►
From YouTube: CNCF SIG Observability 2021-03-30
Description
CNCF SIG Observability 2021-03-30
A
A
A
We
had
the
call.
Last
week
we
found
consensus
on
within
the
toc,
call
that
basically,
everyone
had
good
intentions
and
and
everything
I
don't
know
if
you
saw
that
email
and
the
intention
for
this
call,
is
to
just
walk
through
the
document,
as
we
have
been
trying
and
once
we're
done
with
the
document
just
hand
it
over
to
to
toc
in
its
entirety,
for
them
to
actually
decide
on
on
what
we
come
up
with.
Also
bartek
with
me
like
20
minutes
ago,
that
he
can't
make
a
call
right
now.
A
A
A
A
A
C
A
A
Okay,
attendee
list
seems
to
be
settled
down
and
settling
down.
So
let's
get
started
for
information
of
everyone.
We
had
a
closed
toc
call
last
week
as
announced
it
went
well.
I
do
hope
that
we
will
be
able
to
simply
walk
through
the
document
at
pace
this
time
and
then
we
hand
over
to
toc
speaking
of
toc.
A
A
So
the
consensus
from
the
last
caller
we
just
put
both
of
the
statements
by
by
open
telemetry
and
by
bartek
as
a
link
or
as
a
direct
thing
in
on
top
ted,
wanted
to
write
a
statement
from
open,
telemetry
site.
But
I
cannot
see
it.
I
don't
particularly
care
how
it
is
linked
if
it's
in
the
document
or
or
if
it's
a
separate
document.
A
I
suggest
that,
for
the
intents
and
purposes
of
this
call,
we
just
acknowledge
that
this
exists
and
if
there
is
anything
which
is
in
the
consensus
down
below
speaking
being
blocked
by
the
above,
we
just
point
to
it
and
move
forward
at
speed,
because
again,
sig
does
not
make
any
decisions.
Anyway,
it
is
the
tlc.
So
we
should
optimize
towards
consensus.
Finding
that
being
said
with
number
three,
we
can
either
try
to
continue
with
this
highly
specific
consensus
building
on
specific
parts
of
of
where
we
have
consensus
or
as
an
alternative.
D
I
think
what
ritchie's
present
proposing
is
either
we
can
keep
what
we
have
currently,
which
is.
We
have
consensus
on
the
first
two
and
then
we
have
the
orange
warning,
which
tlc
can
take
a
look
at
or
we
can
attempt
to
resolve
three
and
four.
So
do
we
want
to
try
to
resolve
three
and
four
or
do
we
want
to
move
on
to
number
four?
I
think
that
is
the
question.
Yeah.
E
A
H
F
F
I
F
D
There
are
building
blocks
already
there
right,
like
the
data
model
exists,
there
was
just
a
new
otep
for
the
instrumentation
libraries
that
was
recently
approved,
there's
initial
poc
and
java
done
and,
of
course
the
collector
has
some
logging
support.
So,
given
that
the
foundational
pieces
are
there,
I
mean
I'll,
let
morgan
and
eloi
to
confirm,
but
I'm
pretty
sure
we're
confident
in
in
2022
for
stability.
D
A
J
C
A
A
L
Just
to
make
sure,
while
you're
making
that
update,
I
have
a
contextual
question
here
for
somebody
who
hasn't
been
a
part
of
the
earlier
conversations.
These
statements
that
you're
making
here
these
are
annotations
against
the
due
diligence
document
or
are
they
to
be
considered
to
be
part
of
the
due
diligence
document.
A
Let's
try,
let
me
try
to
to
answer
slightly
differently
what
we,
the
intention
behind
what
we're
doing
is.
We
took
everything
which
we
could
find
from
toc
in
ways
of
of
due
diligence,
documentation
which
is
which
you'll
also
see
in
the
end
of
the
document
where
it
starts
to
repeat
a
little
bit,
but
these
are
the
things
which
we
could
find
and
we
are
walking
through
this
to
find
consensus
for
each
specific
section,
and
then
we
submit
all
of
this
to
toc
for
toc
to
be
able
to
to
walk
through
it
more
quickly.
A
It's
basically
the
the
sick,
saying
that
we
have
consensus
on
this
and
that
part
of
any
due
diligence,
which
is
how
we
did
the
three
due
diligences
before
and
which
is
also
how
we
agreed
to
do
it
like
last
summer,
with
with
toc
as
a
kind
of
experiment,
if
that
actually
reduces
workload
on
tuc
or
if
this
is
more
or
less
work
which
which,
which
is
wasted.
So
that's
the
intention
behind
it.
Does
that
answer
your
question
or
not?.
L
L
What
what
I'm-
and
I
I
this
is
my
first
due
diligence,
so
bear
with
me
a
moment
and
I've
looked
through
the
templates
for
prior
due
diligence
and
and
looked
at
some
prior
due
diligence
and
a
lot
of
the
questions
are
about
the
project,
not
about
the
opinions
of
a
special
interest
group
that
that's
what
I'm
getting
at
with
my
question
is
that
does
this
become
part
of
the
due
diligence
document,
because
the
due
diligence
is
about
the
project,
not
about
others.
