►
From YouTube: CNCF Open Telemetry Community Meeting 2019 10 09
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
CNCF Open Telemetry Community Meeting 2019-10-09
A
A
A
A
A
B
B
C
D
C
A
Okay,
all
right:
well,
everyone
I'm
your
host
Morgan,
a
product
manager
at
Google.
This
is
the
monthly
open,
Salama
tree
community
meeting,
where
we
go
over
the
status
of
all
of
these
special
interest
groups
and
open
Salama
tree
and
discuss
recent
items
of
note
and
news
document.
Our
notes
documents
well,
first
off
these
meetings
are
recorded
and
the
recordings
will
be
posted
online.
Secondly,
the
notes
document
is
shared
in
the
community
calendar
link
and
that's
where
we
post
the
agenda
and
take
notes
on
everything
that
was
discussed
during
the
meeting.
A
So
the
first
item
for
today
is
git
er.
As
you
noticed,
the
main
git
er
chat
room
for
the
open.
Celebrity
community
disappeared
this
morning.
We
quickly
recreated
it
I
reached
out
to
git
er
they're
gonna,
try
and
read
all
the
members
who
weren't
reacted
to
the
new
group
and
they're
gonna
see
if
they
can
capture
the
old
chat
history.
A
It
appears
that
someone
accidentally
deleted
the
chat
room,
Gator
has
a
very
interesting
security
model
or
literally
any
one
of
the
chat
room
can
just
delete
the
entire
chatter,
which
is
definitely
not
the
best
they've
had
a
feature
request
that
they
said
they
will
improve
on
that.
They
have
a
feature
request
to
fix
this,
to
create
an
actual
admin
role,
but
it's
been
open
for
three
years
so.
A
E
E
In
order
to
do
it
seem
kind
of
annoying
to
have
to
go
around
to
every
single
repo
after
gets
created,
think
it
this
access
added,
so
I
was
requesting
that
we
do.
This
is
just
like
an
org
wide
setting
of
being
able
to
have
some
gardeners
or
whatever,
whatever
damina
tizen
role
we
want
to
use
to
employ.
This
person
is
not
like
in
charge
of
the
backlog,
but
nevertheless
to
have
this
access.
F
E
That
we
actually
lower
the
requirements
to
being
an
approver
and
I
felt
like
that
might
maybe
be
except
for,
like
you
know,
repo
explosion,
if
that
happens,
having
a
lower
requirement
for
pruners
might
also
be
fine.
If
we
want
to
just
say
approvers,
are
anyone
that
maintainer
czar
fine
with
there's
already
seems
to
be
some
language
in
there
about
saying
approvers
can
be
project
managers
or
tree
huggers,
but
the
requirements
right
now
are
that
you
actually
have
to
do
a
whole
bunch
of
kokum.
E
It's
like,
like
the
literal
letter
of
the
law,
lists
how
many
like
reviews
and
code
commits
and
other
things
you
have
to
do
to
become
an
approver
and
I
was
also
hearing
from
some
people,
maybe
like
that
was
just
too
onerous.
Seeing
people
get
rejected
for
a
prove
your
status.
You
are
code
committers,
and
it's
just
kind
of
slowing
things
down
background.
What
do
people
think
about
this.
C
Issue
by
the
way,
how
do
you
see
somebody
coming
and
triaging
issues
if
they
don't
know
about
the
work?
So
now
it's
kind
of
following,
what's
really
suggested,
I,
think
you
have
to
have
a
bit
of
context
of
what's
happening
in
that
ripple.
You
know
to
be
able
to
to
triage
the
issues
in
that
vehicle.
Correct,
I,
I,.
E
Think
I
think
that's
fair.
