►
From YouTube: 2023-03-09 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
A
A
C
A
B
A
B
B
B
A
really
easy
little
guy
I'm
still
excited
to
see
you
even
after
you
poke
him.
A
A
Topic
for
people
who
made
it
I
was
going
to
ask
about
putting
the
zoom
chat.
If
anybody
either
of
you
have
any
thoughts,
I
guess
it
was
totally
poor
after
so,
mothers
might
still
show
up,
but
this
is
pretty
much
the
main
thing.
It's
a
discussion
of
the
package
naming
and
whether
we'll
break
them
up
and
have
tracing
metrics
logging
events
all
separate
or
not,
and
what
people
think
would
be
preferred.
I,
don't
know
I'm
kind
of
oh
another,.
C
Face
yes,
I
should
I
should
introduce
myself
I'm,
probably
the
odd
person
I
I
slack
to
you
and
oh
yeah,
individually,
yeah
I'm,
just
I
have
recently
I'm
an
Elixir
person
but
I'm
in
a
situation
where
I'm
not
writing
like
production
Elixir.
But
my
I
work
for
nginx
and
and
like
we're,
really
interested
in
pushing
open
Telemetry
in.
C
As
an
Elixir
person,
I
want
to
use
our
platform
that
nginx
to
maybe
put
Elixir
a
little
more
front
and
center,
sometimes
and
but
I
just
don't
know
anything
about
like
implementing
open
Telemetry
in
elixir
in
production.
So
I
thought
it
would
be
fun
to
I.
Don't
may
not
have
a
lot
to
contribute
because
I
don't
have
a
lot
I'm,
not
in
the
middle
of
anything,
but
I
just
want
to
be
a
fly
on
the
wall
and
hear
you
all
talk
about
things
and
understand.
What's
going
on.
A
Oh,
that's
awesome.
That's
cool
to
hear
Too
the
nginx
is
involved.
Are
you
then,
working
on
open
Telemetry
for
nginx
I'm.
C
Not
on
the
like
the
I'm,
not
on
like
the
nginx
core
team,
they're
they're,
pretty
hardcore
I'm
kind
of
on
the
periphery
I'm
part
of
the
like
I'm,
like
the
technical
end
of
our
like
Community
team,
so
like
I,
do
technical
I'm
like
a
I,
just
stopped
being
a
developer,
so
I've
been
a
developer
for
like
10
years
and
now
I'm
sort
of
supposed
to
be
like
the
voice
of
the
developer
for
like
nginx,
so
that
that'll
be
like,
hopefully
producing
like
best
examples
of
doing
X.
C
You
know
with
nginx
and
like
just
kind
of
trying
to
make
things
a
little
easier,
because
it's
not
the
easiest
thing
to
get
into
if
you're
not
steeped
in
it.
So
that's
the
kind
of
the
plan
I
did
I
did
have
a
question
about
like
what
you're
about
to
talk
about.
If
you're
waiting
for
someone
for
more
folks
to
come
on,
I
recently
did
a
example
with
the
JavaScript
and
unlike
collection
of
libraries,
because
that's
like
the
most
palatable
to
kind
of
everyone
understands
how
to
at
least
read
JavaScript.
C
So
I
did
that
and
this
topic
that
you're
I
asked.
You
actually
asked
the
question:
haven't
gotten
the
answer
yet
on
the
discussions
for
the
open,
Telemetry
JS
Contra,
which
is
just
like
kind
of
the
reasoning
for
having
so
many
packages,
because
one
of
the
issues
I
had
when
I
was
trying
to
teach
people
how
to
make
decisions
about
instrumenting.
Their
code
in
JavaScript
is
like
how
the
heck
do
I
know,
I
need
the
you
know,
standard
standard
definitions
and
what
was
it
the
resource?
You
need
the
resource
pack.
C
Yeah
on
the
JavaScript
is
pretty
extreme,
so
my
assumptions
were
it's
for
ease
of
like
keeping
things
up
to
date
because,
like
the
spec
is
evolving,
and
you
know
some
of
the
semantic
conventions
are
changing,
so
maybe
it
makes
sense
to
have
it
in
a
separate
library
that
you
release
separately
and
maybe
performance
stuff,
like
you
really
want
to
avoid
not
including
stuff
that
you're
not
going
to
use
but
I.
