►
From YouTube: 2021-04-14 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).
C
C
B
Okay,
it's
two
minutes
past
four,
so
I
think
we
can
go
ahead
and
start
if
you're
in
the
call-
please
add
your
name
to
it
in
this
list.
So
once
again,
agenda
is
very
small.
So,
oh
mike
added
something
about
trace
id
ratio
based
sampler
issue
yeah.
I
think
we
can
go
over
that
first,
the
pr
one
is
like
really
small.
So
michael,
can
you
tell
me
what
this
issue
is
about.
D
Somebody
just
asked
me
about
this.
Like
30
minutes
ago,
in
my
company
they
have
a
product
exposed
two
ways.
One
of
it
is
a.net
framework
application.
D
They
have
the
parent-based
sampler,
with
the
ratio
sampler
as
the
inner
sampler
on
the
dotnet
framework
side.
It
works
fine,
they're,
seeing
more
or
less
the
sampling.
They
want
10
percent
net
core
they're
using
the
exact
same
startup
code,
same
sampler
mechanism
and
everything
they're
seeing
zero
percent
sampling.
So
I
just
wanted
to
bring
up.
D
Is
anyone
using
it
in.net
core?
I
did
a
little
bit
of
research
right
before
this.
It
looks
like
our
algorithm
is
taking
bytes
from
the
trace
id
in
the
production
version
of
diagnostic
source.
Trace
id
is
a
guide
and
I
think
goods
are
generated
differently
on
windows
and
linux.
So
I'm
wondering
if
maybe
it's
just
different
algorithms
aren't
working
with
our
algorithm,
but
I
have
no
idea.
I
just
started
poking
around.
B
Okay,
no,
I
have
never
heard
of
this
issue
before
so
it
would
be
a
surprise
if
there
is
an
actual
issue
here.
D
B
Yeah
most
of
the
six
have
followed,
I
think
whatever
was
the
algorithm
in
the
open
sensors
slash
open
tracing,
so
there
is
some
sort
of
consistency
across
implementation,
even
though
it's
not
well
defined,
but
just
coming
back
to
the
problem,
like
did
you
say
it's
only
in
dot
net
core
on
linux
or
net
core
in
general,
no
matter
whether
it
is
windows.
I.
D
Don't
know,
that's
just
the
information
I
have
so
far.
I
said:
how
are
you
deploying
it
and
the
answer
was
it's
a
linux
container,
so
I
don't
know
if
anyone
has
it
working
dot
net
core
on
windows
or
if
it's
linux
only
it
could
be,
it
could
be
a
bug
in
just
that
application.
I
don't
know.
I
was
literally
just
looking
at
it
thought
I'd
bring
it
up.
B
Yeah,
so
nothing
obvious
stands
up
in
my
mind,
but
we
we
can
try
to
see
if
we
can
deploy
something
into
only
our
ca
is
all
run,
but
I
don't
think
we
have
a
targeted
test
which
checks
the
trace,
id
generator
or
trace
id
ratio
based
sampler
being
like
working
or
not.
I
don't
think
we
have
any
targeted
tests.
So
that's
what
I
would
start
with
just
to
see.
If
there
is
any
issue
with
like
for
probabilities
being
off.
B
B
So
if
there
is
a
unit
test
or
if
we
just
add
a
unit
test,
we
probably
don't
need
to
do
anything
else
because
we
already
covered,
like
I
think,
would
do
latest
yeah
and
in
all
three
dot
net
core
versions
as
well.
So
that
would
be
a
good
place
to
start
yeah,
but
it
could
be
interesting.
E
E
B
Okay,
thanks
yeah,
I
want
to
like
do
a
quick
review
of
some
open
pr's,
because
we
do
have
like
quite
significant
the
every
week.
I
try
to
close
it
by
like
five
or
six
we
get
more
so,
which
is
a
good
thing
it
it's
just
that
maybe
not
able
to
keep
up
with
old
one.
I
mean
fairly
big
list
now
so
just
to
go
over
like
some
of
them.
A
lot
of
prs
are
about
metrics
at
least
four
of
them.
B
So
I
think
we
can
ignore
it
for
now,
because
in
next
couple
of
weeks,
we'll
actually
start
merging
prs
for
metrics,
so
most
of
them
would
become
like
not
relevant
anymore,
like
matrix
timestamp
issues,
adding
aggregation
capabilities,
all
those
pr's
would
be
like
like
closable
next
week,
however,
there
are
like
still
some
other
like
huge
list
of
open
peers.
B
We
still
have
so
I'm
just
wondering
like
if
there
is
any
any
any
better
way
of
managing
it
like
would
it
help
if
we
just
ignore
everything
which
is
marked
as
draft
or
like
work
in
progress
and
then
like
try
to
see
if
everyone
can
take
like
one
or
two
prc's
for
the
week
and
see
if
we
can
reach
closure
on
them
by
next
week,
some
of
them
are
really
old.
I
mean
most
likely
like
some
of
them
may
be
waiting
on
me.
B
A
B
Yeah
no
like
when
we
originally
started,
I
think
the
contributing
dog
did
mention
either
use
like
wip
or
draft,
but
then
I
corrected
it
to
just
say
craft,
but
some
people
use
it
use
both,
maybe
like
at
one
point
like
the
draft
feature,
was
not
enabled
in
the
repo
early
stage.
B
I
think
I
mentioned
somewhere
here
that
yeah
I
forgot
like
where
it
is
yeah
yeah,
but
both
are
conceptually
same
like
it's.
A
B
B
So
any
anyone
has
like
some
bandwidth
to
take
up
one
or
two
pr's.
