►
From YouTube: 2020-08-04 .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).
A
B
Numbers
yeah
just
make
sure
you
put
the
number,
so
it's
very
easy
to
count.
How
many
are
there?
B
B
The
agenda
is
fairly
big
compared
to
previous,
so
yeah.
I
think
we
can
start
it's
one
minute.
We
have
most
of
the
folks
here.
C
D
Let's
talk
about
me
yeah,
I'm
I'm
working
on
gke
a
lot
and
but
still
open
telemetry
is
my
like
20
project
here
in
google.
Gte
sorry
is
this.
E
D
Google
kubernetes
engine,
and
mostly
it's
signaled,
it's
a
part
of
kubernetes
that
manage
all
the
containers
on
the
node
and
all
related
stuff.
D
B
B
The
exact
date
will
depend
on
the
data
in
which
dot
net
diagnostic
source
preview
8,
because
we
just
updated
ourselves
to
preview
eight.
There
are
like
bunch
of
interesting
changes
which
we
asked
for
so
we
are
making
progress
there.
I
could
say
like
halfway
through
there
and
we'll
be
working
on
remaining
changes
this
week
and
next,
if
there
are
like,
I
mean,
if
you
look
at
the
milestones,
it
should
be
like
update
now
updated.
B
Now,
let's
take
a
quick
look,
so
this
is
a
next
one,
which
I
have
put
august
26
as
the
date
and
followed
by
two
more
betas,
maybe
one
more
before
number
end,
when
we
officially
plan
to
release
g8
and
most
likely,
this
will
align
with
dot
net
release,
because
this
one
is
going
to
be
aligned
with
a
preview
8,
which
is
the
last
preview.
The
next
rc
one
is
coming
around
september
mid,
so
I
put
10
dataly
put
18
and
then
rc2
on.
I
think
it's
number
13..
B
Oh
yeah,
I
need
to
update
that
thanks
for
catching
that
yeah,
so
ga
is
still
number
30.
That
number
first
second
week
is
when
dotnet
goes
ga,
so
we
should
be.
We
should
still
have
comfortable
time.
It's
like
approximate
date,
because
I
don't
know
exactly
when
dot
net
will
release.
B
It
could
be
like
off
by
a
day
or
two
yeah
I'll
go
ahead
and
change
this,
but
generally
we'll
have
a
beta
every
month
until
we
release
ga,
that's
the
general
idea
and
the
issues
have
been
assigned
correctly
only
to
the
immediate
milestone,
I'm
like
very
defensive
in
assigning.
So
it's
quite
possible
that
we'll
do
much
more
than
like
the
311.
B
It's
just
that
not
I
mean
I'm
just
assuming
I'm
the
only
one
working,
even
though,
like
several
people
are
working,
but
it's
probably
fine.
If
there
are,
if
there
are
any
issues
which,
if
people
are
committing
to,
then
we
can
adjust
the
milestones
correctly.
B
B
B
It
requires
support
from
top
net,
so
we
have
an
issue
opened
in
dotnet
side,
since
we
missed
all
the
previews,
it's
going
to
be
rc1,
so
we
still
have
like
a
small
window
of
opportunity
to
make
that
happen,
but
while
making
that
change,
we
are
also
making
some
changes.
Name
changes
as
well.
We
reviewed
some
feedback
earlier
as
well
about
like
activity
listener
dot.
These
names
are
not
very
friendly,
so
this
is
a
good
opportunity
to
propose
alternate
names.
B
Let's
I
don't
want
to
discuss
names
here.
Please
do
share
the
names
suggestions
right
here
in
the
issue
which
I
have
linked
here.
That's
pretty
much
it
from
my
side.
Now
we
can
go
over
the
rest
of
the
agenda.
E
Yeah
I
had
a
quick
question
about
the
the
update
there
on
the
the
note
for
no
work
on
metrics
planned
on
the
next.
You
know
for
the
next
milestone
or
the
next
beta.
Do
we
plan
to
do
it
and
work
on
that
and
any
of
the
other
betas
coming
up
to
ga.
B
I
would
assume
yes,
but
a
lot
depends
on
whether
the
matrix,
sdk
specs
are
also
ready,
because
right
now
there
is
no
sdk
specs,
so
we
can
only
work
on
apa,
but
it's
kind
of
I
mean.
I
think
it
is
better
if
you
wait
for
actual
spec
to
come,
so
that
it's
much
easier
to
make
progress.
There.
B
Okay,
next
yep
yeah-
I
mean
I
don't
I
don't
the
matrix
spec
for
sdk
is
still
not
there
and
it
is
tagged
as
requirement
for
ga.
So
I
expect
there
should
be
some
progress
there,
but
like
I'm,
not
committing
to
doing
it
in
this
immediate
one,
but
if
there
are
progressing
matrix,
we
definitely
pick
it
up
for
0.6
meter.
B
Okay,
before
we
go
into
riley's
list
of
things,
let's
quickly
cover
two
topics
from
last
meeting
which
we
didn't
get
time
to
do
it.
So
this
was
oh,
I
think
riley
raised
this
one
as
well.
B
Currently,
the
workflow
is
like
we
release
to
my
get
and
you'll
get
packages
to
master
so
like.
If
you
have
10
years
margin
a
day,
we
have
10
releases
to
market,
so
riley
was
wondering
like.
Should
we
completely
get
rid
of
my
get
and
push
everything
to
negatively
and
have
it
unlisted?
Is
there
any
like
right?
You
want
to
talk
about
like
what
you
want
that
way,
as
opposed
to
keeping
my
kid
for
like
early
early
adopters
and
you
get
for
like
actual
public
customers.
B
I
kind
of
forget
like
what
was
your
reasoning
for.
F
Shifting
okay,
I'm
talking,
I
actually
don't,
have
strong
objection
to
like
my
guide
or
new
guide
or
whatever,
like
like
package
repo.