Opinions
on
the
project
usually.
J
Opinion
from
the
syrian,
toc
and
recommendations
are
included
like
at
the
end
of
the
day
as
a
summary,
and
every
section
is
just
read
as
a
story
about
the
project,
the
details
and
stuff
and
all
the
opinions
come
to
the
end.
So
when
the
user
goes
over
the
dog,
they
can
form
their
opinions
and
then
they
read
the
opinions
of
of
the
seek
and
of
the
toc.
J
So
here
we
are
taking
it
like
more
granular
approach,
where
every
section
has
comments,
but
we
can
all
think
of
how
it
will
be
structured.
In
the
end.
It's
definitely
helpful
for
us
as
toc
to
see
the
seek
recommendations
on
each
and
every
item,
but
in
the
end
the
doc
will
still
be
modified
so
to
make
it
like
more
user
digestible,
and
we
can
all
agree
on
the
format.
D
Yeah
and
cornelia
to
actually
answer
your
question
directly
like
this
is
what
we
saw
from
the
the
tod.
Is
that
the
question
that's
asked
to
get
to
incubation.
This
is
open,
telemetry's
response
everything
right
here
and
everything
that's
green
or
orange.
This
is
from
sig
observability,
so
right
now,
sig
observability
is
commenting
on
each
section
determining
whether
there's
consensus
that
these
aren't,
I
mean
technically
there
are
open
symmetry
people
here,
so
they
disagreed.
They
would
state
that,
but
the
official
open
symmetry
response
is
captured
directly
above
okay.
A
Everything
else
is
only
done
in
suggestion,
mode
and
suggestions
and
color
coding,
and
resolving
of
discussions
in
document
is
done
during
the
call
to
to
make
sure
that
we
all
are
on
the
same
page
on
the
specifics.
So
this
is
how
it's
structured,
which
is
largely
copied
from
from
how
promisius
does
it
internally
for
for
the
various
discussions.
E
A
You
nice
so
steve.
If
you
can,
I
know
you
see
I
I
just
restructured
the
sentence
slightly
to
make
it
a
little
bit
more
clear
and,
as
you
saw
I
just
clicked
accepting
on
those
two,
so
sick
observability
has
consensus.
There
is
wide
and
organic
adoption
of
open
telemetry
components
which
are
the
instrumentation
libraries
and
the
collector.
C
A
E
D
D
I
I'm
sorry
to
jump
in
and
I
apologize
if
I've
missed
the
boat
on
questions,
but
I
would
feel
a
lot
more
comfortable
as
an
end
user.
If
there
was
a
fifth
statement
here,
expressing
that
all
three
sub-projects
will
reach
a
maturity
level
as
part
of
the
czech
flag
box
for
getting
past
the
initial
incubation,
because
my
primary
concern
is
making
any
form
of
business
decision
on
the
second
point
looks
very
vague
to
me.
I'm
really
happy
you've
confirmed.
The
third
point
is
2022.
That's
great
news.
I
I
Yeah
because
fundamentally,
we
can
get
a
date
on
the
second
one
and
we
have
the
third
one.
Well.
That
means
that
sure
I
can
understand
you
want
to
push
tracing
through
first
now,
that's
okay,
but
then
we
know
that
at
least
in
2022
the
project
will
ensure
there's
a
maturity
level
across
all
three
sub-projects
that
we're
happy
to
use.
D
Got
it
so
from
from
a
metrics
perspective,
our
goal
is
to
have
something
stable
this
year,
2021.,
whether
that
includes
every
instrumentation
library.
I
mean
I'll
defer
to
ted
eloita,
morgan
others,
but
we
are
definitely
planning
to
have
something
stable
here.
The
data
model
should
be
stable
in
the
next
month
yeah.
I.
K
I
If
you're,
all
in,
if
you're
all
in
agreement,
I
mean
I'll
leave
you
to
decide
the
language,
but
as
an
end
user
I
just
reinforce.
This
would
be
really
really
reassuring.
Just
to
have
a
statement
saying
like
we
do
intend
to
do
all
these
three
things.
This
is
the
kind
of
target
date
and
if
you
could
perhaps
put
that
down,
that
would
make
me
feel
really
comfortable.
A
J
I
I
I
A
A
A
A
J
A
A
I
think
that
this
is
where
we
got
blocked
in
call
one
where
bartik
wanted
to
to
have
his
concern
noted
down.
So
I
would
suggest
that
we
just
write
in
that
the
statement
which
ted
reminder
you
wanted
to
write
and
bartek
and
just
refer
to
this
and
defer
any
decision
to
toc,
because
then
we
don't
have
to
discuss
it.
A
D
J
J
A
D
And
I
think
the
recommendation
on
for
this
diligence
document
the
ask
was
to
move
that
into
a
separate
doc
that
gets
linked
in
because
it's
not
technically
part
of
the
due
diligence
question
and
answer.