Please
speak
up
now,
so
we
can
just
assign
like
someone
if
it
is
like
waiting,
I
mean
if
it's
already
reviewed
and
waiting
for
changes
from
the
original
person
who
submitted.
We
can
like
just
mention
that
hey.
Are
you
still
working
on
it?
If
not,
we
can
just
close
so
so
that,
like
overall,
we'll
have
a
like
much
more
healthier
repo,
rather
than
having
like
a
lot
of
abandoned
peers.
F
Hey
cj
yeah,
I
couldn't,
I
couldn't
help
out.
I've
been
not
doing
a
great
job,
staying
on
top
of
kind
of.
What's
what's
current
and
what's
kind
of
ready
to
move
along,
but
I'd
be
happy
to
jump
in.
I
definitely
have
a
couple
of
working
progress
things
that
I
I
just
need
to
find
the
time
to
move
forward,
and
I
I
think
I
have
a
marked
as
draft
so
I'm
not
worried
about
those
okay
yeah.
So.
B
We
can
ignore,
like
anything
which
is
smart
or
draft.
So
if
you,
if
you
can
like
just
take
like
one
or
two
pr's,
you
can
just
mention
in
the
pr
itself,
hey
I'm
going
to
take
a
look,
so
at
least
it's
better
than
like
being
an
abandoned
pr
and
like
in
some
cases.
It's
it's
clearly
like
like
experiment
like
this.
Is
it's
not
marked
as
thought,
but
this
is
experiment.
We
you
don't
want
to
proceed
with
this,
but
this
is
just
open
just
to
get
some
feedback.
B
Those
we
can
ignore,
but
anything
which
does
not
fit
into
this
criteria,
basically
like
anything
which
is
draft,
ignore
anything
which
says
do
not
merge.
You
can
ignore
and
anything
which
is
about
matrix,
ignore
for
at
least
one
more
week
and
just
take
like
one
is.
I
think
that
if
all
of
the
people
here
I
take
like
one
pres,
then
we
can
get
into
like
below
10
active
peers.
That
would
be
like
much
more
healthier.
B
G
G
B
Did
like
a
try
in
the
spec
repo
there
is
triaging
like
like
one
is
in
exciting
week
to
try
edge
and
the
same
milestones
or
labels
to
the
issue,
but
we
haven't
done
that
here
and
it's
mostly
because
we
we
didn't
have
this
this
much
open
players.
Previously,
it's
only
like
in
the
last
two
months
that
we
started,
having
like
too
many
open
pr's
part
of
that
reason
is
I
I
was
personally
not
heavily
active,
so
like
a
lot
of
them
were
like
left
out.
I
think
it's.
B
The
same
with
lake
allen
was
mentioned
that
we
were
not
like
as
active
as
like
three
months
ago,
so
that's
partially
the
reason
and
then
like
in
general,
like
people
are
submitting
more
pr's
as
well.
So
that's
also
the
reason
so
like
what
like
in
terms
of
priority,
like
I
don't
think
we
have
ever
like
set
aside
any
rules
about
that.
B
But
if
you
have
like
any
suggestions
or
I'd,
be
like
asking
every
meeting
like
if
there
is
any
pr
which
the
other
things
is
important
and
requires
review,
please
mention
it.
So
we
can
prioritize
because
you
have
like
15
20,
open
peers.
You
just
want
to
you
just
have
time
for
one
or
two:
you
want
to
pick
something.
So
if
you
explicitly
mention
one
of
us
in
the
maintain
or
show
up
in
the
sig
meeting,
we
try
to
prioritize
that.
But
it's
just
like
an
informal
rule.
B
Or
so
for
metrics,
I
think
I
I
will
do
the
job
of
like
marking
them
as
metrics,
because
we
we're
not
focusing
on
metrics
at
all
like
for
almost
six
plus
months
now,
but
now
like
in
next
two
weeks.
We
are
going
to
start
working
on
metrics,
so
I'll
do
the
job
of
like
marking,
which
pr's
are
related
to
matrix,
and
I
will
close
them
if
they
are
no
longer
relevant.
So
we'll
clean
all
the
metrics
part
by
next
sig
meeting.
B
But
my
comment
was
more
about
like
we
still
have
like
plenty
of
other
pr's,
a
really
huge
number,
even
if
we
exclude
matrix
and
draft.
B
Or
of
them
alone
I
see
yeah,
some
of
them
are
tied
to
issues.
Some
of
them
are
like
just
peers
without
any
associated
issues,
so
I'll
just
take
like
one.
So
this
is
a
pr
which
is
adding
an
improvement
to
the
reduced
instrumentation,
but
there
is
no
issue,
so
this
was
not
discussed
before,
but
still
it's
appear
and
I
want
to
like
either
tell
the
person
that
hey
this
is
not
accepted,
or
we
had
to
tell
that.
This
is
good
feature,
but
we
just
don't
have
the
time.
B
So
we
need
to
set
the
right
expectation,
so
people
don't
lose
heart
when
they
try
to
submit
pr
basically
trying
to
make
sure
like
everyone
feels
not
abandoned,
because
that's
not
a
good
sign
for
a
report
to
be
if
everyone
feels
sad,
their
peers
are
abandoned.
B
So
we
have
like
certain
peers
which
are
like
not
looked
at
at
all,
so
I'm
just
asking
for
help.
If
anyone
has
bandwidth
to
take
one
or
two
pr,
we
can
try
to
close
more
in
the
coming
weeks
because
the
there
is
a
like
metrics
for
coming
in
two
weeks
and
at
that
time
pretty
much.
