►
From YouTube: 2022-11-30 meeting
Description
Open Telemetry Meeting 1's Personal Meeting Room
B
C
Okay,
I
haven't
gone
through,
the
specs
updates,
I
mean
still
still
not
probably
back
fully
with
from
the
long
weekend
which
we
had
so
I
just
started.
Looking
into
a
few
of
the
specs
discussions
and
the
PRS
which
have
been
raised
but
yeah,
probably
I,
think
I'll,
let
me
go
through
once
again.
C
D
C
But
yeah
apart
from
that,
should
we
directly
go
with
the
issues
and
the
pr,
so
we
can
probably
discuss
with
some
of
the
Mysteries
which
have
been
added
to
start
with,
and
then
we
can
go
to
the
issues
and
appearance
that
will
make
sense.
C
C
A
C
D
A
Did
not
look
exactly
but
I
think
it
tries
a
new
warning
on
some
deprecated
construct
by
default
so
that,
even
if
you
don't
ask
for
a
warning,
you
get
one
and
that
cause
trouble.
So
and
on
top
of
that,
it's
it's
a
warning
not
only
on
code
that
we
own,
but
it's
there
was.
C
B
Fixed
those
warnings
or
not
I
have
raised
the
pr
already
for
cmake
I
I
think
it
makes
sense
to
move
everything
to
in
into
our
own
docker.
B
A
So
the
fact
that
Ubuntu
latest
changes
that
comes
in
a
time
that
we
don't
control
so
yes,
it
creates
some
chaos.
So
if
we
move
to
a
Docker,
we
will
have
full
control
and
then
we
will
not
be
surprised.
However,
when
some
someone
some
developer,
takes
open,
Telemetry
and
try
it
on
his
own
back
into
our
Windows
box
or
Ubuntu
box.
It
should
work
so
we
should
not.
If
we,
if
we
do
everything
in
Docker,
we
take
the
risk
of
it.
We
don't
even
detect,
but
something
is
broken
on
a
given
platform.
A
D
A
Will
have
a
surprise
when
someone
find
a
big
report
if
they
find
one.
D
B
We
we
could,
we
could
control
that
I
mean
we
know
when
Ubuntu
releases
a
new
version
and
we
just
change
our
Docker
and
see
what
happens.
We.
C
Can
we
can
add
a
new
document
properties
you
can
test
and
do
we
know
I
mean
I
haven't
seen
that,
but
is
there
any
any
difference?
In
the
I
mean
the
total
total
duration
of
the
test
or
something
I
mean
with
the
talker,
because
it's
adding
a
new
virtualization
inside
another
visualization
problem
does
it
adds
to
any
latency
in
the
number
of
in
the
test
time
or
something.
C
A
So
one
thing
I
would
like
to
to
propose:
we
can
definitely
have
new
Docker
builds,
especially
for
things
like
if
we
take
the
GCC
for
it
period,
for
example,
that
build
we
installed
so
many
things
on
top
of
a
base
platform,
but
it
does
not
even
look
like
the
base
platform,
so
that
one
makes
sense
to
be
in
docker
if
we,
if
we
could
do
that
and
still
keep
one
one
Linux
one
Mac
OS
and
one
windows
at
least
we
won't
be
surprised
by
by
changes
also
and
another
thing.
A
C
D
C
D
B
Basal,
it
will
affect
the
caching,
it
probably
I
think
it
could
be
better
with
with
my
changes
but
I
I
haven't
touched
the
cash
yet.
B
Yeah
sure
it
will
so
I
will
zip
the
file
compress
the
everything
that
that
comes
from
bazel,
and
we
can.
We
will
cache
that
one
that
tar
file.
C
D
C
C
I
mean
and
I
mean
I
I,
don't
have
any
strong
opinion
in
either
of
keeping
either
of
them.
This
was
just
the
person
I
first
I
mean
the
first
time
we
have
been.
We
have
seen
that
kind
of
this
breaking
of
the
world
environment.
Should
we
move
it
so
or
should
we
wait
for
some
more
time
to
see
if
it's
not
happening
too
frequent,
and
then
probably
we
should
move
I,
don't
want
any
other
issues
coming
with
moving
to
a
Docker,
and
then
we
have
to
fix
up
some
other
topics.