The
question
was
coming
from
the
customer
where
they
want
to
take
the
the
open
telemetry
like
certain
release,
and
currently
we
only
have.
We
only
had
three
releases.
F
I
believe,
and
in
this
way
like
if
the
customer
noticed
some
issue
and
we
fixed,
and
we
released
that
and
either
we
tell
the
customer,
wait
like
for
a
couple
months
until
we
we
have
the
next
version
published
to
new
guide
or
we
tell
them
like,
go
and
remove
new
guide
and
switch
to
my
guide,
and
this
causes
a
lot
of
confusion
for
the
customer.
So
I
want
to
know
like
moving
forward
like
at
the
end
of
the
game.
Once
we
have
ga
version.
F
What's
the
story
like
you
have
hot
fix,
I
think
we
should
just
go
and
release
a
hot
fix
on
you
guys.
If
there's
a
new
preview
version,
we
should
just
release
a
preview
version
on
you
guys
mark
it
as
a
preview
based
on
the
semantic
version.
In
this
way,
it
seems
my
guide
is
only
for
us
for
people
who
are
working
on
this
repo
directly.
I
want
to
know
like
what
what
are
thinking
from
the
other
folks
here.
D
D
I
I
think
I
mean
I
think
one
of
the
issue
is
open.
Telemetry
was
that
we
didn't
do
releases
so
alpha
releases,
often
enough
on
a
new
kid,
so
yeah,
if
you
start
switching
to
like
monthly,
it
shouldn't
be
such
a
big
issue,
but
I
mean
I
might
be
wrong.
I
don't
know
these
are
specific
people
who
asking
for,
like
very
last
update
that
we
need
to
share.
F
No,
I
mean
like
if
we
have
something
that
we
didn't
call
that's
ready
for
a
new
guide
release
and
are
we
going
to
ask
people
hey,
please
remove
new
guide
from
your
your
new
guide,
config
and
switch
to
my
guide
and
that
that's
creating
some
confusion,
and
the
second
part
is
the
question
about
the
long
term.
If
we
have
ga
like.
D
Guide
yeah.
I
think
it
will
be
a
matter
of
escalation
procedure
for
like
a
hot
fix.
So
if
we
believe
that
we
cannot
wait
another
month
till
the
next
release
and
we
need
to
escalate
it
as
a
hot
fix.
F
Yeah
that
used
to
be
a
little
bit
strange,
because
even
if
it's
like
not
ready
for
a
release,
I
mean
for
hotfix
or
any
like
release.
That
means
we
need
to
pay
a
lot
of
attention
for
the
quality
and
if
it
is
not
a
hard
fix,
it's
like
a
preview
feature.
I
think
new
guide
already
have
that
preview
feature.
You
can
tag
a
package
as
preview,
so
people
who
are
brave
enough
they
can
specify
when
they
try
to
install
the
new
web
package
because
specific,
hey,
I
want
preview.
F
F
Is
using
new
guide
and
they're
so
so
far,
they're
happy
with
the
the
1.0
ga
release
and
they
notice
a
small
issue
and
they
ask
we.
We
could
fix
that
and
someone
fixed
that
in
the
new
guy
like
like
in
our
ripple
and
it's
not
it's
not
big
enough-
that
we
decided
to
have
a
hotfix
release,
but
people
would
want
to
try
that.
Are
we
going
to
ask
people
hey
in
order
to
try
this?
F
Our
nightly
building
is
on
my
guide,
so
what
you
can
do
is
to
switch
to
mygat
and
later
if
it's
released
switch
back
or
we
can
tell
them
just
turn
on
the
preview
feature.
So
you
can
take
the
preview
version
and
just
go
and
use
that
and
once
you
feel
confident
and
we
got,
we
got
the
final
release.
You
can
take
that
into
production.
D
So
I
think
it
will
wash
out
the
difference
between
previews
that
we
want
people
to
try
and
preview
that
is
just
nightly
built.
So
if
you
want
to
have
this
separate
category
of
fields
which
are
stable
enough,
that
you
want
people
to
try
them
like
current
beta
2,
then-
and
this
is
like
preview
feature
of
nuget-
that
is
doing
exactly
so
beta
2
is
a
preview
version
of
open
telemetry.
Then
we
need
to
have
my
get
as
separate
to
nuke
it.
D
F
F
I
think
in
this
way,
probably
what
I
would
suggest
is
people
take
a
preview
on
the
github
version
and
they
know
which
version
they're
taking
and
regarding
gut
there's
another
minor
issue
where
my
guide
has
this
automatic
purge
after
certain
days.
I
believe
it's
14
days
so
even
earlier.
If
we
have
too
much.
B
Packages
whenever
we
are
out
of
space
it
rolls
it
like,
it
does
say
like
a
rolling
thing,
so
it
deletes
even
earlier.
If
you
have
too
much
activity
per
day,.
F
F
F
Probably
not
I
really
I
mean
having
a
like
periodic
release
should
be
good
enough
and
for
the
bleeding
edge.
I
I
think
that's
only
like
very
helpful
for
developers
working
on
this
record
directly,
instead
of
like
consumers
like
we're,
pushing
changes
like
crazy,
and
I
don't
think
it
makes
sense
for
customers
to
download
that
package
and
we
actually
look
at
the
download
number
from
myguide.
You
can
see
like
most
of
the
packages.
They
have
zero.
B
Yeah:
okay:
let's
move
to
the
next
one,
let
me
open
an
issue
as
well,
if
not
and
then
discuss
more
there.
B
B
It's
I
mean
I
don't
see
a
problem
with
that,
but
the
fact
that,
like
it's
being
deleted
from
I
get,
is
potentially
an
issue,
because
if
someone
takes
a
dependency
and
when
they
try
to
restore
it,
I've
run
into
that
before
and
a
bunch
of.
B
Get
builds
yeah,
so
let's
open
an
issue
and
continue
this
there,
because
we
have
to
move
to
the
other
ones
as
well.
