►
From YouTube: 2022-02-16 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
A
Just
to
confirm
one
thing
as
before:
can
you
confirm
which
zoom
link
did
you
use?
Is
it
the
one
which
I
pasted
here
or
from
the
calendar?
I
open
it
from
the
calendar?
Okay,
I
have
the
correct
link.
Here.
Yes,
looks
like
there
will
be
another
change
to
the
way
meetings
are
held
so
right
now
we
have
a
dedicated
meeting
for
each
six.
A
We
used
to
have
like
a
shared
meeting
previously
now
we
switched
and
looks
like
we'll
be
going
back
all
the
way
to
the
original
way.
It
was
like,
where
we'll
be
having
the
same
link
shared
between
different
meetings.
So
I
don't
know
why,
but
it
looks
like
706
faced
the
same
issue
as
we
faced
like
they
tried
to
join
the
meeting.
It
says
the
meeting
is
like
scheduled
in
some
future
time
yeah,
but
when
that
happens,
we
can
go
back
to
the
old
way
yeah.
You
can
write
your
name.
A
Okay,
I
think
we
can
start
nothing
in
the
agenda.
So
if
anyone
has
questions
we
can
take
it
now.
Otherwise,
I'll
give
a
very
quick
summary
of
one
point
of
release
and
we'll
go
for
any
open,
prsor
issues.
A
Okay,
no
questions
all
right,
so
let
me
give
the
update
on
the
1.2
release,
so
it's
basically
same
update
just
last
week
now
the
spec
hasn't
moved
come
on.
Let
me
just
open
it
here,
so
there
are
three
issues
in
the
spec
repo.
A
One
of
them
is
the
pr
reaching
this
because
table,
but
it
sort
of
depended
on
these
two
issues
and
they
have
been
open
for
at
least
six,
but
they
are
contentious.
I
think
that's
the
right
way.
I
can
see
like
more
than
159
so
on
one
so
we'll
need
to
just
wait
for
it
and
most
likely
there
will
be
changes
in
the
sdk.
A
I
haven't
looked
at
it
very
recently,
but
based
on
the
initial
version,
we
would
be
having
some
issues,
so
we
would
be
needing
some
changes
in
and
keep
an
eye
and
make
the
changes
as
and
when
this
gets
moved
and
then
give
another
update
next
week.
A
A
Hey
sorry,
I
think
I
have
some
issues
with
my
desktop,
so
I
can
anyone
else
share
this
screen
I'll
and
maybe,
if
you
can
share,
that
would
be
great.
I
don't
think
my
machine
is
in
a
recordable
state.
A
Yeah
thing
all
we
need
is,
I
mean,
since
there
is
no
agenda,
maybe
we
can
quickly
go
over
the
open
pr's
from
the
last
week
on
maybe
issues
just
to
see
if
there
is
anything
which
we
can
quickly
agree
or
not
agree.
C
Sure
I
did
just
add
this
to
the
thing.
If
we
can
just
talk
about
this
quick
yeah
sure
this
was
an
issue
that
was
raised
last
week.
I
think
you
saw
it
see
joe
because
you
pointed
me
to
an
issue
that
a
previous
issue
that
had
been
opened-
and
I
was
just
curious-
what
wasn't
clear
to
me
so
for
everybody's
context.
C
Basically,
the
open,
telemetry
semantic
conveyor.
The
telemetry
specification
should
have
a
an
ability
to
create
a
root
span
when
there
already
exists
a
spam,
so
creating
a
root
span
like
you
know
something
like
this.
C
No
sorry,
where
is
this
person's
well
anyways
created
a
new
yeah
start
start
root
span
is
a
method
that
we
have
well.
It
looked
like
on
the
old
issue.
This
issue
984
it
looked
like
there,
was
going
to
be
some
discussion
with
the
runtime
team
or
the.net
team
to
maybe
address
this
with
yeah.
A
I
don't
think
we
actually
followed
up
on
that,
like
I
know
that
it
was
created
and
ludmila
left
the
this
project
and
like
no
one
else
followed
up
so
yeah
yeah.
I
mean
it
was
mostly
because
no
one
asked
for
it.
Otherwise
we
would
have
noticed
earlier.
C
Would
it
how
likely
do
you
think
it
would
be
like
if
we
were
to
open
up
an
issue
in
the
runtime
repo
that
somebody
would.
A
A
Yeah
I
mean,
if
I
mean,
since
this
is
a
like,
like
I
mean
not
routine
like
this-
is
a
genuine
ask.