C
A
So
this
change
with
GitHub
exposed
a
lot
of
breaks,
but
most
of
them
we
could
have
found
them
earlier
if
we
were
using
more
up-to-date
dependencies.
So
yeah.
C
D
C
Yeah
I
think
that
that's
probably
I
think
we
are
not
good
at
to
keep
at
least
all
the
different
message
to
the
latest
I
think
we
probably
when
something
break,
then
we
realize
it
and
yeah
I
think
if
we
have
more
productive
and
totally
again,
then
probably
I
think
we
may
not
have
in
this
kind
of
issues
and
even
if
we
get
this
issue,
probably
that
will
definitely
surface
out
a
real
problem
and
good
to
have
something
break.
C
If
it's
a
real
issue,
I
mean
I'm,
totally
totally
up
for
moving
to
Docker
and
that
looks
more
safe
and
controlled
environment
I'm.
Just
that,
probably
we
should
to
not
do
it
as
of
no
I
mean
that's
my
personal
opinion
and
again
not
do
it
as
of
now,
just
just
with
one
CI
environment
breaking
for
the
first
time.
As
of
now.
Probably,
let's
see
if
we
can
move
things,
keep
things
latest
and
if
it's
again
happening,
then
probably
you
can
do
a
more
controlled
environment
in
the
docker.
C
Anything
wrong
otherwise
in
the
topper
I
think
that's
pretty
right
way
of
doing
it
to
have
a
more
controlled
environment,
but
if
it's
I
mean
if
something
something
else
breaks
because
of
moving
to
Docker
or
something
else
goes
wrong,
I
mean
I,
don't
know
the
speculation.
Anticipation
is
something
goes
wrong.
Then,
probably
this
this
current
setup
is
working
for
past
couple
of
years
or
something
I
mean
we
haven't
seen
anything
breaking
as
of
till
now.
C
C
C
C
Docker
containers
and
probably
have
this
PR
for
some
time.
Let's
see
if
things
are
with
the
current
setup,
with
moving
to
the
latest
versions,
if
we
can
still
keep
the
things
up
to
date,
then
we
don't
need
to
probably
use
Docker
at
all.
D
D
A
By
the
way,
do
you
know
why
we
support
that
who
is
using
it.
C
Why
do
we
need
photos?
Yes,
I
mean
4.8
was
the
first
GCC
I
mean
4.8
was
the
first
I
think,
probably
Ubuntu
environment,
which
was
providing
GCC
relevant
support
of
four
or
three
days
of
us.
Sorry,
GCC
version
with
the
GCC
11
support,
so
I.
Think
probably
we
just
wanted
to
ensure
that
it
works
on
that.
Okay,.
C
That
support
was
not
fully
it
didn't
support
it
fully.
So
definitely
that
that
person
shows
by
we
have
special
case
for
phone
on
it.
We
can
even
remove
it.
I,
don't
I,
don't
think
any
I
mean
that's
a
compiler
which
is
not
supporting
C
plus
11
to
prevent
all
all
the
features
of
C
plus
11.
So
I
won't
say
it's
fully
compliant
C
plus
11
compiler.
C
D
C
Yeah,
okay,
okay,
let's
keep
it
I
mean,
but
do
you
see
we
should
move
back
to
Docker
I
mean
I,
just
just
thinking
why
you
should
be
we
you're
saying,
because
if
we
are
not
using
a
default
Ubuntu
package
for
that,
we
are
probably
installing
photo
explicitly
and
then
we
are
testing
is.
C
B
A
E
A
C
D
B
One
more
thing:
maybe
we
could
ask
for
some
registry
access,
so
we
could
have
this
image
pushed
somewhere
and
that
would
make
the
appears
much
faster
because
we
are
spending
some
time
just
for
building.
For
example,
grpc
for
this
otfci.
C
C
C
This
is,
we
do
have
some
I
think.
Let
me
see
as
part
of
the
CNC
for
I
mean
open
Telemetry.
We
should
have
got
somebody's
Docker
registry
which
we
can
use.