B
So
the
next
one
is
okay,
again
related
to
releasing
packages
from
open,
telemetry
quantity
proposed
so
right
he
was
asking
who
wants
it.
So
that's
a
good
question.
We
so
far
have
never
released
anything
from
this
repo
and
we
never
had
the
need,
I
would
say
the
only
thing
which
exists.
There
are
instrumentation
for
assure
and
edt
framework
core,
and
now
we
have
stacked.
Oh
yeah,
we
already
had
stackdriver
as
well.
B
There
is
a
good
request
to
bring
something
else,
so
we
need
to
decide
who
on
statement
right
now,
it's
owned
by
the
same
folks
as
the
open
elementary.net
report.
Permissions
were
inherited,
so
what
riley
was
asking
more
was
like
who
owns
it?
Who
is
responsible
for
the
contents
published
from
this
report?
Is
that
what
really
you
are
asking.
F
Yeah
so,
for
example,
who's
going
to
put
the
readme
file
and
define
the
process,
and
if
people
ask
like
questions,
they
put
issues
who's
going
to
trigger
the
issue
and
who's
going
to
decide
when
we're
going
to
release
and
if
certain
thing
got
blocked
and
the
owner
of
a
certain
component
is
not
showing
up.
So
what
are
we
going
to
do?
F
I
asked
this
question
during
the
maintenance
meeting
and
the
spec
meeting
today
and
I
I
believe
that
the
javascript
sig
is
treating
this
country
in
a
similar
way
of
the
main
report.
So
it's
the
same
maintainer
and
the
maintainer
is
doing
the
job,
so
wonder
like
from
this
product,
like
for
the
maintainers
who
who's
willing
to
maintain
this
ripple
and
who
got
the
commitment
here.
B
B
F
Different
so
python
the
maintainers
decided.
Oh,
we
don't
have
enough
energy
to
maintain
this
core
ripple
and
have
all
the
country
packages
so
they're
pushing
the
control
packages
to
its
own
repo.
And
this
way
you
can
say
the
reason
they
want
to
put
that
into
a
different
thing
is
not
because
they
want
to
manage
that
in
two
different
places.
It's
because
they
don't
want
to
manage
that.
You
see.
B
Okay,
so
I
think
there
is
no
central
guidance
from
like
open
telemetry
org.
So
it's
up
to
us
to
do
what
we
want
to
do
with
the
report
like.
If
the
maintenance
have
enough
energy,
we
can
say:
okay,
we
own
it
we'll
continue
to
own
it
in
the
future,
but
otherwise
we
had
to
say:
okay,
it's
like
something
like
experimental
thing
we
release,
but
there
is
no
guarantee
that
we
will
respond
to
issues
or
pull
requests
in
any
any
particular
like
slow
kind
of
thing.
F
Yeah,
and,
and
also
even
if
one
maintainer
has
special
passion
and
energy
that
doesn't
mean
the
other
maintainers
have
that
commitment.
So
we
need
to
have
the
some
like
idea
about
what,
if
that
maintainer
decided
to
quit,
and
are
we
going
to
abandon
this
project
at
all
or
we
can
carry
forward
with
the
other.
F
Behind
contributing
something
else,
I
I'm
thinking
the
the
other
way.
I
think
the
goal
for
the
contributor
is
never
been
like.
We
want
to
promote
something
to
the
main
ripple.
It
is
actually
whatever
doesn't
fade
into
the
memorable,
but
still
we're
seeing
value
and
people
want
to
put
that
in
the
open
telemetry
umbrella
under
that
umbrella,
then
we
put
that
under
contrib
or
it's
like
for
python.
B
I
see
yeah
I
mean
I
don't
really
think
I
personally
would
have
the
energy
to
support
all
the
things
which
can
potentially
come
here
like
because
I
mean
I
don't
really
know.
B
Why
would
we
even
host
stackdriver
here,
because
it
would
be
specific
to
a
particular
company
right,
so
it
should
eventually
move
to
some
google
on
repo,
not
open
elementary
one,
but
I
can
understand
like
if
energy
framework
or
is
kept
here,
then
it
would
make
more
sense
for
us
to
keep
supporting
it
and
support
in
the
same
way
as
the
main
triple
yeah.
But
so
far
we
haven't
defined
like
what
is
that
entry
criteria,
because
stackdriver
and
show
thing
makes
me
one.
B
We
put
it
like
temporarily,
like
in
the
initial
thing,
but
we
never
like
set
aside
any
or
set
forth
any
guidelines
on
what
goes
here.
H
H
B
F
B
Like
any
guidelines
here,
so
yes,
okay,
go
ahead.
I.
D
I
think
last
time
what
we
discussed
is
we
want
to
put
here
like
there
is
a
support
per
se,
not
like
currency,
but
like
promise
that
we
have
a
main
repository
and
we
want
to
put
the
main
repository
all
the
components
that
we
want
to
support
like
actively
and
like
everybody
interested
in
us,
and
here
we
discussed
last
time
that
we
want
to
put
components
that
likely
will
be
shared
among
many
people.
D
So,
if
you
want
to,
I
don't
have
a
strong
opinion
about
stackdriver
where
it
should
be,
but
if
there
is
a
like
component
that
edits
metadata
from
gcp,
that
probably
will
go
here
in
this
repository,
because
it's
a
it's
something
that
is
less
vendor
specific.
I
mean
it's
very
specific
because
it's
hosted
some
place,
but
it's
less
than
the
specific,
because
it's
not
an
expert,
so
it's
not
a
back-end
so
but
ownership
like,
I
think
one
question
we
can.
D
One
thing
we
can
do
is
to
change
the
name
of
packages
clearly
indicating
that
this
country,
folder
or
maybe
maybe
have
like
openclimate.com.exporter
or
dot
config
instrumentation.
So
it's
clearly
indicated
it's
not
it's
less
supported
component
and.
D
D
Yeah,
but
if
bug
is
not
interesting
to
any
maintainer
to
fix,
like
maintainers,
have
no
guarantee
that
this
bug
will
be
fixed,
we
can.