All
of
I
mean
I
can
tell
you
like
myself
will
be
like
heavily
occupied
with
metrics
or
like
issues
which
are
not
related
to
matrix,
would
naturally
be
left
out.
B
So
you
want
to
close
them
before
we
get
into
the
matrix
thing.
B
No
anyone
gets
like,
for
example
like,
even
if
you
are
not
a
maintainer,
you
can
still
mark
your
comments.
It
won't
be
counted
as
an
approver
account
because
you're
not
listed
as
a
official
approver
or
official
code
owner.
It
won't
be
counting
or
it
won't
be,
giving
you
the
green
signal
like
because
yeah
I
mean.
If
I
approve
it,
there
is
a
green
tick
mark,
but
it
doesn't
prevent
anyone
else
from
doing
it
because,
like
people
are
like,
sometimes
personally,
I'm
not
an
expert
in
like
readies.
B
So
if
someone
is
an
expert
in
readings
or
grpc,
you
you
can
just
either
approve
it
or
reject
it.
So
that
gives
like
one
more
data
point
for
me
to
consider.
Okay,
it's
already
revealed
so.
B
Yeah
I
mean
I'm
not
like
I'm
not
saying
that
everyone
has
the
necessary
knowledge
to
do
it,
but
people
who
know
they
can
help
out-
and
I
think,
like
you
probably
know
like
most
about
metrics
now,
so
you
can
help
with
matrix
vrs
as
they
come.
G
B
Do
like
much
damage
like,
even
if
you
say
a
pro
like
it's
usually
like
someone
else,
will
also
review
it,
because
if
you're
not
listed
as
a
approver,
then
it
won't
have
this
tick
mark.
So
some
who
is
maintaining
like
as
of
today,
like
I
think,
oh
like
sergey,
and
yes
against
here.
So
all
the
maintainers
are
today
in
this
call.
So
it's
very
rare
that
all
of
us
will
be
in
accord
like
it's
probably
like
two
or
three
months
since
we
all
are
in
the
same
court
yeah.
So.
B
Yeah,
that's
that's
something
which
we
that's
why
I
was
trying
to
get
to
like.
If,
like
you
are
a
maintainer
or
you
are
an
approver
and
you
know
about
radius
or
you
know
about
jager,
you
can
say
that
okay,
I
I
look
at
this
pr
and
then
I
can
assign
it
to
you
so
like
next
week.
I
can
ask
like:
is
it
good
or
is
it
like
totally
wrong
or
you
can
say
that
okay,
I
don't
have
a
confidence
to
review
it.
Either
way
is
fine.
E
So,
what's
working
in
specifications
very
well,
is
we
currently
auto
assign
opr's
to
tc
members,
and
there
is
a
board,
that's
installed
and
just
circles
through
all
the
maintainers
and
assign
it
to
individuals.
It
works
very
well.
I
mean
you
can
always
reassign
to
other
person,
but
it's
like
it
limits
the
number
of
pr's
to
look
at.
B
I
see
so
like,
but
it
still
means
like
it
is
assigned
to
like
someone
from
the
like
predefined
list,
like
the
approvers
or
maintenance
like.
E
Yeah
and
typically,
it
means
that
this
person
is
not
responsible
for
actually
emerging,
like
the
person
responsible
for
it,
but
does
not
necessarily
mean
that
this
person
have
to
go
through
all
the
lengths
of
commenting
it
like
your
main
job
is
to
make
sure
that
comments
are
addressed,
and
if
there
is
no,
nobody
reviewed
it.
E
You
just
say
in
person
like,
please
find
somebody
to
review
your
pr
or,
like
you,
you
do
your
review
yourself,
because,
as
victor
you
pointed
out
like
you
know,
you
may
not
know
some
specifics
of
this
code,
and
especially
with
some
technology
like
radius
isolates,
you
may
not
know,
what's
working
for
it
as
well,
so
you
may
want
to
ask
person
to
bring
somebody
else
to
review
this,
and
after
like
couple
several
reviews,
you
can
add
your
own
opinion
and
then
merge.
It.
B
Yeah
that
sounds
doable.
It's
just
that
I
did
see
the
whale
expect
works.
I
didn't.
I
never
thought
that,
like
this
ripper
would
have
like
enough
like
requirement
to
automate
it
like.
It
was
always
like,
like
much
smaller
in
scope
than
inspect,
so
I
thought.
Okay,
we
can
just
manage
it
in
the
sig
meetings.
We
can
just
like
ask
people
who
are
in
the
car
like,
like
usually.
E
B
At
least
half
of
the
prover
slash
maintenance
in
the
course,
so
we
can
just
ask
them
hey.
Do
you
want
to
take
one
of
this
then?
Do
it
like
at
least
manually
to
begin
with
the
end,
and
then
we
can
borrow
the
both
idea
from
specs,
where
we
can
assign
today.
G
I
think
see
joe.
I
think
that
if
I'm
just
speaking
for
myself,
if
for
some
reason
I
go
to
the
the
sig
meeting
and
I
see
a
whole
bunch
of
stuff
assigned
to
me-
it
would
be
kind
of
like
on
me
to
kind
of
like
oh
wait,
should
it
be
assigned
to
me
so
that
at
minimum
put
focus
on
individuals
to,
I
don't
mean
to
add
more
work
to
people,
but
that
at
least
have
some
level
of
social
pressure
to
kind
of
look
at
it.
Yeah
I'm
not
trying
to
assign
people
work.
B
I
mean
you
have
a
point
it's
just
like
like
in
this
repo.
Historically,
we
didn't
have
like
that
kind
of
like
flood
of
vrs,
which
are
open
or
left
unattended.
It's
just
a
recent
thing
previously
or
we
even
when
we
were
able
to
do
release.