If
one
approach
is
like
going
to
like,
say,
Python
and
be
like
hey
everyone,
we've
got
two
engineers
on
Python
we'd,
also
like
our
engineering
manager
to
have
a
prove
our
access,
so
they
can
just
help,
manage
the
backlog
and
if
people
are
like
yeah
sure
I
think
that's
like
an
easy
way
to
do
the
approval
I'm
just
noting
that
right
now
our
only
option
is
approver
an
approver
right
now
just
says
you:
you
have
to
actually
be
a
code
contributor
in
terms
of
the
requirements.
I
think.
E
I'm
Terry,
actually
just
maybe
backing
away
from
just
triage,
errs
like
people
who
are
working
on
the
very
cigs
did
people
feel
like
getting
approvers
has
been
difficult
or
do
you
feel
like
you've
had
to
sort
of
bend
the
rules
in
order
to
get
people
to
become
approvers
in
any
way,
or
is
that,
like
mostly
been
been
okay
for
people
to
kind
of
make
actually
hit
the
requirements
spline?
For
us?
It's
been
fine.
Okay,
yeah.
G
I
just
want
to
know
that
became
an
approver
should
be
very
well
balanced
between
being
easy,
so
people
will
feel
welcome
and
easy
to
get.
At
the
same
time,
it
should
be
like
structure
structure
in
such
a
way
that
there
is
no
I,
don't
like
your
kind
of
reason,
not
to
get
an
approval
approval
into
the
very
poor
Xury
and
the
same
time
you
shouldn't
be
able
to
bring
like
whole
bunch
of
your
friends
just
like
YC
around
and,
like
then
it'll
become
unmanageable.
G
So
there
is
like
both
of
requirements
to
balance
and
we've
been
looking
at
how
kubernetes
is
doing
it
when
we
build
a
membership
document,
so
I
mean
I.
Think
this
request
with
the
asha
role
is
very
interesting,
and
we
need
to
do
something
with
that
and
make
sure
it
doesn't
like
it
balance
all
the
requirements
very
well.
Yeah.
E
E
But
the
other
act
option
to
consider
is
like
the
level
of
access.
We're
talking
about
is
fairly
small.
Assuming
we've
got
people
who
are
involved
with
the
project
in
a
long-term
sense,
so
they're
generally
considered
trustworthy.
You
know.
Could
there
be?
Could
that
be
like
an
alternate
way
to
implement
this?
But
I
would
only
suggest
that
if
we
decide
we
want
to
keep
some
like
coding
and
committed
requirements
on
the
approver
role,
which
I
don't
think
is
necessarily
the
wrong
way
to
go.
H
Reiterate
a
point
I
made
in
that
thread
on
issue
thread:
there's
a
significant
investment
from
various
corporate
interests
in
this
project
and
just
bowing
to
the
realities
of
how
engineering
management
works
in
most
places.
It
feels
like
having
some
sort
of
easy
on
ramp
for
the
managers
of
the
engineers
whose
time
is
predominantly
dedicated
to
this
project.
It
would
be
useful.
You
know,
I,
understand
that
we
don't
want
to
take
power
away
from
maintainer
x',
but
I
also
feel
like.
H
There
are
cross-cutting
concerns
that
having
some
dedicated
gardening
staff
so
to
speak,
they
can
go
around
and
is
empowered
to
create
milestones,
and
you
know,
help
out
would
be
useful,
I
think
it's
something
we
could
least
try
and
if
it
doesn't
work
out
hey
we
stopped
doing
it.
You
know
like
we
can
change
things,
then
we
can
change
them
back.
I
tend.
A
A
C
Think
I
think
there
are
interesting
points
into
management
and
and
and
how
how
management
are
are
in
open
source.
I
would
also
want
to
talk
to
Sarah.
Sarah
has
a
lot
of
experience
into
this
because
I
I
don't
know
exactly
usually
open
source
communities
are
driven
by
engineers.
There
are
not
too
many
managers
employed
in
this
and
I
think
this
is
an
interesting
request
to
have
engineers
in
pried
into
into
into
the
open
telemetry
project,
which
I,
don't
think,
is
bad
or
anything
but
I'm.