We
can
just
keep
the
buck
on
the
back
burner.
If
nobody
cares
about
this
buck.
F
The
bug
will
got
automatically
closed.
That's
what
we
discussed
in
the
maintainers
meeting
right.
If
nobody
is
pigmented,
it
goes
we're
not
leaving
that
as
a
like
unclear
state,
I'm
not
sure
how
it's
different
from
this
airport
for
this
question
is
like
we
got
a
few
maintainers
here,
so
my
question
is
sergey:
are
you
going
to
become
a
maintainer
and
do
the
job
here
or
you
think
you're
relying
on
the
others
or
we
don't
know.
D
So
the
main
job
is
to
I
just
publish
packages.
This
is
a
main
job
or
no
I'd
like.
F
B
D
I
gave
you
in
general,
so
we
need
to
like
at
least
initial
push
to
make
it
shippable
and
like
actually
ship.
It
is
very
important
later
like
when
we
will
see
that
something
to
create
and
rotten.
We
need
to
decide
case-by-case
basis.
I
think
that
is.
D
I
mean
this
is
quite
typical
in
open
source
projects
when
there
is
an
issue
and
you
can
get
maintainers
attention,
you
just
like
bank
hard
harder
and
then
eventually
it
will
blow
out
and
like
we
will
need
to
do
something
about
it
and
I
I
don't
know
like
do.
We
need
to
define
a
specific
person
to
do
that
or
we
just
keep
the
same
maintenance
in
our
cougars
structure
as
typical.
D
D
D
I
can
help
with
that.
I
think
it's
it's
important
to
make
it
reusable,
and
I
really
like
that
somebody
came
with
this
mass
transit
library
instrumentation.
I
think
it's.
F
Yeah
so
there's
a
branding
issue,
for
example
like
stackdriver,
I'm
not
even
sure
like
whether
it
should
be
under
like
published
by
open,
telemetry
or
published
by
google
and
and
what's
the
guidance.
If
the
package
has
issue,
should
people
file
an
issue
here
and
if
that
particular
question
about
the
backend
about
pricing,
where
should
they
go
so
all
these
questions?
I
think
we
need
someone
to
maintain
this
and
answer
those
questions
to
bring
the
clarity
sure.
D
I
I'm
with
you
here
so,
let's
discuss
branding
issue
over
email
or
maybe
on
like
maybe
create.
B
An
issue
right
here:
yeah
yeah,
create
an
issue
right
here,
because
there
are
things
which
we
like
I
mean.
If
you
just
exclude
the
google
specific
stuff.
Other
things
are
not
really
tied
to
a
particular
vendor
like
it
should
be
like
more
generic,
and
if,
unless
there
are
like
any
issues
or
we
are
not
doing
any
hacky
code,
we
should
probably
promote
them
to
the
main
repo
itself.
B
I
mean
I
think
I
asked
to
open
it
here
in
this
repo,
but,
like
I
said,
I
didn't
have
like
any
set
guidelines
like
this
is
the
reason
why
I'm
asking
it
it
just
felt
like
we
were
having
too
many
packages
in
the
main
repo,
so
yeah,
let's
open
an
issue.
I
think
I
can
open
one
and
start
discussing
like
what
should
we
keep
here
and
what
should
be
the
guideline
for
keeping
things
in
this
repo,
as
opposed
to
the
main
repo,
and
we
can
also
debate
what
would
be
the
support
strategy
like?
B
Is
it
our
responsibility
and
same
with
branding
should
be
branded
as
open
telemetry
or
create
a
open,
telemetry
contrib
org
in
you
get
to
publish
it
as
well.
Okay,
really,
is
there
anything
else
you
want
to
add
for
that,
so
I
will
create
an
issue
with
the
follow-up,
but.
F
D
B
Okay,
all
right
so
now
I
think
yeah
we
have
a
bunch
of
questions
from
riley,
but
I'll
go
to
eric
for
his
question,
because
it's
just
one
as
opposed
to
ten.
So,
let's
see
so
eric
are
you
on
call?
Can
you
walk
us
through
questions
yeah.
E
Yeah,
it's
it's!
It's!
It's
not
a
question.
It's
it's
just!
I
just
wanted
to
make
a
quick,
quick
announcement
here.
I
think
this
will
be
of
interest
to
some
of
the
people
here
in
this
group
and
maybe
more
of
interest
in
the
auto
instrumentation,
but
we
have
open
sourced
all
of
our
all
of
our
instrumentation,
and
that
includes
our
our
agents.
E
So
if
you're
you're
interested,
of
course
you
can,
you
can
look
here
and
see
see
more
of
the
details,
and
the
other
thing
to
note
here
is
that
we're
just
we're.
I've
been
talking
a
little
bit
about
this
before,
but
we're
making
you
know
public
commitments
to
standardizing
on
open
telemetry,
and
so
that
means
we'll
be
continuing
to
commit
resources
and
new
resources
as
well
too,
and
we
have
chris
chris
ventura
on
the
line
as
well
who's
one
of
the
team.
E
I
Okay,
so
my
name's
chris
ventura
I've
been
an
engineer
at
new
relic
for
a
few
years
now,
primarily
working
on
the.net
agent,
so
working
on
auto
instrumentation,
making
changes
to
through
the
profiling
apis,
adding
instrumentation
for
things
like
wcf,
on
both
the
server
side
and
the
client
side.
I
C
Yeah,
okay,
thank
you
guys,
just
a
question.
Would
it
would
you
like
to?
Would
you
be
willing
to
to
maybe
not
right
now,
but
at
some
point
soon,
to
give
a
little
bit
of
a
like
sort
of
presentation
that
gives
us
a
overview
of
your
agent
architecture
and
technology
that
might
be
more
suitable
for
the
auto
instrumentation
sick,
but
still
a
question.
C
C
I
No,
I
think,
I
think,
half
an
hour's
half
an
hour
would
be
good.