We
had
like
10
or
15
of
number
of
pr's
per
day
for
at
least
a
week,
but
it
was
like
closed
like
very
quickly.
There
is
like
very
tight
feedback.
You
immediately
fix
something.
B
G
B
Like
open
pr
from
bottom
up
and
see
if
anyone
can
volunteer,
so
these
two
are
metrics
so
I'll.
Just
sorry,
this
one
is
metric,
so
I'll
just
ignore
it.
There
is
this
pr
which
was
added
long
back.
It's
just
an
internal
implementation
detail
of
how
our
circular
buffer
works.
B
So
if
anyone
is
volunteering
to
review
it,
I
did
review
it
like
long
back.
I'm
pretty
sure
I
did
review
it.
Maybe
I
even
left
comments
yeah.
So
if
anyone
here
has
the
bandwidth
to
do
this
or
some
interest
in
doing
this,
I
can
just
assign
it
to
you.
I
mean
if
you
come
back
next
week,
saying
that
sorry,
I
don't
think
I
can
do
this.
That's
also
fine,
like
we
can
find
someone
else.
B
Is
this
a
good
use
of
everyone's
time
just
to
assign,
like
one
pr's
like
right
now
in
this
call
and
come
back
and
see
if
you
have
any
improvement
next
week,
so
the
person
ascend
need
not
necessarily
be
the
expert
they
can
leverage
others
or
like
if
it
is
waiting
for
the
person
who
submitted
the
pr
to
some
like.
If
the
review
comments
were
not
addressed,
we
can
simply
mention
this
person
who
submitted
hey.
There
are
some
open
questions.
G
B
Object
like
I
know
that
a
lot
of
other
peoples
are
here.
So
I
just
want
to
hear
like
is:
is
that
acceptable
for
all
of
us
for
all
of
others?
If
I
just
randomly
assigned,
because
that's
what
a
boat
would
have
done
as
well,
so
I'll
just
play
the
boat
myself.
B
B
Yeah
I'll
just
do
manual
assignment.
If
that's
the
case
like,
I
don't
need
to
ask
anyone
I'll,
just
like
randomly
assign,
but
first
I'll
do
the
job
of
like
eliminating
all
the
matrix
and
draft
and
then
start
assigning
and
if
there
are
any
peers
which,
like
the
same
thing,
which
I
asked
as
before,
like
if
you
are
one
of
the
other
of
the
pr,
and
it
requires
like
attention
just
ping
or
mention
again.
B
Any
other
comments
on
this.
It's
somewhat
similar.
We
need
for
issues
also,
I
mean
I
did
a
couple
of
round
of
cleanups
two
months
back
and
it's
it's
long
since
we
did
any
cleanup.
So
this
is
also
another
thing
which
we
could
use
some
help
if
we
can
like
load
balance
between
people,
because,
right
now
it's
150
issue
and
pretty
sure
like
at
least
50
of
them
can
be
straight
away
closed
either
because
they
are
no
longer
relevant
or
they
are
addressed.
B
I
mean
we
have
issues
open
from
like
more
than
two
years.
A
lot
of
them
are
metrics.
I
think
I
I'll
clean
up
the
matrix
one,
because
those
issues
don't
make
sense
anymore
in
the
new
matrix
spec,
so
we
can
just
remove
them,
but
if
anyone
else
like
when
you
are
like
free,
just
look
at
like
issues
really
old
ones,
also
is
fine
and
just
close
them
if
it
doesn't
make
sense.
B
Yeah,
no
strong
preference,
let's
see
how
it
goes
for
like
next
couple
of
weeks
and
if,
if
there
is
no
progress
and
I'll
try
to
use
the
like
this
time,
which
is
basically
our
submitting
to
actively
look
at
pr's,
slash
issues
and
come
to
some
resolution.
B
So
for
now
I'll
just
ask
everyone
based
on
your
free
cycles,
state
like
issues
and
respond
to
them,
either
close
it
or
like
mark
mark.
Like
a
I
mean,
if
it's
marked
as
a
bug,
but
it
is
not
a
bug,
then
change
the
label
to
simply
indicate
what
it
is
like.
Is
it
a
feature
request
then
marketers
enhancement?
B
So
that
way
we
will
have
like
a
much
healthier
repo
rather
than
having
like
hundreds
of
issues,
because
once
we
start
like
shipping
all
the
instrumentations
and
metrics,
we
start
like
seeing
more
and
more
issues,
so
we
just
want
to
make
sure
it's
like
in
a
controlled
state.
B
There
are
like
certain
questions
as
well,
which
I
mean
I
did
see
briefly,
but
I
never
had
the
time
to
look
at
it.
So
I
I'll
try
to
do
like
my
job
like
I'll.
Take
like
five
issues
today
and
try
to
respond
so
and
like
basically
asking
others
also,
if
you
can't
respond
to
some
issues,
yeah
and
we'll
see
like
in
next
week.
If
there
is
some
improvement,
we
just
stick
with
that.
Otherwise
I'll
try
to
use
the
like
submitting
to
go
over
some
issues.
A
B
B
It
takes
some
effort
like
most
of
the
time,
it's
very
easy,
but
like
sometimes
yeah,
it
takes
some
time
to
like
really
review
it
and
try
to
see
if
there
is
a
repro
and
yeah.
B
G
G
G
B
I
think,
as
the
person
who
posted
this
issue,
you
have
the
permission
to
change
it.
So
if
you
know
that
this
is
not
a
bug,
it's
a
rather
a
question
or
like
like
a
feature
request
or
or
just
a
documentation.