Those
are
just
guesses
on
my
part,
as
I
was
curious.
C
A
A
couple
yeah
I
did
notice
because
for
work,
I
work
on
open
Telemetry
in
general,
as
well
as
early
and
elixir,
so
I've
had
the
devil
and
the
JS
side
slightly,
and
they
do
seem
to
have
way
more
packages
than
others
and
but
you
bring
up
resource
and
that's
a
good
good
one
to
mention,
because
that
does
throw
a
bit
of
a
wrench
in
my
packages,
because
where
would
resource
go
it's
used
by
all
of
them,
so
there
would
need
to
be
like
a
common
and
now
I.
A
Remember
that
some
languages
do
who
split?
Have
a
common
I
think
they
name
it
common
like
open,
Telemetry,
common
and
all
of
them
depend
on
that
and
that
would
I
could
put
resource
in
there.
But
the
reasoning
at
least
in
erling
Elixir
would
be
a
mixture
of
it's.
A
Everything
starts
when
the
the
program
starts
up,
so
like
processes
and
Ed's
tables
for
tracing
metrics
and
all
that
stuff,
and
so,
if
you're,
not
using
the
metrics,
you
probably
you
don't
want
those
and
if
there's
any
bugs
in
them,
it
could
hurt
and
I
mean
I
could
have
toggles
for
disabling
stuff.
But
just
the
thinking
is,
people
might
prefer
being
able
to
just
pick
and
choose
which
pieces
they
want,
but
definitely
don't
want
it
to
get
confusing
like
you
mentioned,
and
it
certainly
could
because
of
those
common
pieces
used
between
them.
A
C
A
Like
they
like
what
is
an
SDK
in
this
context,
and
why
am
I,
including
that
at
least
I
mean
it
wasn't
to
me
I'm
gonna,
when
that
name
was
chosen
and
so
just
stayed
as
open,
Telemetry
and
not
open
Telemetry
SDK,
but
then
there's
open,
Telemetry
API,
which
has
an
open
Telemetry
module,
not
an
open,
Telemetry,
API
module,
so
there's
a
little
bit
of
it
would
allow
sort
of
a
clean
slate
without
being
considered
breaking,
because
I
would
publish
the
existing
packages
again
with
all
the
updates.
A
They
would
just
include.
So
if
you've
got
open,
Telemetry
API,
it
would
include
open,
Telemetry,
API
tracing
open,
Telemetry,
API
metrics.
If
you've
got
open
Telemetry
package,
it
would
include
the
SDK
for
tracing
the
metrics,
and
so
it
would
be
that
people
could
still
rely
on
those
existing
packages
names,
but
it
would
still
sort
of
allow
a
clean
slate
of
getting
the
naming
more
in
sync
with
the
the
other
languages
so
yeah.
Those
are
my
two.
C
C
I
just
was
I,
don't
know,
I
was
in
a
position
where
I
was
going
to
be
I'm,
going
to
be
running
a
session
on
it
like
teaching
people
about
it
and
I'm
like
I,
want
to
address
this
because
it's
I'm
explaining
a
lot
around
it
and
then
I,
don't
I
can't
sort
of
like
authoritatively
say
this
is
the
reason
it's
just
like
here.
Are
my
guesses
yeah.
C
As
well
yeah,
just
some
real
one-on-one
stuff,
like
you
know,
we're
just
teaching
folks
we're
kind
of
just
trying
to
push
the
fact
that,
like
an
open
source
standard
for
metrics-
and
you
know,
tracing
and
then
eventually,
hopefully
logs
getting
more
is
just
we're.
C
Trying
to
you
know
just
kind
of
push
folks
away
from
non-standard
things
so
that
they
could
but
yeah,
and
we
want
to
get
open,
Telemetry
we're
pushing
the
core
team
like
hey,
get
open
Telemetry,
and
we
need
like
a
native
module
for
open
Telemetry
for
nginx,
because
the
the
one
that's
currently
out
there
is
a
little
raw.
C
C
It's
in
C
as
far
as
I
know
yeah,
but
it's
very
sort
of
like
like
I,
was
looking
at
the
traces
that
it
produces
and
it
produces
like.
It
goes
through
the
different
layers
of
the
nginx
like
processing,
Pipeline
and
there's
not
a
lot
of
like
I,
can't
remove
stuff,
for
example.