Let
me
see,
probably
if
it's
part
of
I
mean
it's
coming
from
Azure
or
it's
from
so,
but
I
find
remember,
going
through
Reading
it
somewhere
in
in
somewhere
in
open
Telemetry
Forum
for
having
me
talk
a
registry
which
typically,
we
can
use
it,
but
yeah.
Let
me
figure
out
if
it's
there.
A
C
C
To
probably
finalize
on
this
we're
seeing
that
at
least
at
least
we
can,
we
can
have
it
occur
image
for
4.8.
As
of
now
and
let's.
C
Oversee
like,
as
of
now,
let's
try
to
see
how
the
the
GitHub.
C
C
A
D
C
D
D
C
D
C
B
D
E
Think
I
I
saw
that
the
future
Google
tests
will
not
support
C,
plus
plus
even
11
I.
Think
that
includes
the.
This
is
a
4.8
right
and
I
think
we
previously
before
also
like
some
in
our
library
like
appsio,
which
has
a
plan
and
measurer,
and
it
is
already
done
or
it
is
in
apply.
It
only
support
you
just
let's
see,
plus
plus
14.,
so
drop
the
support
for
C
plus
plus
11
I
think.
Maybe
we
can
check
with
oven
to
see
whether
he
has
applying
for
for
upgrade
or.
D
C
Yeah,
but
all
these
again,
all
these
libraries,
though
we
are
not
sure,
are
not
going
to
use
Ctrl
space
11
at
the
interface
layer.
Probably
they
may
be
using
it
internally
and
they
may
not
be
I,
mean
still
with
well
I'm,
just
thinking
still
building
with
Q,
plus
plus
14
and
probably
use
it
with.
It
should
be
still
usable
right
for
the.
A
In
one
of
the
build
breaks
I
saw
last
last
week,
I
think
there
was
a
bill
break
on
the
GCC
408
build
using
some
latest
adult
remember
if
it
was
utter
or
something
else,
and
it
was
breaking
in
some
type
trades
somewhere
using
Thai
internally
by
some
adders
in
some
Library.
And
so
even
if,
in
our
code
we
don't
use
fancy
construct,
the
library
itself
might
have
a
lot
of
yeah.
A
One
thing
so
I
got
distracted
by
the
Builder
break,
but
one
thing
also
I
still
need
to
finish
is
to
split
the
weave.
Stl
support
by
already
is
of
C,
plus
plus
11
14,
17
and
20.,
because
having
things
like
more
fine-grained
will
help
to
to
upgrade
from
C,
plus
plus
11
to
C,
plus
plus
14,
later
I.
Think.
C
C
A
Many
questions,
it's
not
so
much
of
a
combination,
but
if
a
library
is
not
maintained
and
if
there
is
a
sales
bug
in
it,
then
we
will
never
see
a
fix
for
it.
So.
C
I
mean
I
mean
supporting
the
GCC
version,
which
only
supports
implants.
They
don't
have
that
then
probably
you
can
move
to
C,
plus
plus
bottom
or
so.
C
D
A
Yeah
well,
if
we
just
proceed
to
announce
that
we
plan
to
remove
C
plus
plus
11
and
see
if
there
is
some
feedback
and
everyone
screaming
loudly,
but
but
yes
doing.
An
announcement
already
is
a
good
idea.
B
D
E
D
E
D
C
Let's
see
the
traction
coming
on
that,
if
lots
of
people
start
saying
no,
we
don't
want
to
sub.
You
know
we
still
are
using
that
in
C,
plus
plus
level
environment,
then
probably
we
will
discuss
it
again,
whether
it's
the
right
move
or
not.
If
we
don't
see
much
of
the
traction,
people
are
like
not
much
comments
are
coming
and
it
seems
that
people
are
okay
with
the
move,
probably
in
the
next
release,
we
can
give
a
warning
that
we
are
going
to
move
away.
D
A
We
agree:
yeah,
yeah,
I
think
we
we
all
agree.
So
the
next
item
is
very
quickly
I'm,
starting
to
look
at
how
to
implement
SSL
support
for
the
HTTP
oltp
exporter
and.
D
A
That,
basically
I
need
to
have
some
options
which
are
only
available
in
some
more
recent
curl
and
we
trace
the
question:
do
we
do
we
force
people
to
agree
to
Carl
and
probably
not
because
that
will
break
some
more
builds
already
and
if
not,
how
do
we
do
that?