If
you,
then
you
can
tag
it
accordingly,
but
I
don't
think
you
have
the
like
permission
in
github
to
change
tags
on
issues
which
you
did
not
create.
B
B
E
B
Yeah
we
can
make
you
pleasure
and
I'm
pretty
sure,
like
you're,
going
to
be
like
working
on
metrics.
I
suppose
so
you
you
will
be
anyway
busy
in
that
aspect,
but
yeah
I
mean
I
forgot
like
we
did
have
a
treasure
roll,
but
then
we
removed
it.
E
I
don't
know
where
it
got,
but,
yes,
we
did
have
it.
B
Okay,
yeah
thanks
for
I'm
if
you're
looking
for
like
the
immediate
help,
would
be
like
looking
at
any
peers,
which
you
think
makes
sense
for
you,
like
it's
easy
for
you
to
review,
then
just
do
that
as
well,
that
that
would
also
be.
B
Helpful,
okay,
any
other
comments
on
that.
I
I
actually
have
a
couple
of
questions
about
a
individual
couple
of
individual
peers,
so
I'll
just
use
this
chance
to
ask
others
opinion
and
merge
it.
So
this
is
one
of
the
prs
it's
about
sql.
So
michael,
I
did
tag
you
a
little
early
to
see.
B
If
you
have
any
objections,
it
looks
like
you
had
a
comment
about
so
basically,
this
pr
is
eliminating
a
couple
of
options
and
consolidating
into
like
set
db
statement
for
text
and
set
db
statement
for
stored
procedure
and
make
it
available
in,
like
all
frameworks,
dot,
net
and
dot
net
framework,
and
there
will
be
like
no
ops
where
it
doesn't
make
sense.
For
example,
in
dotnet
framework,
there
is
no
way
you
can
record
the
actual
exception,
so
it
will
be
like
no.
B
So
that's,
basically
this
pres
and
the
subject
there
was
a
suggestion
that
it's
still
a
bit
confusing,
because
I
think
I'm
trying
to
find
that
comment
which
michael
affair,
instead
of
having
this
boolean.
F
B
It's
this
one,
so
you
were
asking
like.
Would
it
make
sense
to
use
an
enum
to
determine
to
configure
the
behavior,
should
it
be
like
known
or
just
start
broke
or
start
procaine
text
as
opposed
to
the
current
one,
which
is
like
a
boolean
flag?
Indeed
independent
one
one
for
stored
procedure
and
one
for
test.
B
So
there
is
a
like
some
corner
cases
where
the
current
one
doesn't
make
sense,
but
I
was
asking
like:
is
it
a
broken
issue
or
is
it
something
which
we
can
leave
with
I'm
just
kind
of
re-reading?
My
comment
there.
D
B
Very
ideal,
but
it's
I
think
it's
only
applicable
in
case
of
dotnet
framework.
D
D
B
Understand
question
was
like:
do
you
consider
it
as
like
a
critical
thing
to
how
or
like?
Are
we
okay
to
leave
it,
as
is
because
this
would
be
like
a
somewhat
breaking
change
for
the
dot-net
core
customers,
because
they
are
like
this
pr.
B
D
To
think
this
is
a
work
issue
where,
if
we
allow
users
to
configure
themselves
in
situations
where
they're
going
to
be
surprised
or
confused,
that's
going
to
lead
to
issues
being
open
in
the
repo
which
is
more
work
for
you,
and
I
yeah.
I
really
want
to
get
the
api
so
that
you,
what
do
you
guys
call
it
the
pit
of
success?
B
Okay,
so
it
seems
like
what
we
can
do
is
I
I
let
this
pr
merge
and
I'll
create
an
issue
to
see
whether
we
can
move
from
that
options
to
the
enum,
and
since
these
instrumentations
are
not
yet
in
one
auto
state,
we
still
have
opportunity
to
improve
it.
Let's
see
how
it
goes
for
like
next
few
weeks,
if
it's
still
confusing,
then
we'll
prioritize
converting
these
flags
into
enum.
B
I'm
okay
with
that:
okay,
yeah
I'll!
Do
that,
because
I
want
to
like
make
some
progress
here
as
well.
So
let
let
me
just
do
that
and
come
back
and
create
an
issue
just
to
tackle
migrating
or
moving
over
to
enabs
yeah.
I
really
like
the
inna
my
day.
It's
just
that
it's
like
another
round
of
work,
unless
it's
really
needed,
I'm
trying
to
avoid
that.
B
But
yeah
I
hear
you
like.
It's
definitely
useful
to
avoid
any
confusion
in
the
future.
There
is
another
pr
which
I
want
to
discuss
as
well.
I
thought
it
was
very
straightforward,
but
then
I
realized
it's
much
more
so,
michael
since
you
are
here
and
like
we
have
others
as
well,
so
I
I'm
not
yet
sure
what
is
a
scenario
where
a
dotnet
framework,
customer
or
dotnet
framework
application
would
need
access
to
both
http
web
request
options
and
http
client
options.
B
Can
you
like
remind
me
again
like
why
did
we
allow
the
dotnet
framework
customers
to
access
both
http
web
request
options
and
http
client.
D
D
D
So
when
they're
wanting
to
enrich
it
or
something
that's
what
they're
looking
to
see
so
how
it
works,
how
we
know
it
works
under
the
hood
is
http
client
is
just
creating
http
web
requests
and
firing
them.
If
you
only
use
the
http
web
request
options,
you
will
get
those
calls
you'll
just
get
an
object.
That
might
be
surprising
to
you,
because
your
code
is
using
one
thing
and
your
enriched
call
is
gonna
fire
with
something
you
don't
even
know
is
happening
which
might
be
okay.