I
think
it's
safe
for
us
to
create
that
issue
in
the
runtime
repo.
Now
we
still
have
like
sufficient
time
to
include
it
in
the
dot
net.
Seven,
if
the
runtime
team
agrees
to
do
it,
so
we
are
not
really
plot
by
time,
as
in
dotnet.
Summonship
has
not
sailed
yet,
so
we
have
plenty
of
time
at
least
till
end
of
april.
A
C
Yeah
yeah,
and
that's
that's
basically
what
I
I
told
this
this
person
before
before,
knowing
that
this
new
issue
existed
and
he
said
that
that
was
working
for
him.
So
yeah
there's
a
workaround.
A
So
I
think
one
thing
we
can
do
even
without
waiting
for
any
change
in
runtime
is
we
can
modify
the
the
shim
or
the
wrapper.
A
C
A
Yeah,
and
would
you
be
like
creating
an
issue
in
the
runtime
repo
and
one?
I
think
once
you
create
it,
you
tag
the
right
people,
but
I
can
bring
it
as
a
child
item
to
the
observability
candidates
in
the
dot
net
front
timer
as
well.
A
Yeah
I
mean
I
wanted
to
give
a
quick
update,
so
the
there
is
already
a
parent
item
in
the
triple
dealing
with
all
the
observability
items,
which
are
mostly
asked
from
open
elementary
communities.
So
it
looks
like
there
is
a
reasonable
amount
of
progress
there.
So
up
down
counter.
I
think
that
pr
got
merged
so
like
in
the
next
few
weeks.
We
will
start
seeing
the
daily
builds
containing
up
down
counter
and
we
can
take
a
wash.
A
We
can
probably
like
branch
off
and
then
start
building
the
sdk4,
because
we
currently
operate
with
the
assumption
that
all
counters
are
monotonic,
but
now
we
are
going
to
get
a
new
one,
so,
but
probably
too
early
like
we
can
wait
for
a
few
weeks.
A
Before
I
mean,
maybe
we
can
wait
for
one
point
to
to
be
out
before
we
start
worrying
about
the
next
thing
and
there
were
other
like
asks
as
well,
which
I
don't
recognize,
whether
the
prs
were
merged,
but
at
least
the
apa
reviews
were
done,
and
if
you
look
at
the
issue
there
will
be
youtube
recordings.
All
design
discussions
are
publicly
available,
so
there
would
be
youtube
recordings
in
case.
Anyone
is
interested
in
looking
at
them.
B
A
Okay,
yeah
can
quickly
go
through
open
issues
or
maybe
open
pull
requests.
Whichever
you
pick.
A
How
do
you
want
to
go
through
this
from
the
top
yeah?
I
think
we
can
just
quickly
go
over
the
top
because
we
usually
don't
spend
enough
time
on
pr's
in
this
meeting.
So
would
be
good
if
we
just
spend
few
minutes,
especially
since
we
don't
have
like
other
active
topics,
so
that
will
make
the
meetings
like
more
productive
yeah.
I
agree
so
I'm
from
a
mobile
device,
so
a
little
bit
difficult
for
me
to
see
but
I'll
try
to
comment
as
I
can
see.
Okay,.
C
Yeah,
so
we're
looking
at
the
I
just
opened
up
this
draft
pr,
because
attila
and
myself
were
talking
about
this
bug
that
somebody
raised
in
the
context
of
the
micro
service
example.
So
there
does
seem
to
be
a
goofiness
there.
I
haven't
pinpointed
what
it
is
yet,
but
it
seems
that
the
sdk
doesn't
get
initialized
when
you
might
expect
it
to
yeah
you're
suggesting.
Basically,
if
we,
if
we
just
touch
the
sdk
somewhere,
it'll
it'll,
invoke
that
that's
data
constructor
that
sets
the
default
propagator.
A
Yeah
but
like
in
this
case,
I
think,
like
that
micro
service
example
is
not
called
in
the
sdk
early
enough
right.
So.
A
Comes
we
had
a
very
similar
issue.
If
I
remember
correctly,
it
was
about
in
asp.net
the
very
first
request
when
it
comes.
That
first
request
is
the
one
which
initializes
the
sdk,
but
by
the
time
the
sdk
class
is
initialized.
A
So,
but
after
the
first
one
like
sdk
is
initialized
like
it
will
force
the
ids
to
be
in
the
w3c
format,
so
the
workaround
I
suggested
was
like
to
do
like
some
static
constructor
in
the
user
application
code
and
set
the
format
explicitly.
