►
From YouTube: 2023-01-30 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
B
D
C
C
Some
some
PRS,
which
I
probably
found
relevant
for
us
to
know
or
to
discuss
the
first
one
was
the
add,
the
suppress
tracing
flag.
So
this
is
something
still
open
and
the
pr
is
still
open,
and
this
says
that
there
should
be
a
way
for
exporters
to
disable
the
tracing
based
on
the
flag
which
is
sent
which
is
set
in
the
context.
C
So
it
is
something
different
from
sampling,
something
different
from
HD
SDK
and
different
different
levels
of
probably
disabling
the
for
not
sending
the
spans
like
you
can
either
disable
the
SDK.
That's
something
end
to
end
descending.
We
can
probably
make
the
sampler,
when
you
create
a
sampler
to
to
not
send
that,
but
given
span
based
on
some
logic-
and
this
is
something
just
before
doing
the
export.
C
If
you
want
to
disable
sending
the
spam
and
I
think
it
says
that
it's
not,
it
won't
be
limited
to
spam,
but
this
would
be
something
we
should
be
able
to
suppress
sending
all
this
thing
for
all
the
signals
have
some
kind
of
that,
like
this
very
interesting
all
right.
It
says
that
JavaScript
already
supports
this.
C
And
update
Zipkin,
remote
and
point
preferences
that
something
probably
specific
to
the
semantic
context
attributes
in
Zipkin
I,
don't
think
that
should
be
directly
affecting
us,
but
I
haven't
seen
our
code
so
probably
have
would
be
good
to
really
see
this
I'll
create
a
laser
issue
for
this,
and
probably
we
can
validate.
If
this
is
something
going
to
affect
us,
it
basically
changes
the
semantic
conventions.
C
Attributes
names
to
be
in
sync,
with
the
latest
attributes
latest
semantic
convention:
okay,
I'm
going
to
see
if
it's
going
to
change
anything
else,
Jagger
remote,
sampler
yeah.
This
is
something
I
think.
Probably
we.
C
C
This
is
something
different
from
you
again.
Exporter,
probably
is
something
a
sample
which
we
need
to
add
as
part
of
our
existing
set
of
samples.
So
yeah,
you
know
probably
I,
think
I'll
look
into
this
here.
I
think
I
have
this
P.
This
issue
was
assigned
to
Tom
I
thought
this
is
something
specific
to
Jagger,
but
this
is
not
something
to
the
other
exporter,
so
probably
I'll.
C
So
the
semantic
convention-
sorry,
the
specification.
Probably
this
method
was
somehow
got
missed
out
in
the
specification
for
metric
reader,
but
I
just
variated
our
we.
D
C
Support
Port
Splash
for
automatic
periods,
so
I
don't
think
again.
We
have
to
do
anything
for
that.
So
yeah
that
that's
something
that's
not
something
I
saw
for
the
11th
specification
changes.
I
didn't
find
anything
major.
Apart
from
these
in
last
couple
of
weeks,
we
will
talk
about
but
yeah
anything
else.
Otherwise
you
can
discuss
the
PRS.
B
On
but
otherwise
it's
a
small
cleanup.
C
C
Enable
building,
yeah
I
think
I
put
some
comments
on
that
I
mean
whenever
you
have
time,
Mark
I.
Think
probably
we
could
reply
on
this.
Yes,.
C
This
is,
this
is
ongoing
right.
Yes,.
C
Working
buildings
yeah
build
open,
Telemetry,
C,
plus,
plus
SDK
and
exported
into
DL
and
yeah
I
didn't
get
time
to
fully
review
it
I
plan
to
do
it
this
week,
so
I'll
check
with
Tom
if
he
is
okay,
if.
C
B
B
My
main
comment
is
if
we
are
looking
to
to
ship
only
one
dll
for
entire
open
Telemetry
or
if
we
look
at
a
dll
for
the
SDK
and
dll
for
exporters,
and
my
understanding
is
that
the
storm
wants
to
have
one
just
one:
dll
and
I
don't
see
how
this
can
work.
C
A
C
Yeah
I
mean,
as
of
now,
I
think
this
is
only
one
dll
which
includes
the
complete
SDK
plus
otlp
exporter.
C
D
C
C
I
think
I
was
just
thinking
whether
we
the
way
we
are
doing
it
right
now
for
having
one
dll,
so
I
think
the
change
is
basically
having
one
target
which
creates
a
single
dll.