I'd
just
be
focusing
on
more
of
the
instrumentation
set
of
things
and
how
that
works.
C
Yeah
it
just
you
know,
I
would
be
really
interested
to
see
about
to
learn
about
differences
and
similarities
for
for
the
data
dock
and.
C
Cool,
so
maybe,
since
tomorrow
is
our
first
meeting,
we
might
have
all
sorts
of
like
random
things.
Maybe
if
we
start,
if
you
target
the
week
next
week
so
tomorrow
in
a
week
how
about
then
you
also
have
time
to
prepare
and
everything.
I
Yeah
so
I'm
available
for
tomorrow,
but
if
there's
a
meeting
the
following
week,
I
probably
won't
be
able
to
make
it
I'll
be
on
taking
a
brief
vacation.
C
Oh
okay,
so
maybe
we
then
can
start.
Maybe
then
we
can
start
target
tomorrow
and
check
with
the
other
folks
whether
they
are
okay
moving
some
of
the
items
to
the
week
after
those
who
are
on
the
court
should
they
drink.
I
have
any
position
for.
C
C
Okay,
I
hear
nothing
so
if,
if
that
works
for
you,
if,
if
this
is
enough
time
for
you
to
prepare,
I
bet
like
I
don't
know,
people
who
are
on
on
who
are
also
interested
in
auto
instrumentation,
who
are
on
the
core
right
now.
Will
you
guys
also
be
interested
in
such
a
presentation.
G
C
I
see
that
some
of
the
folks
who
are
typically
in
in
the
in
the
autumn,
instrumentation
space
they're,
not
on
the
core
right
now.
So
that's
why
we
don't
yeah.
B
B
C
I'll
do
that
so,
and
I
also
put
a
send
a
message
in
the
chat
so
that
people
know
that
you're
only
available
tomorrow
and
not
another
day.
So.
J
J
Okay,
yeah.
F
So
I
have
a
couple
of
questions
and
I
and
I'm
trying
to
think
like
how
I
can
organize
those
questions
to
lead
for
a
like
constructive
conversation.
So
my
current
thinking
leads
me
to
believe
that
we
can
start
with
some
tutorials.
As
I
start
to
read
tutorial
and
documents,
I
have
some
very
simple
projects
as
trying
to
demonstrate
with
a
single
file
like
20
lines
of
code.
F
This
is
what
you
can
achieve
by
writing
your
own
sampler
by
writing
your
own
processor
or
any
like
component,
that
you
interact
with
open
telemetry
and
I
start
to
notice
some
design
issue.
For
example,
should
I
pick
this
broadcast
processor,
or
should
I
have
multiple
pipeline,
or
should
I
even
have
different
trace
provider
and
each
one
have
this
own
pipeline
and
do
I
put
the
sampler
in
front
of
all
the
pipelines
or
the
sampler
can
be
put
in
the
special
processor?
F
So
those
are
the
questions
and
I
start
to
see
something
working,
something
not
working
or
something
working,
but
not
as
what
I
expected
or
not.
What
has
the
spike
was
saying
and
something
that
is
working
as
I
expected,
but
different
from
what
the
spike
was
expecting
or
something
that
that's
working
and
it's
more
like
intuitive
for
me,
but
it
is
very
different
from
what
the
the
component
was
designed
for.
So
I
have
a
couple
of
small
prs
and
I'm
trying
to
send
more
prs.
F
F
If
I'm
writing
a
document,
tell
people
how
to
use
the
processor
and
I
give
them
a
self-contained
product
and
if
they
cannot
explain
that
or
I
figured
it's
hard
so
if
we're
building
something
very
hard
for
me
to
explain
to
the
customer,
I
couldn't
even
write
a
document
or
when
I
write
the
document,
I
was
struggling,
whether
I
should
use
this
approach
or
that
approach
and
what's
the
benefit.
I
don't
see
a
clear
thing,
then
that's
a
signal.
F
Something
got
to
change,
or
at
least
we
need
to
discuss-
and
I
have
a
couple
of
these
things
so
I'll
I'll
put
people's
name
here
and
if
you
have
some
opinion.
F
Let
us
know-
and
with
this
I
hope
we
can
have
a
the
end
of
the
game-
is
to
have
a
clear
documentation
and
the
clear
self-contained
tutorial
and
the
tutorial
should
look
straightforward.
So
we're
we're
happy
with
the
outcome
of
the
tutorial
and
also
whatever
we
showed
in
the
tutorial
is
going
to
align
with
the
open,
telemetry
spec.
F
B
F
Yeah
yeah
sure
so
in
this
case,
like
I
have
a
processor
pipeline
and
also
I
have
a
trees
provider,
so
I
started
the
question
because
processor
pipeline
is
not
something
defined
in
the
spec,
so
it's
very
hard
for
me
to
understand
what
like
what
is
the
actual
processor
pipeline
like?
Is
that
a
collection
of
resources
or
that's
just
a
processor
which
has
multiple
underlying
processors
and
read
the
definition
I
kind
of
see
the
processor
pipeline
has
some
idea
about
like
exporter?
F
It
has
a
processor
like
no
the
next
hub,
but
still
is
it's
hard
for
me
to
answer
the
question
like:
why
do
we
need
the
processor
pipeline
at
the
first
place?
Can
we
just
use
different
trace
provider,
or
can
we
use
a
like
processor
instead
of
having
to
invent
another
concept?
So
these
are
the
things
and
then,
if
I
have
like
multiple
processors,
I
want
them
to
run
at
the
same
time
for
every
single
activity
so
which
one
should
I
use.
F
B
Answer
yeah,
the
pipeline
part
is
not
different
in
the
spec,
I
think
dotnet
has
it
most.
Other
languages
also
have
it,
but
it's
I
don't
think
the
word
like
pipeline
is
well
defined,
so
we
the
our
current
definition
of
pipe
training,
is
it
consists
of
processors
and
an
exporter,
so
you
can
add
any
number
of
pipelines.