B
B
D
I
don't
think
so
so
we're
using
that,
like
monkey
patching
code
hooks
into
the
asynchronous
state,
yeah,
okay,
so
http,
client
and
http
web
requests,
on.net
framework
use
the
same
hooks.
They
should,
I
think,
okay,
so
will
the
http
client
options
ever
be
used?
I'm
not
sure
I
have
to
go
double
check.
Okay,.
B
Yeah,
this
would
be
good
to
clarify,
because
I
mean
this-
this
took
a
lot
of
time
to
understand
like
why
this
was
not
working.
So
my
this
pr
was
just
rearranging
the
order,
because
I
I
was
assuming
that,
if
you
are
in
dotnet
framework
you,
you
would
likely
to
deal
with
you're
likely
to
configure
this
one
like
http
web
request,
instrumentation
options
as
opposed
to
http
client
instrumentation
options
like
since
both
are
like
nullable.
B
Typically
people
would
just
use
this
one
and
leave
the
last
parameter.
So
that's
that
was
my
motivation
for
like
making
this
vr
thing
like
yeah.
So
let
me
open
the
original
place.
So
there
is
someone
who
was
trying
to
this
is
http
in
dotnet
framework.
So
this
is
the
http
writing.
So
this
was
like
applicable.
The
first
parameter
is
only
about
configuring.
D
B
B
B
F
I
think
that
there
was
some
like
I
don't
know.
I
think
that
there's
something
about
there's
the
ability
to
use
http
client
but
httpclient.net
framework
uses
http
web
request
behind
the
scenes
or
something
in
in
most
scenarios,
but
I
think
that
there
is
some
way
like
michael
suggested,
where,
if
you
reference
a
nougat
package
directly
like
a
net
standard
version
of
a
nuget
package
directly,
then
you
might
get
the
http
client
behavior.
F
B
Yeah,
so
if
it's
like
not
like
a
real
recommended
way,
then
we
could
like
just
remove
it
to
avoid
this
confusion
so
like
if
you
are
in
dotnet
framework.
The
only
thing
you
have
is
http
web
request
option.
If
you
are
in
dotnet
core,
you
will
get
http
client,
instrumentation
options.
B
Yeah
I
mean
I
I
can
like
do
like
one
more
round
of
investigation
and
come
back
and
modify
the
pr
to
actually
like
remove
it.
So
there
is
no
like
confusion
here,
yeah
yeah,
I
mean,
if
any
of
you
find
our
previous
comments
about
that.
B
Please
send
that
to
me
otherwise
I'll
try
to
in
utkarsh-
and
I
was
doing
some
research
on
it
yesterday
and
we
didn't
reach
our
conclusion
so
we'll
like
spend
some
more
time
and
try
to
like
confirm
that
we
can
eliminate
this
for
http
in
framework.
B
Okay,
yeah,
I
thought
like
this
was
very
easy,
but
then
a
relay
stock
is
not
as
straightforward,
because
this
is.
This
is
potentially
like
much
much
more
in
scope
than
just
like
reordering
something,
and
what
motivated
initially
was
there.
Was
this
command
here
yeah?
So
I
don't
know
whether
anyone
trick
like
seeing
this.
B
So
if
you
have
like
this
scenario,
where
you
have
a
console
app,
which
uses
a
class
library
which
has
the
sdk
and
instrumentation,
if
the
class
library
is
only
targeting
net
standard
and
console
app,
is
like
a
dot
net
framework,
the
console
app
gets
the
dot
net
core
version
of
sql
client.
That's
what
this
pr
was
all
about.
There's
a
repro
but
utkash
and
I
tried
to
do
a
repro
for
the
same
with
http
client.
B
So
you
basically
have
a
class
library
which
uses
http
client
and
have
it
run
from
console,
but
somehow
it's
automatically
picking
up
the
net
framework
bits
and
working
fine,
but
in
case
of
sequel
it
was
like
it
was
picking
the
dot
net
core
bits.
Only
so
that's
what
prompted
the
whole
conversation.
B
And
we
didn't
conclude
that
we
are
still
investigating
so
maybe
like
if
we
find
some
learning
we
will
report
back
next
week
like
since
you
are
in
the
call
like
did.
Was
there
like
any
thing
to
add
to
it
or
like?
We
still
want
to
spend
some
more
time
and.
C
Investigate
no,
I
think,
like
we
still
would
have
to
spend
some
more
time.
I
couldn't
really
figure
out
the
reason
for.
B
Yeah
I'll
try
to
ping
like
someone
else.
If
you
cannot
find
the
answers,
because,
okay,
I
didn't
have
like
any
clue
like
what
was
going
on
so
yeah
we'll
come
back
to
it
and
see
if
that
also
helps
us
like
to
solve
the
http
client
instrumentation
issue
as
well.
So
that
would
be
a
good
message.
B
Okay,
I
don't
have
any
other
items
in
agenda.
Let
me
see
if
there
is
anything
else
yeah
there
is
none
for
now.
If
there
is
anything
which
anyone
wants
to
discuss
right
now,
it
be
happy
you
asked
about
logging
last
week
like
is
there
anything
which
you
wanted
to
specifically
ask
in
terms
of
like
exporter
or
like
logging
support
or
anything
at
all,.
E
No
not
right
now,
my
main
ask
was
if
there
is
any
feedback
on
a
login
data
model
that
we
can
share
from
dotnet
site,
and
I
got
your
answer.
I
think
we
generally
find
so
yeah.
Let's
wait
for
I
mean
I
know
if
you
were
planning
to
work
on
vlogging,
I
would
be
interested
to
at
least
watch
what's
happening.