So
I've
got.
You
know
just
like
a
lot
of
traces
that
just
come
in
that
I.
Don't
need
they're,
they're,
aware
of
this
and
they're
working
on
it,
but
it's
just
like
a.
A
You
know
right
now:
it's
it's
early,
so
yeah,
interesting,
that's
cool!
That's
good!
To
know!
Nginx
is
involved
like
that
I'm,
always
more
hopeful
when
hearing
that,
because
only.
A
Go
ahead,
no
just
some
projects
and
companies
aren't,
as
involved
as
I,
would
hope
or
wish
them
to
be,
but
so
in
hearing
that
that's
awesome.
C
Yeah
and
you've
got
two
folks,
I
think
Steve
popped
in
I'm,
tired,
we're
you've
got
two
Folks,
at
least
on
the
community
side
that
are
heavy
Elixir
Advocates.
A
C
You
might
see
more
Elixir
stop
it's
hard
for
us
to
push
through
because,
like
it's
not
always
an
easy
case
to
make
yeah
but
I
think
I'm,
trying
to
figure
out
a
way
to
slide
it
in
with
other
stuff.
I
do
so
yeah.
C
C
You
know
Dave
McAllister,
Dave,
McAllister,
no
I
think
so
he
used
to
work
for
Splunk
and
he's
our
like
big,
open
source,
Advocate
dude.
So
I
thought,
maybe
maybe
you
didn't
overlap,
but
he.
A
A
Think
no,
it's
fine,
because
the
the!
Unless
did
you
have
anything
specific
you
wanted
to
discuss
or
read.
B
No
I
mean
I've
got
a
thing
so
that
I
could
talk
about,
although,
like
the
consider
new
packages
for
mattress
and
logging
and
tracing
I,
think
that
my
main
comment
on
that
is
the
basically
the
same,
which
is
that
like,
while
the
idea
of
not
having
to
bring
in
the
metric
stuff
appeals
to
me
because
I
just
mostly
don't
like
look
at
okay,
some
will
encounter
a
situation
where
I'm
like
this
is
a
job
for
metrics.
But
for
the
most
part,
events
and
tracing
or
or
where.
B
My,
where
my
mind,
goes.
First,
especially
when
I'm
doing
like
local
development,
I
fire
this
stuff
up
locally
and.
C
C
B
Know
I'm
like
learning
about
the
internal
details
of
various
electric
things
and
so
I
fire
up
Tracy
and
understand,
what's
happening
so
that's
attractive
but,
like
my
main,
like
I,
don't
know,
friction
point
I,
guess
with
the
open,
Telemetry,
erling,
stuff
and
honestly
with
open
Telemetry
in
general.
B
Between
like
I,
like
I,
was
doing
something
wrong
right,
I
was
like
approaching
it
the
wrong
way
in
some
way,
but
I
just
I
always
end
up
in,
like
which
packages
and
which
there's
like
two
different
ways
to
specify
this
value
and
which
structure
am
I
using
and
like
so
that's
yeah
I
can
definitely
see
how
adding
more
more
configuration
points,
especially
at
that
immediate,
just
like
well,
what
mixed
modules
do
you
even
bring
in
like
it
potentially
adds
more
friction
at
that
point
of
onboarding
for
people
who
are
just
trying
to
but
like
I,
don't
know,
this
is
potentially
solvable
documentation,
which
is
also
part
of
why
I'm
like
here
again,
it's
like
what
what's
the
like:
what's
your?
B
C
C
I
I
Echo
your
experience
now,
like
that's
my
my
chief
confusion,
is
just
like
what
what
should
I
put
in
it's
like
riding
that
boilerplate,
because,
like
all
of
the
open,
API
stuff,
it's
you
know
unnecessarily
low
level,
but
every
place
I've
worked
like
we've
had
our
own
internal
library,
that
kind
of
gloms
them
together
and
provides
that
boilerplate
like
we,
we
need
specific
tags.
We
need
specific
things.
You
know
to
be
consistent
across
everything.
So
there's
always
this
need
for
the
but
yeah.
C
Maybe
I'd
also
be
interested
in
knowing
if
there's
stocks,
tasks
that
could
be
contributed
to
because,
like
I
know,
what
I
wanted
like
I
wanted
here
that
here
at
the
top
of
the
readme
like
here.
