being,
you
can
have
export
as
handling
both,
and
we
can
do
the
date
translation
or
in
a
branch
of
how
the
mechanic
is
on
how
we
do
that
doesn't
matter
that
much
but
I
I
think
it's
a
reasonable
approach.
I
I,
just
wanna
be
kind
of
to
be
sure
that
this
was
built
of
having
de
Piz
still
in
play.
B
You
know
I,
think
I,
think
the
API
I
really
see
more
usable
for
people
write
applications,
because
once
the
framework
has
our
offering
that
covers
that
I
think
people
you
jump
to
activity
I,
don't
see
a
reason
for
people
writing
libraries
and
frameworks
to
really
go
to
open
telemetry.
In
that
case
you
know,
but
perhaps
at
the
application
level
it's
a
bit
different.
You
know
so,
and
one
thing
as
I
mentioned
before
this
is
something
that
we
can
prevent,
but
I
I
think
in
this
case
for
applications
I
think
eventually,
people
are
gonna
write.
F
Thing
which
I
was
always
keeping
in
mind
was
it's
always
easy
to
add
things
later,
but
once
we
ship
something,
then
we'll
have
to
keep
shipping
it.
So
I
would
start
with
the
absolute
minimum
to
unblock
us
and
move
forward
and
like,
for
example,
status.
We
can
still
leave
with
that,
but
once
we
add
status
it's
difficult
to
remove
because
we
have
made
in
backward
compatibility
and
things
so
I
would
like
to
get
some
feedback
with
the
tag
saying.
Is
it
talking
or
is
it
not
working?
F
We
can
be
that
will
help
us
make
the
at
least
one
sort
of
one
level
of
integration
with
the
new
activity,
APA
and
open
telemetry,
and
then
we
can
give
feedback.
Okay.
This
is
really
not
going
to
work
like
I.
Can
ask
no
I,
don't
know
team.
You
have
to
fix
this,
but
to
do
that
the
first
step
is
to
do
the
level
1
or
step
1
integration.
K
This
integration
right,
so
my
suggestion
is
currently
we're
taking
this
approach
to
building
side
by
side
we're
not
removing
the
span.
Just
because,
like
we
know
it's,
not
a
hundred
percent
compliant.
This
is
already
like
well
known
so
on
Thomas
side
I
believe
currently
were
following
the
release.
Teickhoff
every
two
weeks
we
leave.
We
release
a
preview
version,
although
like
I'm,
not
sure
whether
this
is
Oryx
yeah,
so
with
that
I
think
he
often
times
have
we
want
to
catch
up
with
the
pace.
K
So
if
we
can
quickly
integrate
in
open
country
and
compare
less
the
best
side
and
give
feedback
to
down
at
him,
then
were
able
to
make
that
much
better,
and
this
is
like
in
general,
were
see.
We're
saying
this
is
a
good
direction
and
want
to
see
that
happening
on
things
that
are
not
blocking
I.
Think
probably
like
move
fast
and
learn
from
that
and
give
that
feedback,
and
given
that
we
all
have
this
communication
channel
I
think
those
are
the
problem
we
can
solve.
I
can.
B
K
So
quick
question,
so
so
what
are
the
real
blocking
issues?
When
you
look
at
CEOs,
PR
I
know
there
are
some
like
status,
not
existing.
So
these
are
the
things
we
know
and
we
think
this
is
not
a
blogger,
because
we
can
add
that
once
we
got
clarity-
and
this
is
something
we're
open,
I.
F
I
want
to
see
like
where
we
are.
No,
it's
probably
like
Mike
heard
some
questions.
Sergey
didn't
approve
or
reject
bralla
had
costumes,
but
I
chomping.
It
has
anything
to
do
with
the
fact
that
we
are
replacing
activity
of
span.
He
had
like
different
minor
concerns.
I
opened
separate
issues
to
tackle
tackle
them.