Just
saying,
I
would
like
to
hear
other
opinions.
C
How
do
we
handle
this
because
definitely
in
a
corporation
management
manager
has
in
the
end,
has
the
full
power
for
a
lot
of
things,
which
is
not
necessarily
the
case
in
case
of
a
source
project,
so
I
think
I.
Think
I
would
like
to
understand
better
how
we
can
structure
it
is,
but
I
do
understand
the
need
of
management
to
be
employed
into
this,
so
it.
A
C
E
A
Next
governance,
elections,
I
added
this
to
the
agenda.
While
we
were
talking
about
the
last
thing
and
Sergey
since
you
wrote
the
blog
posts
about
it,
I
figured
you
can
help
me
out
and
talk
to
everybody
about
it,
but
the
the
elections
are
on
for
the
open,
celebritrees
governance
board,
I,
think
there's
or
additional
seats
being
added.
A
The
cutoff
for
candidates
to
apply
was
today.
There
were
several
emails,
blog
posts
and
messages
about
this
most
basic
game
and
everybody
who
has
a
sufficient
number
of
get
commits.
There's
an
official
numbers
somewhere.
I
simply
don't
know
what
it's
off.
My
head
is
automatically
enrolled.
As
a
voter,
there's,
also
a
form
you
can
use
to
register
if
you've
made
contributions
that
aren't
captured
through
get
connects,
for
example,
I
think
I
technically
had
enough,
but
I
perform
preemptively
in
the
same
spot.
Yes,.
G
It
is
at
least
in
github
post,
and
this
list
contains
all
people
who
you
forget.
There
is
a
commission
and
emails
associated
with
this
commit.
Some
people
have
contributed
a
lot
and
we
have
them
on
the
list,
but
you
don't
have
any
camis
from
them.
So
there
is
a
separate
list
of
like
I.
Think
four
people
in
this
category
on
the
top
of
my
head
is
Tigran
in
this
no
Thomas
I
think
no.
There
is
not.
C
G
G
G
I
mean,
if
you
or
is
it
a
lot
of
come,
is
just
your
github
account
will
be
sufficient
enough.
Okay,
yeah,
we
will
be
watching
in
October
and
I
need
to
be
bring
updates,
but
there
is
a
vote
and
it
will
be
hold
over
three
days
so
people
in
all
time
zones.
You
have
a
chance
to
hear
about
it,
and
would
you
also
accept
late
registrations
for
waters?
If
you
only
found
out
at
these
three
days
that
you
eligible
to
vote,
and
you
really
want
to
cost
you
what
you
can
still
do
that?
G
But
please
don't
ask
for
like
if
you
want
to
register
a
juster
now,
because
it
will
get
easier
on
us
to
send
you
invite
possible
change
anyway.
Yeah
I'm
really
excited.
We
have
so
many
good
nominee,
nominations
and
whole
idea
of
this
governance.
That
we
elected
in
the
beginning
is
that
we
will
bootstrap
with
five
members,
and
then
we
want
to
have
continuation.
G
C
We
double
checked,
if
any
of
you
are
in
the
list
of
we
do
not
have
your
email
list
at
least
make
sure
you
also
register
at
the
forum
and
say
I
am
that
person
so
that
way,
because
the
vote,
the
vote
link
will
be
sent
by
email
and
the
email
that
we
have
and
by
the
way
for
40
grand
for
you,
we
have
your
old
emails,
but
that
should
be
fine
or
the
emails
that
you
had
at
the
moment
when
we
query
in
github
so
make
sure
you
have
access
to
do.
I.
C
G
Planning
to
do
it
in
two
days,
I
believe
again:
I
really
need
to
pull
up
the
document
and
operate
these
days,
Israel,
not
skater.
So
all
the
nominations
got
accepted,
and
now
we
have
yeah
October
14th
discrimination
and
nominees
will
be
published.