Should
we
have
as
a
convenience,
have
a
single
shared
library
for
Linux,
also
for
the
customer
to
single
Target
having
Hotel
dot
so
which
in
the
Linux
customers
can
consume
instead
of
having
multiple
targets,
if.
C
To
have
yeah
I
mean
having
a
single
shared
library
for
complete
Hotel
CPT.
B
B
See
having
one
one
dlf
of
SDK
when
it
comes
to
exporters,
I
mean
if
you
have
two
different
applications
deployed
on
the
same
machine
that
uses
different
exporters,
which
one
you
put
in
the
DNA.
C
C
C
Open
testing
shim
yeah
I'll,
be
taking
this
week
I
think
this
week
there
are
few
PR
which
I
probably
I,
want
to
spend
some
time
on.
I,
don't
think
anybody
from
open
tracing
I
thought,
probably
somebody
from
open
testing
would
be
reviewing
if
they
could
review
it,
but
I
don't
think
at
least
this
guy
is
reviewing
it.
So
probably
let
me
see
if
I
can
find
some
time
in
this
week
to
review.
B
D
C
C
C
D
C
Which,
let's
probably
let's
ping
this
guy?
This
is
a
valid
pixel,
not
even
a
forgotten.
What
was
the
change
in
this.
C
So
if
he's
not
responding,
probably
it's
good
to
close
it.
Okay,
I
remember
from
any
different
browser,
and
let
me
quickly
do
it
from
my
current.
C
D
E
C
C
Got
it
yeah
I
got
it:
it's
not
it's
only
for
civic
and
what
what
was
the
motivation
for
this
fix
for
the
issues
with
PPS
in
the
environment?
Right
if
anything,
changes.
E
One
option
is
that
we
take
the
control
of
the
environment.
We
have
the
CI
on,
but
we
don't
I
mean
we
can
skip
that
part
too.
But
the
other
motivation
is
that
it
speed
UPS
the
OTL
pgrpc,
build
by
by
above
2X,
to
see
how
I
build.
E
So
it's
builds
the
grpc
and
everything
that
is
needed
for
building
open,
Telemetry
and
tests,
and
you
just
load
the
image
which
is
I.
Think
I,
don't
remember
how
big
the
image
is.
C
C
Obviously,
we
cover
C,
plus
plus
version
we
were
using
was
the
latest
or
the
default
latest,
which
comes
with
the
with
the
virtual
machine,
but
now
we
have
kind
of
hard-coded
that
version
so
I,
don't
think
we'll
have
surprises
in
terms
of
failure
of
buildings
with
the
new,
with
any
changes
in
the
build
environment,
but
would
like
to
see.
Definitely
if
it
is
this,
this
change
is
going
to
probably
this.
If
this
change
will
at
least
reduce
the
build
time,
I
think
that
would
be
important
to
see.
C
I
just
wanted
to
I
am
not
not
related
to
this
PR.
As
of
now
I
think
the
Alpine
Docker
image
comes
with
the
grpc
Dev
environment.
It
comes
with
the
packages
for
the
grpc
dev
environment.
If
it
is
something
that
is
something
we
can
use
in
our
existing
builds.
Yes,.
C
E
And
I
think
it
did
one
more
thing,
which
is
not
a
big
issue,
but
when
there
are
changes
in
the
CI
images
on
GitHub
images
for
some
time,
the
images
are
not
the
same
for
everyone.
So
if
I
yeah,
if
I
work
on
some
PRM,
my
Fork
I
probably
get
some
different
image.
B
E
C
May
be
still
using
the
default,
C
plus
plus
version,
which
comes
with
the
compiler,
with
the
with
the
GitHub
virtual
environment
right
and
then
still.
This
will
not
solve
the
problem.
I
mean.
C
C
D
C
B
Same
I
guess
it
depends
on
what
what
quality
we
need
and
what
so
recently,
during
Thanksgiving,
everything
was
chaotic
because
of
a
of
a
change
in
the
Ubuntu
latest
image
that
created
chaos
in
NCR
and
so.
D
B
So
on
another
question
is
I
mean
I
will
get
to
live
without
once
in
a
while.
We
can
also
anticipate
that,
because
I'm
sure
it
was
announced
somewhere.
So
we
can
also
be
aware
of
that,
and
the
other
question
is
especially
for
testing
with
different
versions
of
g-test
and
grpc
and
portograph,
and
everything
else
I
think
we
do
need
to
test
with
multiple
versions
anyway,
because
we
cannot
wait,
let's
say,
for
example,
during
during
an
upgrade
of
Ubuntu
with
a
new
grpc
which
is
deployed
at
the
same
time.