So
even
the
very
first
request
would
be
correctly
using
the
context
in
the
right
format.
So
the
the
command
which
I
made
in
this
pr
was
basically
same,
and
it
really
depends
on
the
application
type
like
what
is
early
enough
yeah.
You
know.
A
This
particular
example,
like
I
I
mean
think
you
know
most
about
this
example
right.
If
you
can
find
a
place
like
really
early
enough,
then
that
would
fix
this
issue.
I
think.
C
Yeah,
I
agree
I
was
hoping
to
find
a
way
to
do
it
without
touching
the
actual
application
code,
but
maybe
maybe
that's
not
going
to
be
possible.
It
looked
like
the
the
link
that
you
gave
was
the
logging
provider,
like
the
constructor
for
the
logging
provider
yeah,
if
there's
another
place
that
we
can
just
do.
A
Yeah
just
access
it
like
this
is
just
accessing
it
so
that
the
static
constructor
within
the
sdk
class
gets
triggered
because
that's
where
we
do
the
propagator
and
that's
where
we
do
the
activity
format
right
and
I
think
that's
where
we
set
up
the
self-diagnostic
logs
as
well.
So
like
someone
has
to
like
touch
it
before.
That
class
is
even
loaded
into
the
process.
A
Yeah
so
I
think
only
action
item.
Is
you
modify
the
example
to
like
initialize
earlier?
Don't
think
I
mean
if
you
find
something
else,
we
need
to
figure
out
what's
a
way
to
solve
it,
but
I
think
it
should
be
solved
by
just
modifying
the
example.
A
Yeah
so
I
think,
like
in
general,
like
I
would
hold
off
anything
related
to
log
record
right
now,
because
we
are
very
close
to
1.2
and
this
vrs
have
like
no
immediate
value,
because
we
will
anyway
won't
be
releasing
any
stable,
otp
exporter
anytime
soon.
So
there
is
a
plan
to
make
the
log
record
like
using.
A
I
mean
not
allocate
all
that,
because
right
now
it
just
creates
a
log
record
class
on
the
heap,
so
we're
trying
to
see
if
we
can
use
some
like
pooling
and
then
reuse
the
logo
code
instances
from
the
poor.
So
it's
basically
like
just
experimenting
with
that
idea,
but
since
it
is
too
close
to
1.2,
I
would
just
say:
don't
merge
anything
unless
it's
going
to
affect
the
like
actual
matrix
thing
so
yeah
we
can
come
back
to
this
afterwards.
C
A
An
object,
pooling
type
of
thing
yeah,
so
I
mean
I
think
that
there
was
a
to-do
item
when
we
originally
intended
created
the
logo
code
like
should
we
model
it
as
a
structure,
or
should
we
do
it
as
a
class
similar
to
activity
and
we
ended
up
using
class,
but
that
obviously
means
for
every
log
statement.
We
are
going
to
allocate
a
log
record
in
the
heap
yeah
so
trying
to
see
if
we
can
avoid
that
to
make
things
much
faster.
A
Yeah,
I
think
we
can
merge
test
efficient
approval,
so.
C
A
Think
it
it's
still
the
up,
you
need
to
probably
bring
the
master
to
it.
I
mean
main
branch
to
it.
Oh.
C
A
A
Yeah,
that's
a
lot
of
file,
but
it's
like
really
like
simplifying
things,
not
really
changing
any
actual
functionality,
so
it
should
be
safe
to
go.
C
A
A
A
It
looks
like
that's
the
existing
behavior.
I
don't
know
whether
we
should
spend
time
and
change
it
or
just
leave
it
like
that,
but
just
don't
document
it
as
public
as
because
for
other
two
providers,
the
meter
and
tracer.
A
We
explicitly
say
what
are
the
things
which
you
can
change
after
the
provider
is
being
built
for
matrix.
It's
nothing
like
once
matrix
provider
is
built.
You
cannot
change
anything
and
that's
by
design
and
for
tracing.
We
only
allow
processors
to
be
added
nothing
else,
but
it
looks
like
for
logging.
We
pretty
much
allow
everything.
A
It
has
some
risks,
so
I
mean,
if
it's
already
part
of
the
sdk
from
like
1.0,
then
we'll
just
leave
it
like
that.
But
maybe
we
can
just
do
like
some
tweaks
or
don't
know
whether
it's
considered
breaking.