So
each
of
this
pipeline
contains
one
processor
and
an
exporter.
F
Yeah
so
yeah
once
we
have,
I
I
think
people
will
look
at
the
the
example
and
see
like.
Is
this
something
clear?
Is
this
something
we
really
want
or
how
we
can
make
that
clear?
Also,
I
can
give
some
example,
for
example,
if
I
have
a
simple
exporter
that
opens
a
file
like
predefined
file,
name
like
under
the
root
drive,
something
like
that,
and
I
just
simply
take
the
activity
and
sequentialize
them
and
put
them
in
a
file.
F
Then
I
have
a
problem
if
I
need
to
have
multiple
pipelines
to
in
order
to
have
different
samplers,
but
in
the
end
I
still
want
all
the
data
to
get
merged
into
the
same
file.
I
cannot
have
two
exporters
because
they
will
compete
on
the
file,
so
in
the
end,
I'll
have
one
exporter
that
owns
the
file.
Of
course
you
can
debate,
we
can
have
two
exporters
that
have
an
underlying
shared
file,
but
that
doesn't
seem
to
be
very
intuitive.
F
So
in
this
case,
if
I
have
one
exporter,
can
I
use
that
exponent
for
multiple
pipelines
or
processors?
At
the
same
time,
according
to
spec?
I
cannot
because
the
exporter
seems
to
be
owned
by
the
processor
and
when
the
processor
got
finished,
the
filter
will
got
flashed
and
then
disposed,
and
the
problem
is:
if
you
use
that
for
multiple
processors,
should
we
give
the
user
an
error
like
the
next
one,
who's
going
to
register
the
exposure
we're
going
to
tell
them?
Oh,
you
got
an
exception.
F
The
exposure
is
already
taken
and
you
cannot
take
that
again
or
should
we
change
the
behavior
make
better
reference
content,
so
the
next
one
will
actually
add
a
reference
count
and
then
the
question
is
who's
going
to
call
the
flash
the
first
one
or
the
last
one
of
all
or
all
of
them
or
is
based
on
the
reference
counting
behavior.
F
F
B
F
From
the
spec
itself-
oh
no,
okay,
so
I
I
think
eventually
we
need
to
drive
that
from
the
spec
side,
but
it's
a
chicken
egg
question,
so
I
I
will
start
by
putting
some
simple
document
and
the
example
code
and
of
course
the
example
will
not
be
perfect,
because
I
have
so
many
questions.
So
I
put
to
do
item
and
all
my
questions
in
the
example
and
then
I
can
try
to
solve
the
example
by
saying
hey.
F
We
can
do
this,
we
can
address
this
question
and-
and
we
need
to
change
this
particular
api
or
the
sdk
and
see
people
like
that,
and
then
we
have
the
discussion
and
once
we
know
where
we
want
to
be,
we
can
compare
it
with
the
spec
and
get
back
to
the
spec.
In
this
way
I
have
something
very
concrete
to
drive
the
spec
change.
F
D
Yeah
last
time
I
thought
about
it.
There
is
a
blog
post
that
I
wrote
about
use
cases,
so
I
was
trying
to
approach
it
from
a
scenario.
How
end
users
will
see
will
use
open,
telemetry
and
will
enrich
telemetry
that
is
produced
by
open
imagery?
It
doesn't
have
multiple
experts
scenario,
but
it's
it's
also
very,
very
good.
I'm
not
sure
how
much
we
need
to
optimize
for
like
for
the
scenario
when
multiple
processor
chains
go
into
a
single
exporter.
F
D
D
Crush
it,
if
we
don't
support
it
right
now,
and
it's
not
possible
to
do
it,
and
it's
should
be.
It
may
be
something
that
we
can
postpone
for
later
version
unless
we
like,
I
think
we
need
to
approach
it
from
like
scenarios
first
and
stability,
so
I
totally
agree
with
stability
aspect
of
it
like
it
shouldn't
crash.
If
you
just
combine
it
different
play
a
different
way,
it
shouldn't
be
crushing,
but
I
also
don't
want
to.
D
F
We're
trying
to
solve
yeah,
so
it
seems
like
from
the
user
perspective.
I
don't
know
how
to
use
that,
for
example,
if
you
look
at
just
that
pr
serial
will
open
up
pr.
So
you
can
see.
I
added
three
pipelines
and
my
question
is
when
we
finished
that
before
the
application
got
finished,
the
old
hell
object
will
get
this
post
automatically
right
before
we
leave
this
closure
yeah
and
and
how
many
this
post
are
we
going
to
get
for
the
processor
we're
going
to
get
two
or
three
or
one
or
zero.
D
And
the
user
scenario
would
be
that
in
this
post
we
need
to
clear
up
some
some
resources
to
open
like
files,
yeah.
F
H
F
Yeah
exactly
so,
you
got
three
this
post,
but
the
problem
is:
if
you
remove
a
and
b,
what
are
you
getting?
I
expect
you
got
one,
but
I
got
nothing
so
you
see
like,
like
I've,
been
working
on
this
product
for
for
about
like
too
long.
I
feel
I
don't
have
idea.
So
how
would
expect
the
user
to
have
idea?
And
if
we
give
them
something
like
this,
when
they,
when
they
implement
this
correctly,
they
registered
and
they
didn't
see
the
behavior.
Then
can
we
explain
that?
D
F
F
F
B
F
So
you
see
like
instead
of
finding
a
bug,
I
think
we
have
some
systematic
thing
and
I
want
to
put
those
tutorials
because
they're
going
to
be
helpful
anyways
at
the
end
of
the
game.
We
need
to
give
people
a
tutorial,
so
they
can
follow
this
and
build
their
own
components.
And
by
having
this
and
putting
all
the
questions,
I
can
start
to
use
this
to
drive.
Hey
like
this
is
not
what
we
want
and
how
we
can
change
that.