The
pattern-
maybe
not
the
top
but
like
here,
are
the
packages
that
exist
in
the
maybe
Elixir,
open,
Telemetry
ecosystem
and
like
here's,
the
breakdown
and
here's.
What
you'll
need
for
like
if
you're
just
looking
to
do
metrics
do
this
and
then
here's
like
the
map
and
the
philosophy
behind.
Why?
A
So
yeah
it
looks
like
I
haven't,
made
a
list,
I
really
thought
there
was
one
somewhere,
that's
either
I
or
someone
else
made,
but
I'm
not
doing
a
quick
search
on
the
issues.
I'm
not
seeing
one
so
I
might
need
to
make
one,
but
it
would
also
I
mean
it's
sort
of
if
I
make
it
all.
If
I,
you
know,
try
to
come
up
with
Docs
I'll,
probably
miss
important
things
like
that
like
because
I
already
know
like.
Why
is
it?
A
Why
is
it
this
until
I'll
forget
to
write
those
down
so
I
mean
anything
you're
interested
in
contributing
is
awesome.
Documentation,
wise
I
know
that
it's
Broad
and
it's
really
helpful
to
have
specific
instances.
I'd
say
a
little
narrow
where
it
would
be
I.
Think
of
big
difference
between
what
we
have
currently
and
say.
An
app
signal
or
New
Relic
is
integration
in,
say,
Phoenix,
Ecto
and
all
those
like.
We
have
the
libraries,
but
we
don't
have
docs
around
that.
B
A
The
website
or
anything
so
having
somewhere,
because
a
lot
of
people
they
don't
want,
they
don't
necessarily
even
need
to
do
a
low-level
creating
their
own
spans.
They
would
get
away
and
get
a
lot
of
use.
Just
out
of
you
know
having
Ecto
and
Phoenix
instrumented,
but
then
they
they
start
out
and
they're
like
where
the
hell
do.
I
go
it's
like
oh,
okay,
they're.
They
hopefully
find
the
libraries
that
we
have
in
contrib,
but
still
there's
no
ducks
like
on
the
website
or.
B
Anything
yeah,
so
maybe
a
really
like
there
is.
There
is
the
example
project,
but
maybe
a
really
Phoenix
oriented
guide
of
like
so
you've
got
a
brand
new
Phoenix
project
and
here
are
the
steps
to
get
it
emitting,
and
this
is
what
they'll
look
like
like
I
ask
about
this
because,
like
I,
can't
I
I
cannot
make
promises,
but
I
just
did
this
work
right.
I
just
got
it
loaded
up
fresh
in
my
head.
C
B
You
know
right
now
what
I
do
is
I'm
I,
I,
write
stuff
anyway,
right
I
have
a
newsletter
like
I'm
looking
for
more
workers,
well
I'm
low-key,
looking
for
more
work
as
a
contractor,
but
until
that
until
that
starts
happening
right,
like
I'm
like
okay,
well,
I'm,
making
content,
what's
some
good
content
to
make,
and
so
yeah
I'm
I'm
in
a
position
where
I
actually
could
probably
pick
up.
B
Any
mechanism
around
keeping
them
updated
right,
so
that's
kind
of
what
I
wish
existed.
Honestly
is
a
guy
that
had
also
some
just
like
just
a
little
bit
of
here's,
a
script
that
validates
that
this
guide
still
works
kinds.
B
B
This
stuff
yeah
getting
that
into
the
project,
seems
like
it
would
be.
It
would
be
nice
and
just
in
the
repo.
A
That's
yeah
a
common
issue:
I've
noticed
with
people
coming
in,
they
they're
used
to
either
other
languages
or
other
or
actual
vendor
services
and
doing
stuff
for
them,
which
I
would
like
for
open,
Telemetry
early
Elixir
to
do
for
them,
but
the
getting
the
you
know.
It's
not
yet
prioritized
because
there's
so
much
low.
B
B
Tricky
right
because,
like
you
know,
for
instance,
honeycomb
has
these
like
beautiful,
step-by-step
guides,
for
you
know,
going
from
zero
to
you've
got
a
basic
kind
of
project
instrumented
in
the
languages
that
are
the
big
languages.