If
we
stop
respecting
that
changes
after
the
provider
is
built,
maybe
we
can
do
that.
I
don't
think
it
would
be
considered
a
breaking
change.
A
I
mean
we
have
tools
to
catch
us
if
we
are
breaking
so.
Let's,
maybe
I
can
try
something
on
that,
because
that
would
be
at
least
consistent,
but
there
are
like
several
nuances
when
you
deal
with
this
thing
because,
if
like,
if
you
do
change
something
after
it's
built
like
it's
possible
that,
like
you've
been
like
some
in
between
state,
so
I
thought,
like
I
added
some
unit
tests,
but
maybe
like
I,
I
had
wrong
memory.
I
I
didn't.
A
C
He's
actually
advocating
for
not
allowing
updates
to
the
configuration
so
he's.
A
Yeah,
okay:
the
choice
was
explicitly
made
in
like
tracing,
because
even
the
spec
allowed
that
flexibility,
you
don't
need
to
allow
any
changes,
but
if
you
want
it's
up
to
individual
languages,
so
the
spec
really
is
very
flexible.
We
don't
really
need
to
do
anything
but
yeah.
This
is
something
which
we
can
come
back.
I
mean
it
should
not
affect
this
test
because
it's
already
existing,
but
we
can
come
back
and
see
if
we
can
figure
out
a
way
to
like
prevent
that
changes
from
being
affecting
the
existing
provider.
E
So
I
guess
for
the
test
I'll,
follow
the
pattern
as
this
and
then
probably
I'll
file
a
to
do
to
like
clean
up
the
existing
things
and
including
this
afterwards.
A
To
worry
about
an
existing
broken
thing
if
it
is
broken
or
so
yeah
continue
to
fix
this
test,
I
didn't
I
mean
I
will
review
it
definitely
but
fixing
the
original
issue
leave
it
as
a
separate
issue
and
just
link
from
this
pr.
So
we
know
how
to
keep
track
so.
E
C
Yeah,
mainly
just
the
log
exporter.
C
B
A
Yeah,
okay,
enthusiasm.
B
A
Because
I
think
sweet
like
indian
go
ahead.
E
Oh,
I
was
just
saying
I'll
push
a
new
iteration
for
this.
A
Yeah
yeah,
I
think,
since
these
are
like
very
targeted
tests,
we
should
really
do
the
thing
which
triggers
that
scenario.
So
if
you're
testing
formatted
message
then
do
something
which
will
trigger
a
formatted
message
to
be
created
and
click
always
do
it
in
two
apis.
So
one
is
the
current
one
which
you
are
using
the
logger
dot,
log
information
or
log
warning,
and
then
there
is
a
row
api,
the
log
dot
log.
A
I
think
there
would
be
existing
test,
which
we
added
like
some
time
back,
so
I
would
suggest
that
for
any
testing
related
to
logging,
I
would
add,
like
two
tests,
one
is
using
the
the
bare
I
logger
dot
log
api,
then
the
other
one,
which
is
an
actually
an
extension
method
on
my
blogger.
So
this
looks
like
just
doing
the
extension
method.
So
if
you
look
at
the
sdk
test,
the
other
file,
where
you
added
some
unit
test
for
the
public
setter,
I
think
it
has
examples
on
using
the
raw
api.
A
So
it
would
be
good
to
cover
both
because
that's
the
absolute
pair
api.
Everything
else
is
like
things
which
are
built
on
top
of
it.
A
I
think
utkarsh
added
like
some
tests
covering
that
like
in
those,
so
maybe
you
can
check
with
him
if,
if
you
need
help,
but
otherwise,
I
can
also
help.
A
A
Yeah,
okay,
yeah,
sometimes
vccl,
doesn't
pick
changes,
so
one
nice
way
is
to
like
really
go
and
start
a
new
pair.
Otherwise
there
is
a
like
github
command,
which
you
can
use,
but
I
don't
remember
what
that
was
something
like
eccla
slash
check
and
it
will
re-trigger
it.
A
A
Get
a
chance
to
read,
but
if
anyone
has
some
context
on
that,
that
would
be
great,
because
this
is
about
the
matrix
one.
Okay,
it's
only
specific
to
prometheus,
maybe
like
until
do
you
have
enough
context?
Can
you
share
what
ideas.
E
E
C
E
No,
no,
not
this
one,
I'm
going
to
try
to
find.
I
think
it's
somewhere
in
the
description.
E
Can
you
go
to
files,
yes
changed?
Please.