So
merely
we
took
a
couple
days
for
ourselves
just
to
clear
out
the
lease
and
Ansan
like
put
it
like
in
nicely
format
and
ask
additional
questions
if
needed,
and
then
we'll
post
it
October
14.
So
it's
Monday
I
believe.
C
One
comment:
I
want
to
make
about
the
election
process.
If
you
know
somebody
who
applied
just
tell
them
that
we
will,
unless
they
tell
us
not
to
right
now,
we'll
publish
their
appeal.
I
think
Tigran
asked
us
about
this
separately
and
we
responded
to
him,
but
anyone
who
applied
if
they
put
something
in
their
Bo,
that
they
don't
wanna,
be
shared.
Please
make
sure
and
update
otherwise
we'll
publish
that
on
our
list
of
candidates,
yeah.
A
Okay,
the
next
topic
I
had
added
this
as
well
just
an
update
on
Cube
Khan
and
the
observability
summit.
We
talked
about
this
last
month,
so
there's
not
a
lot
of
updates.
Today.
I
know
of
at
least
two
talks
that
have
been
accepted
for
opens
lemon
tree.
Possibly
more
and
we've
also
been
granted
some
done
during
the
keynote,
though
I
don't
know,
if
we've
actually
planned
what
we're
gonna
do
with
it
quite
yet
Ted
did
you
want
to
speak
a
little
bit
about
the
observability
summit
that
you're
hosting
the
day
before
you
come?
Oh.
E
Yeah
I,
just
just
a
shout
out
since
this
is
an
audience
that
I'm
sure
is
interested
in
it
if
you
are
going
to
keep
Khan
on
day
zero.
Just
the
day
before
the
start
of
cube
Khan
we're
gonna
have
an
observability
conference,
some
really
cool
talks.
It's
just
a
single
track
for
a
single
day
and
I
think
it's
250
bucks
on
top
of
your
keuken
registration.
E
A
H
H
Basically,
it
doesn't
appear
that
any
of
the
CNC
f
zoom
counts,
have
API
access,
so
no
web
hooks
no
off
BOTS,
no,
nothing!
It
yells
about
account
admin
access,
so
I'm,
mostly
putting
this
on
here
as
a
generic
plea.
If
anyone
knows
people
at
CMC
F
that
can
alleviate
this,
because
it
seems
bonkers
to
me
that
that
would
be
the
case
because
I
believe
kubernetes
uses
CN,
CF,
zooms
and
I
can't
imagine
that
they
handle
archiving
manually.
F
G
G
H
B
H
Problem
is:
is
that
the
zoom
accounts
don't
have
the
level
of
access
they
need
to
use
the
developer
API,
so
I
can't
create
a
web
hook.
For
you
know,
a
meeting
being
over
I
can't
create
like
a
integration
on
the
account
that
using
OAuth
or
anything,
it
just
seems
like
that
permission
isn't
available
to
these
hotel
zum
accounts.
H
A
F
So
yeah
I
could
talk
to
Chris
or
Dan,
because
they're
definitely
want
to
get
more
involved
in
hotels,
so
yeah
great.
F
B
I
A
Alternative
is
we
just
have
the
host
of
each
meeting
be
responsible
for
it,
like
I'm,
happy
to
do
it
for
the
community
meetings,
but
you
know
15
a
month,
probably
shouldn't
be
assigned
to
any
one
person.
Oh
yeah
15
a
week
15
a
week.
Sorry,
and
do
we
want
to
shift
to
a
policy
where
the
the
one
person
from
each
meeting
is
responsible
for
it?
Just.
H
E
F
H
G
J
D
C
C
Probably,
today,
I
just
left
two
minor
comments
on
that
thing
and
then
the
rest
of
the
work
is
going
to
be
a
bit
rewritten
and
we're
gonna
try
to
make
progress,
though
the
advantage
will
be
that
now
it's
becoming
way
smaller
and
way
easier
to
review
and
stuff,
so
yeah
well,
I
think
I
think
it
was
a
combination
of
things
like
last
week.