So
the
O'neill
thing,
but
the
only
option
I
can
think
of
so
far,
is
to
keep
every
feature
behind
the
feature
flag
like
or
Wednesday,
for
the
asynchronous
change,
for
example.
A
But
then
we
have
a
risk
that
we
have
so
many
internal
terms,
so
many
different
features
with
so
many
different
flags
that
it
will
add
something.
Some
complexity
as
well.
A
C
C
A
So
I
haven't
checked
with
most
likely
yes,
so
this
is
only
when
you,
when
you
provide
the
certificate
as
a
string
directly
and.
A
C
C
C
C
D
A
This
is
a
typical
typical
example,
so
this
is
a
new
feature
and
if
someone
is
using
that
on
an
old
platform,
for
example,
VCC
408,
it
won't
the
code,
Library
will
not
be
available
for
that.
So.
D
C
A
Another
question
becomes
whether
this
thing
is
a
warning
and
we
continue
or
if
we
make
it
a
failure.
So
let's
say
so.
C
We
should
make
it
sorry,
yeah,
sorry.
A
C
I
mean
I'm
thinking
like
this
way,
like
I
mean
just
seeing.
If
practically
it
is
possible.
If,
if
somebody
is
giving
the
the
the
the
certificates,
both
the
CA
and
the
client
certificates
in
the
in
in
string
and
the
version
is
less
than
7.71,
we
can
at
least
ensure
that
the
handshake
will
fail
right,
that
that
can
be
done.
That
I
think
I.
A
Don't
know
because.
A
C
C
A
B
C
C
D
C
C
And
and
the
user
is
giving
the
certificate
path,
then
then
we
don't
need
a
latest
version
right.
D
C
A
It
will
be,
it
will
be
there,
because
it's
it's
not
a
new
feature
in
card
I
mean.
D
C
D
C
We
do
have
GitHub
issues
because
curl
is
not
the
right
choice
for
all
the
platforms.
This
is
not
not
really
directly
related
to
what
you
have
been
working
on,
but
probably
just
no
I
forgot.
It
could
have
gone
down
somewhere
down
in
the.
C
C
C
Yes,
so,
basically
that
that's
how
our
current
implementation,
at
least
the
way
way
we
started
I
have
to
still
see.
If
we
still
are.
We
use
the
factory
pattern
to
support
the
HTTP
connection,
where
we
can
inject
any
of
the
client
libraries,
so
I
think
just
like
color.
It
should
be
possible
to
inject
any
of
these
libraries
from
both
Apple
and
Windows
okay,
so
so
that
that's
how
right
now
the
implementation
is
kind
of
factory
pattern
which
should
be
able
to
inject
or
any
of
these
libraries
yeah.
C
C
Yeah
I'll
be
away
at
least
I
mean
that
that's
the
tentative
right
now
probably
I
think
is
probably
the
third
week
of
December
from
17
till
25th
or
something
I'll
be
away.
D
A
C
C
C
E
C
Right,
yeah,
yeah
and
sorry
about
that.
I
think
that
was
something
because
of
me.
I
didn't
really
adjust.
I
didn't
realize
that
we
have
two
different
Scenic
targets,
I
mean
so
so
there
are
the
cmic
targets
which
are
generated
by
default,
and
then
we
also
have
the
different
CM
Target
names,
which
goes
to
the
click
config
file.
Well,
I
think
that
that
somehow
break
break
the
things.
I'm,
really
sorry
about
that,
but
yeah,
but.
E
E
Yeah
one
more
thing
is
about
this
package
package:
yeah
I
was
trying
to
like
make
metrics
in
the
how
the
new
implementation
that
default
he
invested
package
for
1.7
the
same
synthesis
and
build
issue
I.
Think
in
you
know,
1.7
code
base
learn
some
folder
in
the
in
the
oh
The
Matrix
folder,
which
header
header
folder
like
something
called
aggregator.
But
the
folder
is
important:
there's
no
no
file
on
the
data
folder,
so
they
say
package
compliance
so
actually,
without
making
Source
fixed,
we
have
to
count
release
a
with
the.