So,
and
you
know
I'm
I'm
honeycomb
is
what
I'm
most
familiar
with,
but
this
is
true
of
all
vendor
stuff,
like
if
you're
a
vendor
and
you're
writing
a
guide
for
configuring
open
Telemetry.
B
B
It's
both
what's
great
about
open
Telemetry
and
what
drives
me
just
nuts
about
it
is
that
so
much
of
this
stuff
is
generic
across
projects,
and
it's
designed
to
be
that
way
and
then
you
get
into
a
place
where
it's
like
this
doesn't
quite
fit
the
idiom
around
here,
but
well,
that's
how
it
that's,
how
it
be.
A
Yeah
trying
to
thread
the
needles
now
watch
yourself,
but
I
hope
this
helps
with
going
forward
when
people
come
in,
because
those
configuration
will
even
be
able
to
be
shared
like
between
tutorials,
like
people
would
be
able.
B
B
A
B
Yeah
yeah
I
mean,
like
I,
said,
like
all
the
stuff
there's
this
like
there's
this
trade-off
because,
like
all
the
stuff,
that's
like
frustrating
to
me
about
open,
Telemetry,
I
can
also
immediately
see
how,
in
the
long
run
it's
great,
but
just
like
at
the
cost
of
some
friction
for
me,
like
learning
it
for
the
first
time,
yeah.
A
And
yeah
just
early
adoption
in
general
because
right
you
know,
hopefully
we
will
have
all
these
ducks
and
libraries
and
stuff
and
be
as
easy
to
use.
As
you
know,
if
you
dropped
into
honeycomb
or
dated
dog,
but
Auto.
A
B
A
Oh
yeah
yeah
I
mean
some
of
it.
Yeah
we're
never.
B
C
B
Oh
just
in
Java
world
like
offending
instrumentation,
is
the
the
cloud
native
build
packs
that
build
various
applications.
They
actually
bring
in
the
instrumentation
for
Java
as
just
an
entirely
different.
It's
not
even
part
of
the
code
base.
They
just
bring
in
a
demon
and
figure
it,
and
then
the
like,
and
that's
really
nice
in
a
build
pack
contest,
because
the
build
like
the
build
system
can
just
give
you
instrumentation
yeah
anyway.
C
It
looks
like
it
has
good,
like
patterns
around
Telemetry
like
internal
to
the
app
at
least.
You
know
if
you're
using
the
Elixir
Telemetry
stuff
I
mean
that's
always.
My
knee
jerk
is
like
I
just
want
to
write,
Telemetry
and
elixir
and
then
plug
it
into
whatever
I
I
want
and
I
can.
C
You
know,
put
my
Ecto
whatever
Ecto
Telemetry
generates
through
some
Transformations
and
then
throw
that
out
to
open
Telemetry
I
mean
that's
what
we
did
with
betadog
and
stuff
like
that,
when
we
used
it,
I
used
to
work
for
pager
Duty,
and
that's
where
we
that's.
What
we
did
is
just
like,
have
a
big
Telemetry
consumer
subscriber
file
and
do
it
from
there,
and
that
was
nice.
It
kept
the
code
clean,
yeah.
C
A
A
C
Well,
I
might
keep
an
eye
on
this
issue
and
do
some
experimentation,
and
it
looks
like
when
I
have
some
spare
time
and
make
a
comment.
If
I
have
an
opinion.
C
I'll
keep
an
eye
out
for
documentation
stuff
around
that
or
if
I
can
be
helpful.
All.
A
Right,
yeah
I'll,
try
I'll
make
a
an
issue
around
that
as
well
of
what
I
think
is
missing.
Certainly
one
thing
is
metrics,
but
I'm
hoping
I'll
get
to
that
because
I'd
like
to
pick
the
first
shot
at
it,
but
that's
major
missing
piece.
B
Yeah
I
guess
one
other
thing
that
occurs
me
with
this
issue
is
like
to
what
degree
are,
and
it's
like
impossible
to
know
this,
but
to
what
degree
are
instrumented,
Elixir
and
erlang
applications
being
deployed
in
environments
where
there
are
applications
written
in
other
languages
in
that
same
environment
in
that
same
operational
environment?
Just
because,
if
that
was
like,
if
we
live
in
a
world,
for
instance,
where
every
time
erling
and
elixir
are
deployed,
they're
being
deployed
alongside
and
like
by
an
Ops
Team,
that's
also
configuring
stuff
in
other
languages.