E
Right,
I
just
added
a
sanity
check
using
the
uri.try
create
to
check
whether
first
it's
an
absolute
api
for
the
http
listener,
prefixes,
and
whether
it's
using
http
or
https,
and
the
user
filed
a
bug
saying
that
they
are
using
the
wildcard,
for
they
were
yeah.
They
were
using
wildcard
previously,
and
this
check
breaks
the
scenario
and
I
looked
at
whether
they
are
like
apis
in
the
uri
class.
E
That
can
check
that
will
accept
the
wildcard.
But
there
is
no
current
api.
That
accepted.
Can
we
go
the
issue?
The
user
said
the
bug
yeah
user
said
the
documentation
said.
The
wild
card
is
valid
in
the
network
that
belonging
yeah,
yeah.
C
D
E
Right
and
yeah,
I
was
debating
myself
whether
I
should
drop
the
checks
and
my
check
was
noah
from
the
runtime
team
and
he
was
suggesting
that
if
we
have
to
like
align
the
tracks,
we
can
add
some
logic
on
the
add
prefixes
which
is
similar
to
this.
But
I'm
not
sure
whether
we
should
have
that
check
or
not.
And
riley
has
some
suggestions
in
the
pr
as
well.
A
Okay,
so
this
is
like
something
which
I
think
I
have
to
do
like
somehow
more
because
maybe
like
michael
did
some
research
there.
We
don't
know
whether
the.
A
Sorry
I
had
a
network
issue.
Sorry,
it
got
dropped
and
come
back
here.
Yeah
thing,
one
comment
which
maybe
michael
or
riley,
made
like
few
months
back.
That
is
the
non-middleware
option.
Is
it
like
ready
for
production?
A
That
is
not
clear
whether
the
http
listener
is
like
really
meant
to
be
like
a
production
level
server
or
is
it
just
meant
for
like
local
testing,
not
not
for
like
any
real
thing,
so
that
it's
related
to
this
question
right
now
and
do
we
expect
people
to
use
it
in
production
and
based
on
that
we
could
decide?
Okay,
if
it's
just
for
testing
yeah
star
or
that
should
be
fine.
A
Okay,
yeah,
I
think
we
need
to
discuss
a
little
bit
more
I'll,
also
read
the
pr
and
make
some
comments,
maybe
like
there
would
be
like
some
spec
about
it.
Just
don't
know
about
the
timing,
because
the
I
primitive
spec
prometheus
export
anyway
need
to
follow
whatever
is
laid
out
in
the
specification
when
it
lands.
So
we
have
like
plenty
of
time
to
fix
that
anyway,
it's
quite
likely
going
to
be
coming
after
the
sdk
stability.
C
I
hear
these
yeah,
let's
see
code
style,
I
haven't
looked.
A
At
some
of
these
myself
yeah,
I
think
those
are
like
relatively
easy.
We
should
be
able
to
just
merge
it.
I
mean
if
you
have
sufficient
approvals,
don't
think
there
would
be
any
discussions
needed.
C
A
So
there
were
like
couple
of
action
items
which
we
discussed
like
last
week
like
I'll.
Just
from
my
memory
like
we
had
to,
I
think
we
closed
on
the
options
for
now
the
otlp
exporter
and
ellen.
You
will
be
working
on
that
and
then
there
is
a
status
related
update,
which
I
will
send
an
update
later
so
having
like
just
sharing
that,
I
haven't
done
that
yet
so
I'll
be
sending
it
later.
This
week.
C
A
Oh
okay,
I
thought
sorry,
I
didn't
realize
that
it
was
ready
for
review.
So
maybe
I
I
yeah.
A
C
I'll
remove
this
stair
label,
so
it
doesn't
get
closed,
but
yeah.
This
should
be.
This
should
be
ready.
I
think
you'd
approved
it
back
in
the
day,
but
yeah.
A
A
Okay,
I
think
maybe
only
the
purse-
oh
yeah,
you
remove
it
right:
okay,
okay,
yeah
yeah
I'll
check
this
out,
because
we
I
mean
quite
likely.
We
want
to
release
the
otlp
exporter
along
with
the
sdk,
because
it's
kind
of
pointless
to
have
the
sdk
without
at
least
otp
exporter
for
metrics,
because
prometheus
is
anyway
going
to
be
late.
C
Yeah,
if
we,
if
we
land
this
one,
you
know
get
it
into
the,
how
we
want
it
I
can
there
are
so
I
think
it
I'm
forgetting
now
I
think
the
console
exporter,
maybe
the
in-memory
exporter,
has.