F
B
Idea
is
like
generally
look
at
all
the
extensibility
points
offered
by
all
the
building
blocks,
which
customers
can
use
open,
elementary
things
which
we
ship
from
this
repo.
B
And
see
if
it
is
see
if
it
makes
sense
from
a
customer
perspective
and
see
if
it
is
like
simple
enough
or
even
possible
for
customers
to
like
incorrectly
write
it
somehow
and
shoot
themselves
in
the
food,
so
yeah
like
just
an
initial
attempt
to
list
down
the
issue,
we're
not
solving
it
like
today,
but
at
least
this
is
like
an
entry
point
where
we
are
listing.
All
the
possible
issues
where
things
are
not
clear
like
pipeline
is
not
a
very
different
thing.
B
So
how
do
we
explain
and
the
end
goal
is
like
what
would
be
the
end
goal
like?
Will
operations
and
start.
F
The
end
goal
for
this
practice
is,
we
have
a
list
of
questions
and
then
we
can
decide.
Are
these
good
questions
we
want
to
tackle
before
ga
or
something
we
don't
think
is
super
important,
as
sergey
mentioned,
like
having
multiple
exporters?
Probably
is
the
ones
that
scenario
that
we
don't
have
to
solve.
F
D
Right
yeah,
if
it
would
be
stable
package
that
we
validated
over
years,
I
would
say
it
needs
to
go
to
website,
but
with
the
current
pace
of
changes,
I
totally
agree
that
this
kind
of
samples
will
benefit
from
being
compiled
as
part
of
the
solution.
Yeah.
F
Yeah,
so
this
is
more
like
driving,
we
know.
In
the
end,
we
got
to
have
documents
and
simple
projects,
telling
people
how
to
build
certain
components,
and
if
we
have
issues
we
got
to
be
transparent.
We
know
the
issues
today
and
then
we
can.
We
can
see
how
we
can
address
those
issues
and
or
this
can
be
an
indication
where
we
are.
B
B
Yeah
I
mean
like,
if
we
think
of
this,
as
what
are
the
things
which
customers
can
do
other
than
just
like
consuming
it,
then.
Yes,
we
need
to
cover
that
also
because,
technically
we
don't
enforce
the
usage
of
this
sdk,
they
can
create
their
own
sdk
as
the
extreme
extensibility
point:
yeah,
okay
yeah.
So
let's
do
this
I'll
go
through
this
and
maybe
like
instead
of
comments,
let's
put
to-do's
yeah,
you
already
have
to
lose
here
right,
okay
and
yeah.
B
F
B
Which
I
I
haven't
found
answer
yet
so
yeah,
it's
not
really
well
defined
in
the
spec
as
well
and
in
if
you
use
the
pure.net
activity
apa,
it's
still
confusing,
because
if
you
have
like
multiple
pipelines,
one
says
assemble
it
in
and
the
other
gauge
is
wandered
like
yes,
but
there
is
no
yeah.
That's
it.
F
Yeah-
and
I
think
it's
fine-
that
the
spec
is
not
clear
on
all
these
specs,
like
I've,
never
seen
like
even
for
the
c
plus
plus
language
spec.
It's
not
it's
not
perfect.
I
think
that's
so
qualified,
but
it's
not
okay
for
an
sdk
to
tell
people
for
a
common
scenario.
If
you
use
that
something
promised
by
the
spec
like
you
can
have
multiple
providers
and
we
don't
know
what
would
happen
so
you
got
to
try-
and
this
is
bad
because
there's
no
there's
no
safety
guarantee
for
the
customer.
F
At
least
I
want
to
tell
them
like
use
this
like
be
aware.
This
is
not
something
that
is
well
not
tested,
so
you're
on
your
own,
and
this
one
is
something
you
can
rely
on,
or
at
least
we
know
if
you
use
this
you're,
not
you're,
going
to
shoot
yourself,
that's
still
better
than
like
having
a
big
ic,
clubhouse
api
and
everything
is
like
a
jungle.
D
Yeah
one
important
aspect
of
this
exercise
or
maybe
separate
exercise
like
I,
I
really
like
the
sensibility
model
exercise,
but
also
how
do
we
make
impulse
because
how
customer
configure
and
solution?
So
sometimes
there
is
a
problem
of
like
somebody,
configuring
pipelines
and
then
after
that,
adding
new
data
collector
and
then
nothing
works.
D
So
maybe,
if
you
can
yeah
clear
up
those
scenarios
as
well,
where
we
register
everything
and
make
sure
that
we
clearly
articulate
it
in
documentation
and
maybe
have
some
proper
login
enabled
for
the
scenarios,
and
it
will
also
be
very
helpful.
B
F
B
B
So
in
in
case
of.net
we
go
for
the
propagation
one,
so
it
means
activity
is
created
with
a
propagation
and
we
still
send
that
activity
to
processors
and
exporters.
So
that's
like
one
limitation.
That's
why
you're
seeing
like
these
things
in
console,
even
even
if
we
say
don't
stumble
it
in.
D
And
yeah,
I
think
I
think
specifications
have
a
note
about
experts.
Yeah.
B
An
issue
which
we
opened,
which
like
makes
it
clear
that
our
sampling
implementation
is
not
up
to
the
spell
like
90
of
them.
We
should
be
able
to
solve
it
except
the
last
one.
There
is
one
issue
which
is:
I
mentioned
it
in
the
beginning.
This
issue
is
also
about
sampling,
like
when
sampler
returns.
Apart
from
returning
the
assembly
decision,
it
can
also
return
a
set
of
tags,
additional
attributes
which
is
supposed
to
be
attached
to
the
activity
itself,
but
currently
there
is
no
way
to
do
that.
B
So
there
are
a
bunch
of
broken
things
with
sampler
and
yeah.
One
of
the.
What
right
you
hit
here
is
one
of
them.
What
rumula
hit
here
is
completely
different
one,
and
this
one
is
recurring
support
from
dotnet.
This
one
is
just
honestly:
we
just
need
to
rewrite
the
assembler
to
take
into
account.