B
So
then
matching
the
open,
Telemetry
conventions
would
be
really
important.
A
Guess
I
don't
actually
know
about
Elixir
as
much
in
their
link
everywhere.
I've
worked,
it's
been
that
that
it's
one
out
of
it's
niche
area
is
using
erling
or
elixir
in
the
company,
and
everything
else
is
you
know
something
completely.
A
B
A
B
Seems
that
seems
compelling
that
seems
reasonable
to
me.
I
also
know
that,
like
I
mean
I
think
that
in
Elixir
it's
a
little
bit
more
common
for
people
to
just
have
an
Elixir
application
because
of
Phoenix,
but
even
the
people
I
know
like
this
is
a
tiny
sample
size,
but
the
people
I
know
who
are
doing.
B
That
kind
of
thing
also
often
have
a
little
bit
of
rest
running
around,
or
at
least
like
have
that
in
their
back
pocket
as
something
that
they
might
put
in,
though
I'm
looking
and
it
doesn't
look
like,
there's
a
rust
opens
Telemetry
project.
Oh.
A
There
is
oh.
A
They
also
have
they
have
a
package
just
called
tracing
yeah.
C
B
B
A
B
And
that's
what
like
in
in
the
scenario
that
I'm
thinking
that's
the
one
that
I
would
want
if
I
was
if
I
was
using
this
pattern,
because
so,
if
you're
like
well,
if
we
have
something
that's
computationally
intensive,
we
will
have
Elixir
orchestrate
it
and
then
we'll
just
you
know,
access
the
rest.
I
think
you
can
even
like
write
I,
think
you
can
just
do
the
bindings
directly
and
like
write
a
function
in
Rust
that
gets
called.
A
B
B
B
A
B
B
A
B
C
I'm
picking
up
Rusty
one
thing
that
was
interesting,
though,
from
like
a
developer
experience
perspective
I
feel
like
a
lot
of
the
stuff
I
like
from
Elixir.
Is
there.
You
know
I
mean
like
in
a
systems
programming
language
like
the
using
cargo
and
using
and
the
type
of
Errors
you
get
in
the
command
line
and
stuff
I
I
was
really
like
nodding,
my
head,
along
with
that,
like
yeah.
This
is
very
different
from
but
I'm
used
to
my
day-to-day,
but,
like
all
of
the
ways
that
you
are
trying
to
help
me
are
very.
C
It
feels
Advanced
yeah.
A
C
I
just
love
the
I
love
the
error
messages
the
way
they
handle
it
like
you
get
these
compiler
warnings
and
they're
just
like
hey.
It
looks
like
you're
doing
this,
and
this
is
like
a
really
common
thing
that
people
run
into.
You
probably
meant
this
and
then
you're
like
okay,
that
just
saved
me,
I
didn't
have
to
jump
out
to
a
browser
or
anything
that
you
just
helped
me
and
I
love
that.
A
C
Well,
it
was
nice
to
meet
y'all
I
might
jump
out
because
I
have
to
go
to
a
meeting,
but
thank
you
very
much
and
I'll
keep
thank
you
for
your
work
and
I'll
keep
checking
in
and
if
you
want
to
ping
me
about
I,
don't
know
kind
of
y'all
have
interesting
ideas
for
Elixir
and
nginx
ping
me
and
I'll
we'll
see.
A
A
B
A
Right
yeah,
but
thanks.
A
And
thank
you
for
future
work.
I
hope,
yeah.
B
Yeah,
we'll
see
we'll
see
yeah
yeah.
That
may
be
my
next.
My
next
thing
that
I
pick
up
is
trying
to
write
taking
a
crack
at
it,
at
least
so.
B
Let
you
know
perfect:
yeah
yeah
yeah
cause
I'm
I'm
happy
to
try
something
but
I'm,
also
very
happy
to
like
work
off
of
footwork
off
of
a
even
just
a
general
yeah
like
this
is
the
direction
because
you
mentioned
last
week
that,
like
getting
Elixir
Elixir
adoption
is
something
that
you're
interested
in
yeah
and
I'm
like
well
yeah.
Getting
some
just
straightforward.
Tutorials
would
probably
help
with
that.