B
If
we
wait
until
then
to
to
discover
it,
it
would
be
too
late
because
we
will
have
already
people
complaining
that
it
doesn't
work
with
the
latest
grpc
and
the
latest
Ubuntu.
So
we
most
likely
need
to
do
builds
with
something
fresh
which
is
more
recent
than
what
is
on
the
default
image
for
for
search
and
search.
Always
so
because
of
that
we
needed
to
have
our
own
compiler,
or
maybe
with
a
recent
one
on
our
own
grpc
on
g-test,
throwing
closely
what
is
latest.
C
D
C
E
B
E
I
mean
it's
not
a
constant
thing:
I
will
just
when
there's
a
new
grpc,
we
upgrade
to
the
new
grpc
yeah,
but
we
do
that.
I
mean
we.
We
don't
get
it
as
a
surprise
from
GitHub.
E
Very
cool:
no,
it
comes
with
the
image
so.
C
E
Yes,
that
that
is
just
remote
that
will
be
removed
from
the
CI,
the
build
part
okay.
But
if,
if
we
upgrade
this
the
grpc,
the
this,
the
image
will
be
changed
to
the
new
grpc
as
well.
E
E
I
think
it
was
reduced
from
10
20
minutes
to
10
minutes.
It
was
something
like
that.
It's
I
don't
remember
long.
It
was
long
ago
not
sure
if
I
wrote
on
the
pr
description
did
I.
D
C
I
mean
if
it
increases
definitely
increases
the
grpc,
at
least
for
the
grpc
cases.
For
but
this
this
will
only
bring
these
your
things
locable.
So.
C
E
C
D
C
C
E
E
Yeah
and
we
will
be
a
bit
faster
when
we
remove
the
grps
sorry,
the
anchor
for
Windows.
C
A
Is
the
current
PR
for
Linux
only
or
that.
A
E
So
Windows
one
we
we
could
improve
by
caching.
Also
the
VC
package
installed.
That's
the
slow
part,
I
believe.
C
D
B
I
tried
to
check
in
some
searches
that
I
generated
locally
and
then
I
got
an
email
of
insult
from
GitHub,
saying,
hey.
You
are
checking
in
Secrets
because
you
have
a
pirate
key
checking
in
you.
Well.
B
It
was
basically
a
warning
hey.
You
should
not
do
that
so.
Well,
it's
a
it's
a
test
key,
so
it
doesn't
matter,
but
instead
what
I
did
is
I
generally
I
can
now
generate
the
key
on
the
Fly
during
testing.
Okay,.
C
B
In
CI,
I
was
able
to
generate
the
keys
for
for
the
test
and
then
the
next
part
is
to
try
I,
also
a
small
test
times
and
I'm,
currently
trying
to
well
sort
of
a
logs
from
the
test
client
to
to
do
the
test
itself
so
making
some
progress
well,.
B
So
this
one
on
CI
will.
C
C
A
A
B
If
we
I
tried
that
earlier,
if
we
check
in
the
certificates,
there
is
about
complaining
that
we
should
not
check
private
keys.
B
Well,
over
I,
looked
at
other
people
are
doing
that.
The
open,
Telemetry
collector
also
has
some
test
keys,
so
you
can
check
in
in
a
repo,
but
I
tried
to
avoid
that.
A
C
You
must
what
do
you
suggest
because
I
probably
we
discussed
that
you're
going
to
split
this
PR
and
then
I
am
waiting
basically
to
get
a
split,
smaller.
B
B
Yes,
I
will
speed
the
pr
for
all
the
environment
viable
part
before
that.
I
need
to
address
a
couple
of
comments
that
I
went
raised
on
that
that
code,
so
that
I
don't
have
to
to
be
together.
C
C
Okay
and
here's
Tom
quickly
on
this
on
the
dll
pr.
A
Okay
for
this
one
I
think
I'm
I
witnessed
from
all
their
comments.
I
think
I
resolved
or
most
of
them,
maybe
one
of
them
I'm
still
looking
at
it
yeah
for
free
to
to
take
more
look.