I
was
in
New,
York
I
didn't
have
time
to
look
at
this,
but
I
think
I'm
doing
a
good
job
with
Josh
right
now
and
his
is
very
responsive.
Okay,.
J
D
I
So
a
couple
updates,
so,
first
of
all,
it's
now
called
collector.
No
longer
a
service,
we
renamed
that
the
name
was
causing
a
confusion.
After
discussing
with
the
community,
we
decided
to
rename
it
to
a
collector.
That
is
cheering
yes,
the
github
repository
and
all
go
package
names
are
also
renamed
from
open,
telemetry
Darsh
service.
Don't
let
me
to
director
so
if
you're
importing
this,
you
will
need
to
update
your
import
path.
I
Number
two.
We
released
collector
alpha
0.2
last
week,
which
means
that
the
configuration
file
format
is
now
considered
stable.
We
at
least
we
aim
not
to
make
breaking
changes
to
the
format,
but
it's
still
alpha,
but
we
aim
for
that.
The
next
release
0.3
will
aim
to
have
stable
public
interfaces
for
packages
that
are
intended
for
consumption.
I
This
is
important
for
people
who
create
custom,
builds
of
the
collector
and
and
also
we
decided
to
align
the
collector
releases
with
spec
releases
number
wise,
so
the
alpha
0.3
will
be
released
in
November
and
if
specs
ik
finalizes
the
open
flame
tree
protocol
early.
That
release
will
also
include
the
open,
telemetry
collector
sorry,
protocol
implementation
in
the
collector,
but.
C
I
I
C
G
G
So
I
hope
that
situation
will
get
better
when
we
entered
like
this
week
and
by
the
next
week.
So
if
you
have
energy
to
healthy
specifications,
please
help
us,
as
Jigar
said
specification,
sometimes
block
other
like
broken
limitations,
because
people
were
expecting
to
see
what
needs
to
be
granted
and
how
like
how
to
handle
questionable.
Appreciations,
okay,
yeah.
E
And
just
a
simple
process
in
case
people
don't
know,
Sergey
has
been
doing
a
great
job
of
organizing
everything
by
milestones
so
on
both
the
specs
repo
and
the
Odette's
repo.
You
can
search
by
milestone
and
select
point
two,
and
you
will
see
the
open
issues
that
are
like
specifically
related
to
this
release.
I
think
those.
C
H
G
E
Suggest
it's
really
the
metrics
that
are
the
thing
we'd
like
to
see
in
point
two,
because
that's
the
one
part
of
the
API
it
seems
like
SIG's
ever
looked
and
to
put
effort
into
like
everything
else.
People
already
have
something
so
soon
as
metrics
gets
done,
I
would
recommend
releasing
point
two
and
everything.
C
C
D
Who
wanna
call
her
go?
Clay's
part
is
done
for
me
to
release.
Metrics
is
still
what
being
worked
on
a
few
changes
that
are,
as
Marvin
mentioned,
he
and
Josh
talked
about
it,
and
those
changes
are
gonna
happen
so
again
for
in
this
probably
not
going
to
be
possible
for
matrix
world,
probably
21
for
go
as
well.
I.
E
G
D
K
L
Okay,
Python
I
think
it's
usually
Chris
lip
so
where
largely
in
the
same
place
as
JavaScript,
we
did
the
code
I
wonder,
release
last
week
and
then
you've
got
I,
went
to
work
everything
now
for
the
two
alpha
release.
We
do
have
a
problem
like
all
the
SIG's
have
this
problem.
Now
that
we
have
milestones
in
the
spec
repo.
L
Let's
say
you
know,
for
example,
metrics
exporters
will
be
ready
for
the
like
the
o2
alpha
yeah,
but
none
of
this
is
written
yet
so
when
we
said
josh
is
working
on
the
SDK,
the
SDK
means
a
lot
of
things
right.