B
Yeah,
I
think
we
want
to
like
do
a
couple
of
work
items
which
was
open
like
michael,
was
looking
at
providing
now
there
is
a
tr
about
this
is
about
like
converting
from
one
pillar
to
another.
Basically,
you
have
logs
and
you
like,
attach
the
log
record
into
the
activity.
If
there
is
an
activity,
so
basically
converting
logs
into
activity
and
technically,
there
is
like
other
way,
also
possible,
like
from
activity
you
can
like
activity
events.
B
If
your
backend
doesn't
support
event
model,
you
can
just
convert
that
into
log
and
send
it
to
your
logging
system.
So,
but
that's
one
thing,
but
while
working
on
that
aspect
we
added
support
for
scopes.
B
The
eye
logo
scopes.
There
are
some
issues,
so
we
haven't
sold
it.
I
was
just
having
that
conversation
with
michael
earlier.
We
don't
have
a
good,
perfect
solution
yet
so
we
would
be
like
I'm
very
much
interested
in
solving
that
problem,
because
many
people
were
asking
like
how
do
I
enrich
the
logs?
There
is
no.
B
I
mean
there
is
no
easy
answer
that
scopes
was
considered
the
right
answer,
but
it
looks
like
it
has
some
issues
because
it's
tied
to
the
execution
context
and
depending
on
whether
you
use
like
a
processor
or
exporter
in
the
async
fashion,
like
if
you
use
a
batching
processor,
then
by
the
time
you
reach
your
exporter,
you
already
lost
the
contacts,
so
you
cannot
really
do
anything
with
the
scope.
So
so
that's
one
area
which
I
think
I'll
be
spending
some
more
time.
Michael,
is
definitely
working
on
that
as
well.
B
So
that's
the
only
thing
which
I'm
like
actively
aware
of
in
loging
said
and
alan
made
it
a
demo.
I
think,
two
weeks
back
about
an
exporter
which
exports
into
otp
protocol
from
the
logo
code
yeah.
So
these
are
the
updates.
I
think
we
agreed
that
we
will
not
like
ga
release
this,
because
it's
the
data
model
is
not
yet
stable,
but
we
can
do
a
separate
like
known,
stable
version
of
the
log
exporter.
B
So
again
like
is
there.
Like
anything,
I
mean
anyone
from
google
who
who
is
interested
in
using
and
giving
some
feedback,
because
we're
not
like
allen
was
raising
this
concern
like
we
don't
know
whether
it
is
the
right
approach
to
use
the
otlp
exporter
to
export
logs
and
then
have
the
ot
collector
export
it
to
your
backend
choice
or
should
be
used
like
some
other
technologies
like
since
for
logging.
There
are
many
other
things,
so
I
just
don't
have
the
answer.
B
So
I'm
just
curious
if,
like
these
are
some
of
the
problems,
which
you
are
also
trying
to
find
answers
to.
E
Yeah
there
are
many
issues
and
conversion
from
logs
to
activities
to
something
that
also
discussed
on
slack,
and
we
will
be
talking
like
they
will
be
talking
a
little
bit
confidence
about
that.
But
in
any
sense
my
main
driver
for
this
question
was.
I
wanted
to
understand.
E
If
otop
data
model
is
good
enough,
so
we
can
start
finalizing
it,
maybe
not
right
now,
maybe
a
little
bit
later,
mostly
because
we
had
this
issue
with
metrics
when
data
model
wasn't
stable
and
it
caused
a
lot
of
other
things
to
not
come
in
the
time.
So
since
log
model
was
around
for
a
very
long
time-
and
I
know
that
some
people
already
using
it-
I
was
thinking-
maybe
we
can
collect
more
feedback
and
understand
what
we
can
do
and
maybe
we
can
finalize
it
way
earlier.
E
B
E
Yeah,
thank
you
for
update.
I,
I
will
take
a
look
at
issues
that
you.
B
G
Ahead,
mr
okay,
so
this
is
a
question
for
the
group:
I've
been
working
with
noah
in
the
net
stuff
and
tracking
the
metrics
api.
You
know
spec
development,
so
the
question
for
the
group
is,
I
have
some
code,
that
is
a
re
interpretation
of
the
current
api
spec
and
I
have
this.
In
my
you
know:
branch
of
noah's
branch.
The
question
is:
does
the
does
this?
Sig
would
be
interested
in
seeing
that
in
the
metrics
branch
or
and
if
so,
then
I
can.
G
You
know
check
in
you
know
you
know,
do
a
pr
for
whatever
I
have
so
that
people
could
see.
You
know
an
interpretation
of
it
or
if
you
guys
are
not
interested,
then
I'll
just
continue
to
work
in
you
know
the
other
branch
and
you
know,
along
with
noah,
on
the
dotnet
portion.
So
so
that's
the
question.
B
I
mean
it's
not
like
throwaway
code
ready.
You
intend
to
like
use
it
sometime
in
the
future
right.
G
Well
that
I
think
that
depends
on
a
few
things,
starting
off
with
I
think,
for
net.
I
think
the
the
current
structure
is
that
we
will
likely
just
adopt
the
net
api
per
se,
so
I
don't
know
when
the
net
api
is
said
and
done
how
closely
I'm
sure
be
very
close
to
what
the
api
spec
metric
spec
has
been.
You
know
is
written,
but
as
it
stands
today,
it's
not
exactly
the
same
form
for
it,
and
noah
is
continuing
to
refine
that.
G
So
in
the
meantime,
the
question
was:
would
anybody
be
interested
in
in
just
a
interpretation
of
the
current
metrics
api
spec?