A
If
there's
comment,
I
think
the
last
time
I
mentioned
there's
there
could
be
a
potential
leak
in
a
shared
PDR
right
and
that
one
was
and
that
one
was
fixed
in
there
in
the
latest
VR
I
added
the
default
Editor
to
the
construction
of
shared
PTR
from
real
pointer
in
in
our
no
STD
header
file
share
PTR,
so
that
should
fix.
D
A
D
A
For
catalog,
we
don't
use
that
a
lot,
so
I
mean
the
impact
should
be
minimal
or
if
you
prefer
I
think
I
can
still
like
to
separate
in
the
past.
For
for
their
build,
I
mean
if
it.
C
A
Yeah
yeah
this
file,
yeah
I,
would
like
to
avoid
like
to
check
the
macro
Dr
built
here
and
I.
Think
this
should
also
work
for
the
default
for
non-dial
case,
but
in
if,
if
you
prefer
like
to
separate
it,
that
could
pass
I
think
that
that
can
also
be
down
here.
A
C
A
B
D
B
B
D
C
Be
reviewing
it
probably
this
week,
I
mean
it's
okay.
If
it's
okay
to
to
keep
it
open,
I
mean.
A
A
C
E
Should
we'll
be
closing
this?
Yes,
yeah.
C
Be
sure,
let's
quickly
see
they
have
been
few
issues
I
think
spans
set
attribute,
is
allowing
duplicate
key
attributed
when
using
otlp
exporter,
yeah
I
think
I
saw
it
a
bit.
This
would
be
different.
It
will
not
be
easy
to
implement,
looks
like
because
we
directly
see
realize
or
if
you
directly
copy
everything
in
the
Proto
data
sections
I
mean
all
the
attributes.
C
Unless
until
this
we
have
a
method
to
get
a
list
of
all
the
attributes
as
having
any
methods
search
in
the
attributes,
I
didn't
I
didn't
see
that,
but
it
would
be
good
to
see
the
total.
You
know
files
at
least
in
attributes,
if
you
have
any
anything
to
do
a
search
or
a
given
thing
or
if
it
is
possible.
Otherwise
we
have
to
read
all
the
attributes
and
do
that's
on
that
which,
for
every
for
every
set
attribute
call,
we
have
to
do
that
which
will
not
be
visible
but
yeah.
C
B
C
C
D
E
C
D
D
C
C
C
C
C
The
in
The
Matrix
aggregation,
okay,.
C
E
C
D
E
C
C
C
C
It
should
be
the
NPC
unit,
sorry,
it
will
not
automatically
read
with
the
latest
changes.
It
will
not
automatically
automatically
use
as
erpc
I
mean
use.
The
otlp
here
see
even
if
grpc
is
installed,
erpc
Dev
is
installed.
It
will
not
automatically
enable
the
grpc
exporter
with
our
research
changes.
Okay,
so.
C
B
D
C
B
A
D
A
D
D
C
Okay,
we
probably
just
let's
probably
go
to
the
miscellaneous.
If
there's
something
important
to
discuss
before
that
release
1.8.2.
This
should
be
good
to
have
a
new
release.
Now
right,
I
mean
yeah.
A
B
D
B
C
Let
me
let
me
do
something
for
Matrix
I
think
some
of
the
changes
we
may
have
to
split
it
in
different
releases.
Some,
not
all
the
changes
can
go
together,
so
probably
have
a
status
model.
Minus
Milestone
would
help
here
so
probably
I
think
as
of
now
extend
it
so
that
it
looks
fine,
it
should
not
show.
C
Okay,
thanks
for
reminding
this
Mark
and
I
think
quickly,
there
was
something
which
probably
has
been
ongoing.
The
Outreach
this
is
this
is
this
is
basically
I
mean
in
cncf,
it's
specifically
cnca,
but
now
they
have
started
to
take
it
for
open
Telemetry.
Now
they
started
accepting
any
news
requests
which,
as
an
open,
Telemetry
anyone
in
Opportunity
Community,
they
want
to
have
any
internship
they
want,
they
want,
they
have
some
features.
They
want
any
new
interns
to
develop
into
three
month
internships.
C
So
it's
a
paid,
CNC
cncf
will
pay
those
interns,
but
if
any
of
us
have
any
such
feature
related
to
open,
Telemetry
related
to
opportunity,
C
plus
plus,
please
we
feel
it
would
be
good.
Any
paid
in
turn
can
work
on
that,
and
one
of
us
can
work
as
a
mentor.
I
think
that
could
that's
would
be
something
which
we
can
add
that
feature
as
part
of
Outreach,
so
something
something
of
probably.