It's
not
just
the
implementation
with
the
API.
It's
also
all
of
the
stuff
that
we
buried
in
the
SDK,
which
includes
the
data
model
and
exhalation,
and
all
of
that
is
going
to
be
more
work
than
people.
Imagine
right
now,
like
I,
don't
think
it's
realistic!
L
D
D
L
D
L
Think
we're
gonna
need
to
start
distinguishing
between
spec
milestones
and
implementation
milestones,
because
right
now
you
know,
we've
got
a
date
for
one,
but
if
the
spec,
you
know,
let's
say
just
barely-
hits
their
milestone
target
date
and
they
they
have
one
day
to
spare.
That
means
we
have
one
day
to
implement
them.
So
I,
don't
you
know
it's
I.
E
Think,
as
far
as
like
messaging
around
the
project
goes,
it's
realistic
for
us
to
announce
project
wide
stuff
around
like
this
spec
releases,
but
as
far
as
like
implementations
go
I
feel
like
that's
gonna,
be
all
over
the
map.
Right,
like
wind,
go,
get
something
out
the
door
versus
Python
versus
Ruby
versus
PHP,
which
just
started
last
week
like
it
probably
won't
be
realistic
to
have
some
big
official.
This
is
the
release
of
everything
and
all
the
languages
kind
of.
E
But
my
main
issue
with
that,
because
we
thought
about
that
before
it
I'm
sure
that
people
have
opinions
by
the
thing
I've
noticed
is
right.
Now
those
different
observability
api's
tend
to
have
some
amount
of
shared
surface
area,
so
there's
some
concepts
that
are
shared
across
them
and
particularly,
we
want
to
be
like
correlating
these
various
kinds
of
observations,
and
so
that
made
me
feel
like
it
would
be
like
a
step
too
far
to
at
this
point,
be
like
they're
all
treating
them
like
they're
separate,
independent
verticals
is
like
not
necessarily.
L
C
Final
and
we're
going
to
be
holding
a
tracing
of
metrics,
we
should
we
should
discuss
if
there's
a
gate,
I
think
I
think
we're
a
bit
far
from
one.
We
know
that
until
I
still
think
we
have
features
to
solve
for
one.
We
know
aspects
to
define
for
1.0,
even
for
traces.
So
once
we
will
get
closer
and
this
will
become
a
problem,
we
can
yeah.
E
B
L
H
Think
we're
much
in
the
same
place
as
other
SIG's.
We
do
have
work
going
in
some
stuff.
Around
benchmarking
came
in
this
week,
which
is
good.
More
integration
with
the
node
ecosystem
tracing
has
been
in
a
good
place
for
a
while.
So
that's
that's
already
out,
but
we're
kind
of
in
the
same
boat
with
everyone
else
on
metrics.
H
A
C
A
A
L
A
C
H
It'll
be
the
first
cig
meeting
tomorrow.
So
if
you're
interested
in
working
on
the
website
or
involve
the
website
stuff,
please
show
up.
One
thing:
I
did
notice
is
we
need
to
get
the
meeting
notes
on
the
calendar?
Invite
but
yeah
and
Constance
I
think
you
mentioned
documentation
and
the
website
is
just
really
looking
for
quick
start
guys.
H
H
H
C
H
There
is
some
stuff
that
is
in
progress.
Josh
wrote
a
QuickStart
guide
for
go
that
I'll
be
trying
to
repurpose
into
that
there's
some
c-sharp
QuickStart
guides
already
is
so
I've
just
been
traveling
and
doing
a
lot
of
other
writing
and
stuff
this
week
and
haven't
had
a
chance
to
get
to
it
but
hoping
by
the
end
of
the
month.
We've
got
all
that
fleshed
out
in
terms
of
quick
starts
for
tracing
at
least.