C
D
D
A
Some
someone
just
fired
a
PR
fixing
a
bug
that
I
caused
actually
in
the
make
file
for
open
Telemetry
portal
and
also
recently
with
otlp
grpc
utils
file.
A
This
was
added
into
the
common
library
for
the
grpc
exporter
and
not
into
the
TLP
grpc
export.
Sorry,
so
it
was
added
to
the
otlp
common
library
and
not
to
the
hotel,
pgrpc
Library,
so
very
I
have
notified
the
issue
yet,
but
there's
a
build
break
today.
If
you
try
to
compile
only
with
the
HTTP
exporter
and
not
without
the
grpc
and
the
break
is
that
the.
D
A
C
A
We
can
definitely
ship
as
part
of
the
1819
East
Van.
If
we
do
one.
D
C
C
C
B
D
A
Internally,
because
I
I'm
I
started
to
build
on
when
it's
at
zero
and
I
just
found
that
so
I
will
find
the
issue
with
the
details.
C
D
C
E
C
B
B
A
B
B
In
the
kubernetes
Ingress
packets
repository
I
will
check
that.
E
E
B
C
D
C
A
We
do
just
do
181
release
to
four
small
breaks.
I
will
then
wait
on
your
removal
for
the
experimental
semantic
conventions,
because
it
should
be
removed
later.
Okay,.
C
Okay,
let's
quickly,
probably
in
the
interest
of
time,
I
think
we'll
just
quickly
go
to
the
existing
issues
open
daily
protocol,
yeah
I
think
that's
already.
The
pr
is
raised.
C
A
A
So
I
think
the
use
case
makes
sense.
However,.
E
A
Are
many
things
someone
may
want
to
set
for
Father,
Fred
and
I'm
thinking
of
some
federal
core
storage
keys
that
I
might
be
using
myself,
I'm
thinking
of
named
networks?
Things
like
that.
So
if
we
have
a
solution
where
the
user
has
some
sort
of
callback
where
we
can
execute
any
code,
they
like
on
thread.
D
A
D
C
I
was
thinking
the
other
thing
probably
I
mean
if
we
can
start
supporting
users
to
come
up
with
their
own
thread.
Library
I
mean
they
may
be
using
some
some
some
other
site,
Library
internally
or
some
thread
Factory,
which
uses
which
is
creating
their
own
threads.
Probably
if
you
can
allow
them
to
add
that,
as
as
the
as
the
optional
parameter
and
then
let
let
them
figure
out
how
how
they
are
creating
those
threads,
they
just
passes
the
thread
and
we'll
use
that
20
to
create.
C
We
will
you
just
use
those
threats,
we'll
run
our
code
in
that
thread.
So
something
like
this.
If
we
can
do
it
well,
this.
C
Case
so
we
we
want,
we
won't
be
bothered
what
kind
of
threat
Library
they
are
using.
What's
Affinity,
whether
it's
speech
rate
or
anything
else,
they
just
give
us
a
handle
for
that,
we'll
use
that
handle
to
run
our
code
and
let
them
let
them
worry
about
what
kind
of
threat
option
they
want
to
use
it
in
that.
C
That
will
also
help
us
to
support
the
scenarios
where
customers
may
be
using
their
own
threat,
Factory
internally
for
their
own
application,
and
they
don't
want
a
library
like
open
Telemetry,
to
expire,
new
threats
from
within
ourselves.
Let
them
control
how
which
or
how
to
span
the
threads.
Some
people,
probably
more
generalized
here,
yeah.
C
Exactly
yeah,
so
if
they
have
some
kind
of
thread
pool
and
they
can
pass,
we
can
provide
some
custom
interface,
how
they
can
pass
it
and
then,
let's
probably,
they
can
use
that
custom
interface
to
pass
us
the
handle
for
that
and
we'll
use
that
to
create
a
thread
from
the
thread,
pool
and
use
that
to
run
our
power
code
base
game.
So
that's
that's
something
I
think
probably
more
genetic
solution,
at
least
for
this
case.
C
C
C
More
details,
but
my
thoughts
for
in
this
probably
I
just
wrote
some
three
four
liners,
but
let
me
add
some
more
details
in
this.
B
C
C
D
C
Oh
yeah
I
think,
let's,
let's
probably
in
interest
of
time,
I,
think
any.