C
Both
interns
also
and
as
a
community
also
we
can
get
benefited.
So
this
is
the
I
mean
I.
Remember
that
this,
because
so
this
is
operator,
this
is
open
to
to
take
any
any
such
features
which
we
want
to
add,
and
also
the
etw
exporter
was
done
through
this
Outreach
and
one
of
the
intern
has
did
the
development
for
that.
So
I
think
that
it.
D
C
A
good,
probably
so
definitely
a
good
benefit
to
us
to
have
that
something
getting
implemented.
So
yeah
is.
C
C
Us
sorry,
which.
C
No,
it
can
be
anywhere,
it
can
be
anywhere.
It
does
not
matter
which
location
so
probably
I,
think
mentors
and
in
terms
they
can
set
up
well
once
it
is
get
so
so
as
a
mentor.
If
we
have
some
feature,
we
can
probably
propose
that
feature,
and
then
it
once
it
gets
approved
the
intern.
There
will
be
already
set
of
internal
interns
who
would
have
applied
for
the
internship,
so
they
will
reach
out
they
will.
C
They
will
probably
reach
out
to
the
for
those
proposals
and
somebody
would
be
interested
would
be
assigned
to
that
or
we
we
may
have
as
a
mentor.
We
may
have
multiple
interns
and
then
we
have
to
select.
We
may
have
to
interview
them,
I
know
for
etw
exporter,
they
have
a
couple
of
couple
of
interns
and
they
were
interviewed
and
then
one
of
them
were
selected,
but
it
does
not
matter
the
location.
It
can
be
anywhere
geographically.
How.
C
C
There
definitely
they
would
be
reaching
out
directly
to
cncf.
D
C
Any
such
internship,
because
it's
a
paid
internship
that
definitely
applicants,
would
be
there
yeah,
but
eventually
for
for
that
feature,
if
there
are
interests
by
the
applicants,
I'm
not
sure
whether
that
the
interest
is
given
by
applicant,
so
directly,
cncf
will
assign
it
that
these
are.
The
set
of
this
is
something
which
you
can
directly
take,
but
in
the
in
the
end,
CNF
cncf
will
tell
that
these
These
are
the
set
of
interns.
B
For
the
let's
say,
there
is
a
feature
to
implement.
Is
it
the
intern
that
come
up
with
a
feature,
or
is
it
us
to
come
up
with
a
feature
and
an
intern
will
try
to
work
on
it.
C
So
so
we
have
to
propose
the
feature
degree
as
in
any
of
the
open,
Telemetry
Community
member
can
propose
the
picture
and
it
has
to
be
accepted
by
cncf
as
the
valid
as
a
feature
which
will
be
definitely
benefiting
open,
Telemetry.
Once
that
is
accepted,
then
the
interns
would
be
assigned.
Probably
multiple
interns
would
be
assigned
for
that
feature,
and
then
we
have
to
select
one
of
them.
Okay,
so
V
has
been
put,
server
has
proposed.
That
feature
has
to
select
so
the
right
candidate.
For
that.
B
And
then,
and
the
inventor
is
mentorship
after
that,.
C
So
probably
I
think
you
can
reach
out
to
you
can
go
through
this
page
already
mentioned
in
that
link.
You'll
get
more
information
on
this
on
FAQs,
and
but
it's
not
something
specific
to
seek
if
it
can
be
something
different
from
I
mean
like
Mark.
If
you
want
to
propose
some
some
feature
which
is
specific
to
Java
or
something
you
can
do
that,
it's
not
something
specifically
because
anybody
can
do
that
as
a
community
of
open
television,
Community
member,
oh
I,.
C
C
A
C
C
C
B
Yeah
well,
I
tried
to
to
restart
the
build
already
or
which
one
which
one
is
it.
I
saw
some
failures
when
basically
GitHub
fails
to
download
something
and
complains
about
yeah.
A
C
A
C
Okay,
I
think
that's,
that's
all
I
have
I
forgot
to
say
there
is
there.
Is
you
can't
open
Telemetry?
Also,
it's
part
of
new
phone
conference,
some
sometime
in
in
February
I
have
to
check
what's
the
dates
and
it's
in
Europe,
so
I
mean
any
any
of
you
who
are
interested
or
any.
Anybody
is
interested
to
give
any
talk
or
in
your
talk,
2023.
C
C
You
can
close
the
meeting,
probably
you
can
meet
next
week.
Thank
you
for
joining.