It
was
about
single
process,
so
those
should
not
be
blocking.
F
D
B
B
F
All
I
need
is
like
some
approval
and
like
Sergei
or
Mike
to
message,
because
I
cannot
approve
my
own
TR,
so
someone
else
has
to
approve
it.
So,
as
I
mentioned
in
step,
one
which
is
going
to
be
followed
by
a
bunch
of
things-
and
this
is
where,
like
all
over
to
helping
because
we
need
to
change
exporters
as
well,
but
the
one
good
thing
is:
we
already
removed
a
non-standard
or
non
open,
telemetry
stranded,
sorry
for
choice
of
words,
anything
which
are
vendors,
vendor
specific,
are
moved
out,
so
the
job
is
relatively
easy.
F
You
only
need
to
worry
about
like
a
smaller
site,
but
like
really
the
first
step
is
to
get
this
in
and
then
we
can
start
like
moving
other
things.
We
still
have
like
open
costumes
like
what
should
be
the
default
behavior
like
should
open
telemetry
by
default,
subscribe
to
all
activity
sources
or
selected
ones
thing.
We
can
like
wait
for
that
little
later,
but
right
now
the
current
model
is
obtained.
You
have
to
explicitly
say
which
activity
sources
your
thing
I
just
want
to
keep
it
as
simple
as
possible.
F
F
K
And
I
also
think
this
is
important
for
us
to
unblock
the
performance
test,
because
we
have
some
initial
discussion
about
like
the
performance
trade-offs
when
we
design
the
API,
and
we
have
some
untoned
scenario
once
this
got
merged.
I
can
I
can
hook
off
the
drill
exposure
and
do
the
performance
test
so
well,
we'll
come
back
and
and
see
like
whether
we
need
to
make
change
and
the
earlier.
We
can
decide
on
what
change
we
need
to
make
in
the
donut
API.
The
more
confidence
will
have
okay,
yep.
F
That
makes
sense
yeah,
so
this
appear
and
I
like
there
is
a
peer
there's,
an
issue
which
I
opened,
which
is
comparing
the
current
differences
again
so
just
to
make
it
into
a
Google.
Docs
kind
of
things
have
CC
to
comment,
and
edit
I
will
do
that.
We
I
mean
it's
like
a
comparison
of
what
is
currently
available
in
dotnet
versus
what
is
what
is
the
open
elementary
spec?
F
B
F
And
will
how
the
recording?
So,
if
he's
happy
once
the
recording
is
in,
then
we
can
go
ahead.
Otherwise
we
still
have
it.
They
call
with
him
to
clarify
things
that
has
Sayed.
Is
there
any
questions
on
this
topic
because
thing
I
am
kid
and
I
was
writing
a
like
a
like
documentation,
kind
of
thing
describing
the
steps
and
the
direction
I
have
already
explained
it,
but
you'll
be
having
a
post.
F
It's
it's
already
like
Ruby
would
be
everyone,
it's
just
that
we
didn't
to
post
it
as
an
issue,
so
there
will
be
some
more
communication
on
this
side.
It's
mostly
like
things
which
are
already
like
public
surveys.
Blog
already
covered
it,
but
you
just
want
to
like
formally
prostrate
as
an
issue
here.
So
people
can
see
like
what
are
we
reading
a
query?
Not
shipping
beta.
Yet
I
think
you
mean
an
announcement.
Oh
yes,
oh
it
would
be
like
an
issue
again.
We
can
mark
it
as
I.
F
K
K
Necessarily
I
a
moment:
donut
him
officially
announced
this
preview
version,
preview,
5,
I,
believe
and
saying
we
have
this
activity
API.
That
is
a
starting
point
and
eventually
I
want
to
align
with
open
country
API,
and
then
we
can
have
people
try
this
and
we
can
have
open
country
support
some
minimum
support
there.