Now
there's
no
sdk,
there's
not
you
know,
there's
not
much!
You
could
do
it's
just
a
well.
It's
an
interpretation
of
the
of
the
api
spec
as
it
stands
today.
B
G
You
know
completely
independently
of
everything
else
and
attempting
to
to
quote
write
a
api
based
on
the
spec
so
and
I'm
following
that
spec
very
closely
in
terms
of
what
the
api
spec
is
asking
for
and
so
forth,
and
so
this
is,
you
know
this
is
like
I
said,
an
interpretation
of
the
api
stack.
B
So
I
was
thinking
like
the
api
would
anyway,
whatever
api
we
have
in
matrix
branch
would
anyway
be
like
thrown
away
and
replaced
with
the
one
from
dot
net.
So
what?
What
is
the
intention
of
like
having
api
in
this
repo?
Is
it
just
to
like
have
something
to
work
on
until
the
time
comes,
for
when
dot
net
will
actually
give
you
a
build,
correct,
matrix,
api,
correct
yeah?
I
think
I'm,
okay
with
that,
we
can
like
start
working
on
it
like
whenever
you
are
ready.
G
B
I
think,
like
once,
we
actually
start
working
on
it
pretty
much.
Everything
in
the
metrics
like
including
sdk
might
be
like
removable,
because
if
this
was
written
based
on
the
old
right,
that's
very
likely
the
case.
Yes,
that's
correct!
So
if
that's
the
case,
I
mean
I
was
anyway
plan
to
do
that,
like
two
weeks
from
now
when
like
when
we
start
getting
like
some
bits
from
the
dot
net,
but
if
you
want
to
do
like
it
earlier,
I
am
totally
fine
with
that.
B
Just
submit
pr
against
the
matrix
branch,
which
shows
like
new
set
of
instruments
which
match
the
current
spec,
and
in
fact,
if
you
want
you
can
delete
it,
but
then
deletion
means
you'll,
delete
all
the
way
to
sdk
tests.
Examples
prometheus
exporter
all
the
way.
G
So
well
so
so
I
I,
I
don't
think
I
understand
what
you're
saying,
but
I
think
the
metric
spec
is
not
ready
for
all
of
that
at
the
moment,
because,
even
even
after,
like
a
week
or
two
weeks
from
now
at
most,
what
we
will
have
in
the
spec
is
basically
the
api
portion
with
maybe
the
instruments
predefined,
and
we
could.
You
know
someone
could
quote,
read
the
spec
and
interpret
and
implement
a
version
of
the
spec,
but
there's
still
no
sdk
and
there's
still.
No.
G
You
know
anything
that
you
could
really
do
with
it,
and
so
the
process
of
tracking
it
and
continuing
to
do
that
is
is
an
ongoing
work.
And
so
I
I
I
don't
think
it
I
would
like.
I
don't
think
I
would
want
to
just
you
know,
remove
or
destroy
whatever
current
you
know.
Metrics
code
is
right
now,
because
what
you
guys
have
is
a
working
version,
albeit
you
know,
deprecated,
but
still
fully
working.
G
B
Yeah,
so
I
think
if
you
allow
me
like
one
more
week,
I
think
I
need
to
spend
like
some
more
time
to
see
what's
the
state
of
a
spec
and
maybe
like,
if
you
can
delete
this
and
start
over.
That
also
is
probably
fine,
because
a
lot
of
things
changed
since
we
wrote
this,
so
it
might.
B
Like
easy
to
just
delete
this
and
start
over
because
otherwise
I
mean
the
only
issue
is
like
I,
I
was
delaying
it
is
when
I
started
doing
it,
I
want
to
add
the
new
api
parallelly
or
at
the
same
time.
So
that's
why
I
was
saying,
like
we
started
like
around
third
week
of
april
when
at
least
the
api
from
a
matrix
api
from
a
specs
perspective
is
marked
like
somewhat
like
good,
to
go.
G
B
Back
to
this
problem
like
next
week,
and
then
we
can
start
deciding
like,
should
we
like
know
this
first
and
then
submit
your
pr's
or
should
we
keep
this
and
do
like
metrics?
We
do
like
when
you
submit
pr
like
submit
it
under
the
folder
like
metrics.
We
do
so
that
part.
I
think
we
can
like
discuss
next
week.
I
haven't
really
thought
about:
what's
the
exact
logistics
on
how
to
do
it,
so
we
we
can.
I
I'll
ping
you
offline
and
we
can
figure
out
what.
B
Of
course,
we
should
get
there.
Yeah
yeah,
we
initially
decided
that
matrix
code
will
be
going
into
matrix
branch
and
we
already
set
up
the
ca
and
cd
to
ci,
basically
targeting
this
branch,
but
that
I
said
after
I
did
that
there
was
no
work
in
the
metrics
at
all
like
zero,
so
it
it
just
like
is
abandoned.
So
I
need
to
like
get
back
to
it
and
see
if
it
is
time
for
it
for
us
to
like
fully
remove
this,
like
we
don't
think
we
ever
implemented
the
observer
part.
B
B
Yeah,
I
will
think
you
offline
victor
to
like
sure,
of
course,
what's
the
best
way
to
like
organize
this
work
and,
of
course,
report
back
in
the
sick
meeting
by
next
week
and
personally,
like
I'm
like
going
to
start
work
on
metrics
like
mid
or
maybe
like
another
two
weeks
from
now.
So
until
then
like
it's
whatever
victor
has,
is
what
we
have.
B
Questions
okay
looks
like
we
don't
have
anything
else
to
discuss,
so
we'll
stop
give
back
three
minutes.
Thank
you.
Everyone
we'll
see
next
week.