A
A
Yeah
I'll
definitely
make
this
a
top
issue
to
make
topiar
to
review
them,
because
this
is
important
for
one
point:
maybe
could
you
do
one
favor
by
adding
a
1.2
milestone
to
that
one?
So,
like
oh
yeah
yeah,
I
mean
I'm
expecting
that
like
sometime
in
the
next
two
weeks,
it
will
happen.
So
I
want
to
be
like
very
careful
with
what
we
merge.
So
by
adding
the
tag,
we
are
explicitly
announcing
that
okay,
this
is
intended
indeed
for
1.0.
A
So
the
spec
report
did
something
different.
They
used
a
milestone.
Sorry,
they
used
a
tag
which
says
like
required
for
ga
or
like
allowed
for
ga
something
like
that.
We
don't
have
such
tags,
so
we
can
probably
use
the
milestones
for
now.
Okay,
yeah,
okay
sounds
good,
yeah
yeah,
and
if
you
open
the
milestones
I
mean
I
still
don't
have
my
machine
yet
so,
if
you
go
to
the,
maybe
how
do
you
open
all
the
milestones
yeah?
So
like
the
description,
it
still
says
it's
not
done
because
of
the
spec.
A
That's
why
it
is
looking
like
it's
past
due
for
three
months
but
yeah
I
mean
if
there
is
any
better
way
of
like
so
many
people
are
asking
like
when
are
we
releasing
1.2?
So
I
generally
ask:
please
keep
an
eye
on
this
issue
or
this
milestone
and
we'll
update
it
as
we
get
more
clarity.
A
A
A
So
that
is
like
not
happened
on
original
plan
dates
on
noita.
So
if
anyone
asked
I'm
still
sending
them
this
issue,
sorry,
this
milestone
yeah.
A
Yeah,
so
I
did
like
some
more
cleanup
like
a
couple
of
weeks
back,
but
I'm
going
to
be
doing
that
again
this
week.
I
think
michael
has
that
histogram
bounce
pr
already
there
it's
just
that
we
haven't
had
a
chance
to
review
it,
so
that
should
get
closed
with
that
pr
and
maybe
for
the
measurement
value
I
did
made
appear,
but
then
later
decided
it's
a
bit
more
complicated
than
just
doing
evaluation.
A
A
Yeah
but
it's
a
little
bit
more
complicated
than
that:
it's
not
a
plane
check.
So
maybe
I
have
a
prs
associated
with
this
one
which
got
closed
anyway.
A
Yeah,
so
my
current
thinking
is,
we
should
leave
it
as
open
issue
and
come
back
to
it
once
we
have
the
initials
table
and
the
main
reason
is
I
mean
one
of
the
reason
is
we
are
going
to
allow
aggregations
to
be
specified
using
views
in
future,
so
that
would
change
how
we
validate
like.
Do
we
still
validate
that
and
throw
away
a
histogram
if
it's
negative,
even
if
the
aggregation
says
it's,
let's
say
uptown
counter,
so
those
combinations
needs
to
be
like
carefully
reviewed.
So
that's
a
bit
challenging.
A
It
simply
says
like
you
can
use
view
to
override
any
aggregation,
but
there
are
like
some
ramifications
of
that,
because
when
you
allow
a
histogram
to
be
aggregated
by
default,
the
negative
number
should
be
ignored.
But
when
you
take
a
histogram
and
the
view
says
aggregated
as
a
gui,
then
negative
values
make
sense.
So
I
don't
know
how
to
deal
with
that
part.
Yet
so
that's
one
part
and
second,
is
like
purely
like
numerical
things.
A
A
Yeah,
so
my
thinking
is
this
is
not
something
which
we
would
like
need
to
block
1.21
so
put
that
command
and
remove
that
1.2
milestone
from
this
issue,
because
this
is
like
a
fairly
intense
one.
We
need
to
spend
some
time,
discuss
it
and
probably
like
decide
to
do
it
after
one
point,
not
probably,
I
would
definitely
say
this
is
something
we
should
do
after
one
point,
because
if
you
want
to
support
the
change
in
aggregation,
that's
that
itself
is
very
complex.
A
So
so
it's
very
unlikely
that
we'll
have
time
to
include
such
a
fairly
big
change
into
the
1.21.
So
we
can
still
do
1.2
yeah
keep
the
scope
reduced
and
then
we
can
add
more
things.