People
can
come
and
try
and
give
us
feedback
that
has
to
have
a
much
better
understanding
for
or
open
collaborative
project
to
catch
that,
because
I
assume,
like
mandalas,
announced
this
trivial
5
release.
K
B
Question
just
in
theory:
we
have
plenty
of
time
since
its
November
the
time
but
activity
source
and
the
change
for
activity,
probably
you'll
not
be
on
those
high
bar
later.
That
requires
a
lot
of
things
to
change
right,
because
I
think
it's
assumed
it's
kind
of
new
and
we
have
a
little
bit
of
linear
as
if
we
need
later
changes
there.
G
G
I
mean
that's,
not
necessarily
a
hard
boundary
I
think,
but
it
starts
like
the
bar
goes
up
significantly
to
say:
oh
someone
has
already
deployed
this
into
production
they're
using
this
API
and
now
you
want
to
go
and
and
like
break
it.
So
you
start
needing
pretty
compelling
reasons
to
do
that,
whereas
if
we
do
it,
you
know,
let's
say
during
you
know
this
month
during
June,
probably
even
during
July
I.
Think
it's
much
easier.
We
can,
like
you
know
in
July.
You
can
just
say
hey.
We
got
this
preview
feedback
turns
out
like
this.
G
L
Yeah
I
think
I
agree.
No
I
I
would
funnel
this
in
as
part
of
preview
feedback,
rather
than
the
build
Conference
feedback,
because
hey
I,
think
everything
for
build
conference
is
almost
locked
in.
So
that
might
be
a
difficult
change
to
bring
in
right
now,
as
in
this
might
be
a
difficult
change
to
even
announce
in
the
conference.
So
I
would
rather
keep
this
as
something
we
do
repeatedly.
As
we
announce
previews
and
as
we
announce
dotnet
core
I
mean
dotnet
5,
yeah
yeah,.
G
So
my
understanding
is
all
the
work
that
we
did
is
part
of
preview
for
a
preview
for
has
not
officially
released
yet,
as
best
I
can
tell,
but
I
think
it's
supposed
to
release
you
know
within
the
week
or
something
like
it.
It
should
be
up
very
soon
on
nougat
and
if
you
want
it
even
earlier
than
that,
there
should
already
be
builds
on
my
get
the
only
go
from
there.
So
the
only
took
a
dependency.
K
Yeah,
the
the
the
thing
I
heard
is
on
May
the
Seventeen's
it
all
you
know
he,
the
public
knew
god
okay.
This
is
why
I
want
to
take
this
as
a
powerhead
to
ship
open
country.
So
we
can
get
feedback
as
early
as
possible,
because
the
early
we
can
cut
out.
We
have
more
time
to
refine
the
API
and
we
can
have
more
discussion.
K
F
K
B
K
For
some
of
the
simple
things
like
basic,
like
counters,
min
marks,
aggregation,
those
are
things
we
can
potentially
explore
for
since
I
have
more
complex
like
magic
spree
aggregation.
How
do
we
decorate
the
matrix
data
eating
like
like
the
contacts
we
have
distributed?
Contacts
this
probably
are
much
bigger
table
8
and
in
like
personally
I'm,
not
very
optimistic
that
we
can
achieve
everything.
I
think
it's
technically
like
unlikely
to
happen,
but
we
can
make
some
progress.
Give
some
idea
about
metrics
for
about
very
simple
metric
CGI
that
we
could.
K
I
F
Okay,
got
it
yeah
I
mean,
like
you
said.
It's
very
unlikely
given
as
follows:
Horseman
in
APA
itself
is
not
stable,
even
the
basic
point,
which
was
like
counters
that
that
could
change
like
last
week
so
yeah.
Yes,
let's
see
how
it
goes
yeah.
If
they
are
like
stable
enough,
then
maybe
we
can
answer
it,
but
for
now
I
I
mean
I'm.
Just
sharing
my
status
I
try
to
keep
focus
on
tracing,
so
we
can
have
something
ready.
Yeah.