►
From YouTube: 2021-09-13 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
C
A
B
A
B
Yeah,
oh
hi
laden
yeah,
so
I
I
I
was
at
microsoft,
working
on
on
tracing
since
2016.,
we've
added
some
like,
let's
say,
open
tracing.
There
was
no
open
telemetry
at
this
point,
into.net
some
modified
version
of
it,
and
ever
since
I've
been
working
on
the
on
azure,
monitor
instrumenting,
some
of
the
azure
libraries
and
I've
instrumented
httpclient.net.
B
Then
I
move
to
something
else,
and
then
I'm
currently
back.
I
work
on
azure
as
the
case,
so
my
main
area
is
the
instrumentation
of
those
libraries
and
now
is
up
in
telemetry.
They
are
actually
instrumented
as
open.
Telemetry.
We've
done
it
two
years
back
when
there
was
no
open
telemetry
released.
Yet
we
started
with
open
senses
and
then
it's
all
exploded
yeah.
C
You
have
a
question-
maybe
you
guys
already
talked
about
this,
but
what's
the
what's
the
difference
between
this
meeting
and
the
other
more
specific
semantic
convention
meetings
or
the
instrumentation
meetings.
B
Yeah,
I
think
this
is
like,
like
that.
We
have
three
currents
there
right
there.
One
on
thursdays
focused
on
messaging.
C
B
One
on
tuesday
is
about
friendly,
it
seems
we
believe
we
have
enough
semantic
conventions
and
instrumentation
issues
to
have
three
meetings,
and
we
will
also
have
http
questions
discussed
all
over
the
place.
So
it's
it's.
I
think
it's
the
first
time
we're
having
this
instance
or
maybe
the
second
time.
C
A
C
Do
you
have
anything
to
discuss,
or
should
we
just
call
it
for
this
week.
A
A
doing
new
to
like
I've,
been
working
on
open
telemetry
for
a
year
and
a
half
on
a
small
step
in
israel
called
the
spectre
and
I
implemented
many
instrumentations
in
a
node.
So
I
I
learned
all
the
tough
issues
and
the
questions
and
the
problems
of
processing
spends.
So
I'm
really
interested
in
improving
it.
And
then
I
saw
you
on
the
messaging.
B
Yeah,
I'm
kind
of
always
curious.
This
instrumentation
is
such
a
controversial
and
ambiguous
topic
that
it's
it
you
never
know
what's
best
and
and
a
couple
of
years
ago
I
felt
very
lonely
in
this
area,
and
I
am
sure
there
are
a
lot
of
people.
I
I
didn't
know
how
to
find
them
so
actually
happy
to
hear
any
like
bigger
picture.
Learn
something
from
people
like
you.
A
Yeah,
I'm
very
happy
myself
because
sometimes
reading
those
documentations
can
be
so
confusing
and
so
frustrating,
because
the
real
issues
that
you
face
when
you
try
to
implement
with
the
patching
really
difficult
to
enter
yeah.
So
what
are
you?
What
are
you
working
on
now?
Are
you
still
instrumenting
stuff
for?
B
We
are
well
on
azure
as
the
case
we
kind
of
we're
kind
of
stuck
with
instrumentation.
We've
done
some
work
and
it
needs
a
bunch
of
improvements.
So
we
we
implemented
conventions
our
owns,
because
there
were
no
conventions
at
that
point
and
we
are
currently
like
working
looking
for
the
way
to
like
move
conventions
forward.
So
we
can
onboard
onto
the
open,
telemetry
thing
and
also
like
we
don't
know
what
to
do
with
them
like
we
don't
have
a
lot
of
adoption,
because
this
is
all
preview
experimental
stuff
like
that.
C
B
Right
yeah
yeah
on
our
instrumentation,
because
conventions
are
experimental
like
we
don't
get
too
much
users
using
open
telemetry,
yet
yeah.
A
I
personally
think
it
will
gain
speed
very
fast.