I
mean
to
match
the
spec.
D
Yeah
and
even
very
simple
scenarios
that
I
mentioned
in
a
blog
post
that
I
posted
people
want
to
sample
out
some
requests
to
like
house
page,
for
instance,
it's
not
possible
today
with
activity,
because
url
is
not
available
so
like
we
need
to
try
out,
try
and
implement
this
scenario.
It's
very
basic,
but
it's
still
not
trivial
to
implement
and.
G
B
Need
to
give
some
other
solutions,
those
filter
and
enrichment.
I
think
we
have
issues
open.
I
think,
like
michael
did
like
one
attempt
to
fix
it.
We
haven't
really
closed
on
it.
Yeah
I
mean
it's
there
like.
We
have
to
have
easy
way
for
customers
to
filter
out
things
also
rain
and
enriching
yeah
yeah.
H
B
G
B
Mean
we
don't
want
to
discuss
all
of
them
here,
yeah,
it's
this
one,
but
yes
thanks
riley
for
like
trying
this.
Let's
get
like
more
of
this
and
figure
out
what
are
the
other
things
which
can
be
misleading
and
then
work
our
way
into
like
fixing
it
in
the
sdk
and
eventually
getting
clarity
from
the
specs
itself.
F
Yeah,
so
if
I
could
summarize
this,
I
I
think
the
goal
of
practice
is
to
start
from
the
user's
perspective.
We
have
three
personas
the
developer,
that
simply
takes
the
open,
telemetry
api
or
the
sdk,
and
use
that
in
their
application
and
the
library
owners
who
don't
own
any
application.
They
simply
take
the
api
or
activity
api
and
instrument,
their
application,
and
also
the
the
owner
of
certain
plugins
like
different
vendors.
F
They
develop
something
for
open,
telemetry
and
I
think,
given
we
have
these
customers,
we
want
to
give
them
some
apn
components
that
are
like
set
them
for
success.
So
number
one
thing:
is
they
got
to
be
performant
reasonably
like
performant
number
two?
Is
they
got
to
be
simple
for
people
to
fall
into
the
path
of
success?
Number
three
is
the
behavior
should
be
well
defined
and
it's
not
something
like
people
try
and
come
to
us
and
we
look
at
each
other
and
say
we
don't
know.
B
All
right,
I
think
we
covered
all
the
topics,
so
my
just
want
to
remind
everyone.
There
is
an
issue
on
in.net
where
there
is
a
possibility
for
us
to
influence
the
name,
some
of
the
names
which
were
previously
like
marked
as
confusing,
so
we
have
a
I'm,
not
saying
we
have
a
good
chance
of
making
it
happen.
It's
just
a
proposal
if
it
gets
approved,
then
we'll
have
it
otherwise
we'll
have
to
live
with
it.
So
this
is
our
like,
probably
the
last
opportunity
to
influence
dot
net
five.
B
So
please
look
at
this
issue
and
share
your
feedback.
There
right.
B
One
one
last
thing
like:
if
there
are
things
if
there
are
open
issues
which
any
one
of
you
are
working
on
and
you
can
commit
to
make
it
happen
in
the
next
video,
please
ascend
my.
I
don't
think
anyone
can
yeah
just
pick
me
in
jitter
and
we
cannot
say
milestone
so
who
is
looking
at
the
report?
They
can.
B
They
get
a
clear
idea
of
what
they
can
expect
in
beta
2
and
beta
3
and
beta
4,
because
right
now
I've
been
like
very
conservative
in
adding
issues
into
a
particular
milestone
but
yeah.
If
there
are
issues
which
you
are
working
on,
I'm
confident
that
it
can
be
done
in
this
time.
Let's
like
assign
the
right
milestone,
so
people
can
know.
B
B
It's
not
confirmed,
but
they
there
will
be
a
dot-net
design
review
sometime
like
this
week
and
the
decent
review
has
to
approve
it
and
then
yeah.
So
we
don't
know
whether
donald
to
fix
that
or
not
yeah.
I
mean
it's
currently
tagged
as
5.1,
and
it
says
it
requires
api
review,
because
this
is
a
breaking
change
compared
to
the
previous
one,
which
was.
B
Yeah,
okay
thanks
and
the
recordings
are
like
posted
in
youtube.
Like
whatever
happens,
it's
completely
all
the
decisions
on
this
from
dot
net,
it's
all
public,
so
anyone
can
listen
to
it.
I
think
the
last
decision,
I
think
I
sent
the
link
to
mike
michael.
If
anyone
is
interested
in
knowing
how
dotnet
decides
to
take
it
or
reject
it,
we
have
the
video
at
the
end,
so
yeah,
okay,
I
think
I
can
see
greg
already
put
some
comments.
Yeah.
B
B
Yeah
this
one
like
the
meeting
about
like
daughter,
api
review,
happens
like
every
week,
also
like
they
review
all
the
proposals
to
apa
changes.
So
this
week
we
expect
to
get
discussion
on
like
this
particular
item,
which
please
I
ask
from
open,
telemetry.
C
C
B
Yeah,
let
me
figure
that
out,
but
either
way
the
video
is
public
that
I
know
for
sure,
because
it's
published
to
youtube-
and
I
shared
it
already
with
some
of
yeah.
So
I
can
ask
if
it
is
accessible
for
everyone
from
outside
as
well,
and
then
I
can
find
how
to
get
you
invaded
otherwise
I'll
just
share
the
video.
At
the
end
I
mean
it's
not
shared
by
me.
It's
all
already
shared
automatically
by
the
dot
net
repo
yeah,
I'm
trying
to
figure
out
like
where
it
is.
B
Yeah
thanks
for
yeah,
I
mean
it
will
be
looked
at
by
the
dog
team.
So,
okay,
all
right.
That's
it
thanks
everyone,
and
most
likely
we'll
see
you
again
tomorrow
for
the
instrumentation
thing:
okay,
bye-bye.