A
Rest
of
the
things
are
like
just
general
things
catching
like
reviewing
and
making
sure
the
public
ap
is
like
good
enough,
so
I
am
still
doing
that
whenever
I
am
like
free
from
other
issues.
So
I
expect
that
to
continue
until
we
actually
do
1.4.
A
A
Yeah
see
a
new
member
sveta,
like
I
don't
regret
seeing
you
before.
So
maybe,
if
you
want
to
say
hello
to
all.
B
Yeah
yeah
I'm
joining
this
meeting
for
the
first
time,
hi
everybody.
This
is
shweta
from
cisco,
so
I
I
I
started
contributing
an
open
telemetry
in
the
dot
net
contrib
project
and
also-
and
I
had
one
one
of
my
prs
merged
and
one
of
my
pr
and
review
currently
so
I
just
joined
the
meeting
to
know
about
like
where
we
are
in
this
one
and.
B
A
Contains
instrumentation
libraries
which
are
like
far
from
being
stable
so.
E
A
That
our
attention
is
mostly
on
the
metrics
this
year
I
mean
not
this
year,
it
started
last
year,
so
so
yeah.
We
don't
really
actively
look
at
these
issues.
Unless
there
is
someone
brings
it
up,
but
is
there
any
pr
which
you
would
like.
B
Yeah
this
the
172
that
I
created
yesterday,
this
is
for
detectors.
A
A
A
Yes,
yeah,
okay,
this
is
very
interesting
because
does
it
I
mean
how
did
you
submit
this
br
before
we
shipped
the
I
resource
detector,
because
that
was
internal
as
of
today
morning.
B
A
A
So
can
you
now
open
the
pr
which
you
opened
in
the
contract
so
which?
How
did
you
that
do
the
resource
detection.
B
A
Yeah,
okay,
so
you
don't
need
to
do
this
one,
since
we
already
officially
merged
that
here.
So
all
you
need
to
do
is
just
wait
for
one
day.
I
think
it's
already
one
day,
so
there
will
be
a
daily
build
from.
A
E
A
You
just
use
this
package
to
use
that
daily,
build
okay,
and
then
you
should
be
able
to
like
do
this
in
the
official
way.
B
A
And,
like
probably
on
the
call
show
she
made
the
other
pr,
so
I
I
made
a
few
comments
just
to
write
an
example
for
how
do
you
write
a
custom
resource
detector,
but
since
you
already
figured
it
out,
you
probably
don't
need
that.
You
don't
need
to
wait,
but
I
would
still
prefer
to
have
a
example
document
which
says:
hey:
how
do
you
write
resource
detector,
and
this
is
how
you
do
it
so
that
should
come
anyway?
A
You,
you
could
also
help
with
the
review,
since
you
already
have
some
experience
in
that
area.
A
We
expect
to
have
like
more
like
people
contributing
the
detectors,
because
I
think
you're
doing
it
for
docker.
So
I
expect
like
there
will
be
more
for
aks
azure
aws
on
google
cloud
and
maybe
like
other
things
as
well.
B
A
In
contributing
more
or
you
just
did
like
docker
say
one
off
or
are
you
interested.
B
A
Think
could
you
create
an
issue
and
I
mean
if
you
create
an
issue,
you
can
select
self-assent
in
the
control
repo.
So
you
can
create
an
issue
in
the
control
panel
or
like
I
can
create
one
and
assign
it
to
you
as
well.
Yeah
yeah
works
for.
C
A
Yeah
so
when,
before
you
like,
actually
start
implementing
more
just
create
issues
for
each
and
every
resource
detector
you
intend
to
create,
so
maybe
you
can
get
some
help,
like
utilities
also
she's,
also
on
the
call
she's
also
trying
to
help
with
that
aspect.
So
just.
B
A
That's
the
answer.
Yeah.
Unless
the
spec
says
you
should
have
something
in
the
main
ripper
and
you
would
always
keep
it
in
the
country.
I
think
we
kept
one
in
the
main
repo,
the
environment
variable
and
telemetry
sdk.
So
those
are
in
the
main
repo,
because
stack
requires
us
to
do
that,
but
everything
else
is
going
to
be
in
the
control.
C
And
for
new
things,
I
know
we've
talked
about
this
in
the
past.
You
know
removing
the
the
give
trib
namespace
for
new
things.
Do
we
want
to
just
start
that
going
forward.
A
Yeah,
I
think
we
can
think
we
need
to
like
clean
it
up.