A
A
A
So,
can
I
ask
you
about
the
messaging
the
messaging
specification
like
they
talk
a
lot
about
problems,
but
nobody
mentions
any
solutions
like
how
do
you
picture
the
the
the
semantic
conventions
for
messaging?
What
do
you
think
should
be
changed
from
how
it's
currently
working?
B
B
Yeah
sure
so,
yeah
no
worries
at
all.
So
like
the
the
the
message,
creation
and
publishing
they
should
be
separate.
Let's
say
you
create
a
bunch
of
messages
and
you
send
them
in
background.
They
are
totally
independent.
The
same
link
is
independent,
so
the
create
the
context
creation
should
be
baked
into
the
basically
the
constructions.
It's
a
local
operation,
and
currently
we
are
representing
it
as
a
span.
A
Because
when
we
link
from
the
processing,
currently
we
link
to
the
to
the
published
span,
but
if
it's
a
batch,
there's
multiple
processing
or
linking
to
the
same
span
and
this
pen,
it
cannot
hold
like
message
id
or
a
conversation
of
the
id
or
anything
message
specific,
because
it's
not
correlated
to
a
method.
It's
correlated
to
a
publish
which
is,
I
totally
agree
with
you
about
it.
B
Cool
yeah,
nice
and
then
I
think,
what's
going
to
change,
is
maybe
some
naming,
but
I
don't
think
we
we
can
or
maybe
we
will
just
create
a
bit
more
clarity
everywhere.
I
think
the
current
specification
is
very
bad.
A
I
personally
think
that
most
of
them,
like
messaging,
is
mainly
either
a
key
or
a
pub
sub,
or
more
often,
it's
pops
up
with
cues
attached
to
it
and
there's
no
way
to
express
it
in
the
current
conventions,
because
you
have
to
like
set
the
destination
kind,
which
is
which
is
either
a
topic
or
a
cue,
and
then
many
systems
don't
fit
in
this
this
screamer.
So
I
I
think
I
encounter
it
a
lot.
A
Yeah,
like
once
you're,
actually
writing
this.
This
implementation,
you
encounter
all
the
problems
and
the
issues
yeah.
Most
of
the
do.
You
use
those
links
like
when
you
display
and
process
the
spans.
A
Yeah,
I
agree
with
you:
it's
it's
not
easy,
it's
I
I
I
own
the
aws
sdk
instrumentation.
I
can
maintain
it
and
every
week
someone
says
it's
not
working,
there's
no
context
propagation,
like
they're
not
linked
to
each
other,
and
I
have
to
tell
him
this
is
the
specification.
This
is
how
it
works,
and
then
they
are
okay.
B
A
Yeah
so
we
started
it
like
a
year
and
a
half
ago
because
our
clients
needed
it
like
they
wanted
the
tracing.
D
A
A
Yeah,
it
was
just
missing
and
it's
very
problematic
to
tell
someone
to
have
tracing
without
his
cloud
provider
sdk
being
instrumented,
but
I
haven't
seen
any
actual
instrumentation
for
node.
B
B
We
are
thinking
that
maybe
we
can
move
it
but
like
we
want
to
keep
some
control
over
it
and
regardless
varies,
but
where
you
keep
it,
it's
gonna
be
very
painful
to
keep
dependencies
up
to
date
and
stuff
like
that.
A
A
Yeah
cool
so
very
interesting
to
talk
nice
meeting.
You.
B
Yeah
there
is
one
tomorrow,
but
if
ted
is
not
there,
I
don't
know.
Hopefully
we'll
have
more
folks
to
talk
about.
B
A
Yeah
yeah
I'll
be
very
glad
to
discuss
the
issues.
If
you
want
to
discuss
something
like
the
meeting
on
thursdays,
they
talk
a
lot
but
they're
not
touching
the
the
solutions
yet
and
I'm
really
interested
in
the
solutions
not
on
the
like.
Also
in
the
problems,
but.