I
would
have
some
time
to
do
it
after
we
do
the
1.2.
So
that's
what
we
maybe
discussed
like
two
weeks
back
we'll
do
the
cleanup
after
1.2,
because
we
might
need
to
move
some
instrumentations
as
well
from
the
main
repo
to
the
country
brown.
A
It's
a
fairly
big
task.
It's
not
simple
task
of
like
lifting
some
code
and
putting
it
somewhere
else,
because
we
might
be
relying
on
some
of
that
instrumentations
to
do
some
integration
and
test
for
that.
Okay,
so
we
want
to
like.
Do
it
like
slightly
very
carefully
not
slightly
so
we
can
come
back
and
address
it,
but
for
now
like
we
can
keep
it
in
contrib
ship
it
as
non-stable
before
we
do
the
very
first
stable
that's
when
we
should
pay
very
close
attention
and
make
sure
it's
name.
A
Everything
is
correct
if
you
want
to
remove
the
contrib
name
that
should
be
done
before,
but
I
led
to
your
point
like
if
you're
starting
a
new
project
yeah.
Maybe
that's
like
right
time
to
use
the
name
without
the
contrib
word
in
the
project.
A
Yeah,
I
don't
know
whether
any
of
the
automation
is
relying
on
that
just
don't
reflect
yeah.
Maybe
you
can
leave
it
I'll.
When
I
review
the
pr
I
can
put
some
comment.
E
A
There
are
like
some
github
actions,
maybe
it's
relying
on
the
word
contra
being
there,
so
we
want
to
make
sure
like
it's
not
broken
when
we
just
change
it
for
one
project
as
opposed
to
changing
it
for
everything,
yeah,
okay,
yeah.
For
now,
let's
keep
the
name
as
before,
and
when
we
do
the
renaming
we'll
do
it
like
a
bulk
renaming
for
all
of
them
and
make
sure,
like
everything,
still
works,
all
right.
A
Yeah
and
like
when
you're
ready,
I
mean
once
you
update
this
pr
to
use
the
official
interface
just
mention
it
in
the
slack.
So
you
can
get
like
more
folks
to
look
at
it.
Yeah.
D
A
You
may
find
it
a
little
bit
difficult
to
get
reviewers
because
we
already
have
like
very
less
people
for
the
main
repo.
So
I
will
try
to
be
like
as
helpful
but
delays
in
general
because
we
are
just
trying
to.
B
The
thing
is,
I
feel,
like
everybody
is
really
quick,
because
I
opened
up
a
pr
in
this
repo
and
I
pinged
the
corresponding
person
and
I
got
it
merged.
A
B
B
A
But
I
was
generally
mentioning
that
the
maintenance
of
the
main
repo
and
myself
are
two
of
two
of
the
maintainers
and
there
are
like
few
approvers
as
well,
but
the
general
goal
is
to
ship
1.2
stable,
which
has
the
metric.
So
that's
been
the
priority
focus
since
last
march
or
april,
so
most
likely
we'll
be
at
least
spending
more
time
there
in
the
next
at
least
two
weeks,
and
once
we
do
that
and
we'll
come
back
and
have
more.
A
But
if
you
really
need
help
with
anything,
you
can
mention
in
the
slack
channel
or
bring
it
to
your
signature
next
week.
Also,
if
you
don't
get
to
it
sooner
than
that.
A
Because
there
are
other
peers
in
this
repo
which
are
also
like
similar
state,
it's
been
open,
but
not
enough
people
to
review,
so
it's
progressing
slightly
slower
than
it
takes.
It
is
supposed
to
be
so
like
the
source
it
is
should
be
relatively
fast
because
there
is
not
much
controversial
topics
to
it,
so
it
should
be
like
relatively
easy,
so
try
to
make
it
happen
really
fast.
A
Anything
else
or
anyone
else
with
any
questions
so
sweda.
If
you
want
to
write
your
name
to
the
attendees
list,
that
would
be
great,
so
just
keep
track
of
who
is
joining?
I
mean,
if
you
really
want
no.
A
Okay,
thank
you
yeah,
so
you're
from
like
cisco,
like
I
saw
some
news,
that's
fun
is
being
acquired
or
something
is
that
you
are
an
employee
of
splunk
or
you
are.
A
A
I
was
a
bit
curious
to
see
like
people
from
known
like
apm
vendors.
So
that's
why
you
also
you.
A
Okay,
yeah:
okay,
yeah!
I
guess
that's
it!