So
I
think
that's
still
an
option.
Open
action
item
is:
does
it
stay
at
the
top
of
the
dock
or
does
it
move
somewhere
else.
A
J
A
A
A
A
C
E
D
Correct
nice,
I
also
think
it's.
I
also
think
his
point
is
that
the
first
bullet
will
be
when
we
mark
the
data
model
stable.
The
stable
notion
for
the
data
model
does
not
mean
ga
of
open
telemetry.
I
think
what
he
means
is
only
when
this
one
is
done.
Will
the
project
have
traces
and
metrics
bga?
Is
that
your
reading
eloisa.
E
C
E
C
A
E
A
Document
that
the
project
has
a
fundamentally
sound
design
within
obvious
critical
compromises
that
will
inhibit
potential
widespread
adoption.
That's
another
one
of
those
where
we
got
bogged
down
in
call
one.
I
would
suggest
that
we
simply
copy
and
paste
the
consensus
from
section
number
four
and
try
and
find
consensus
on
that
statement.
E
A
E
C
C
C
C
N
A
A
A
A
J
J
J
E
J
N
O
Sorry,
I
have
a
quick
question
on
that
note:
is
it
generally
except
expected
that
the
signals
and
the
collectors
can
be
decoupled
or
is
it
like?
Is
it
possible
to
actually
use
the
tracing
without
the
collector
yeah?
Okay?
So
so
it's
not
expected
that
you
would
use
the
collector
like
you.
Can
you
can
use
whatever
collector
you
want
as
like,
if
it's,
if
it
supports
the
signals.
D
You
don't
even
need
a
collector,
like
the
instrumentation
could
send
direct
to
a
back
end.
So
there's
flexibility
in
choice
like
if
there's
a
reference
architecture
where
you
could
like.
We
would
typically
recommend
it
because
you
probably
want
to
like
have
infrastructure
correlation
with
the
telemetry
data
being
emitted
from
your
application,
but
it
really
comes
down
to
business
requirements.
So,
if
you
didn't
want
to,
you
could
just
take
a
dependency
on
an
instrumentation
library
and
send
direct
to
a
back
end.
If
you
wanted
to.
O
O
C
A
K
D
A
Oh
yeah,
I
absolutely
do
want
to
because
this
is,
if
yeah
thank
you,
I
I
marked
it
green.
I
know
we
had
consensus
coverages
and
it
was
positive.
Thank
you.
This
is
else
my
system
breaks
down
so
yeah.
Here
we
revisit
stuff
which
we
had
already
resolved
in
call
1,
and
I
suggest
we
do
not
revisit
any
of
those.
C
D
E
A
E
G
N
A
A
D
A
K
D
E
D
So
the
the
link
is
to
the
specification
directly
like
the
specification
lists,
a
lot
of
stuff
resources,
semantic
conventions.
What
have
you
I
don't?
I
didn't
list
all
the
sub
things
I
mean.
The
same
applies
to
the
collector.
The
collector
has
a
bunch
of
receivers.
Exporters-
extensions,
like
I
just
listed
it
by
major
component.
We
can
make
it
more
specific
if
that's
what
people
would
like
or
we
could
just
link
out
and
say
these
are
the
three
major
blocks
and
there
are
of
course,
sub
blocks
within
these
blocks.
J
K
E
E
E
A
M
N
A
That's
the
only
thing
which
stood
out
to
me
because,
having
just
done
the
dance
of
itf,
they
will
want
you
to
be
super
specific
about
every
single
thing,
and
it
might
be
that
you
need
to
split
up
into
several
rfcs.
So
do
you
want
to
put
in
here
what?
Specifically
you
want
to
release
under
itf
or
just
as
a
general
statement
of
intent.
D
K
A
D
J
A
A
J
A
A
J
A
A
A
E
A
A
A
A
A
N
A
A
D
I'm
assuming
we'll
do
that
as
part
of
actually
implementing
it,
like,
I
think,
projecting
when
we
don't
even
have
the
design
of
the
data
model
finalized
yet
is
kind
of
challenging.
We
can
make
a
note
that
clearly
ensuring
that
the
stateful
nature
of
features
is
properly
architected
would
need
to
be
addressed
as
part
of
implementation.
D
Well,
I
mean
I
read
it
as
its
current
state.
Currently
there
is
no
state
whatsoever
like
right.
Now
there
definitely
isn't
in
the
future.
There
might
be.
I
gave
a
different
example,
which
is
like
a
dispatch
that
might
be
required
for
say
the
logging
use
case
and
then
clearly
cumulative
counter
like
metric
type
stuff.
You
might
also
need
state
neither
one
of
those
exists
today.
A
D
D
E
I
mean
richard
as
you
as
you
highlighted.
You
know,
the
current
work
that
is
ongoing
in
the
prometheus
work
group
for
open
telemetry
is
going
to
address
that
we're
working
on
the
design
right
now.
So
would
you
like
to
see
more
detail
there
once
you
know
we
can
link
to
the
existing
design
docs
that
are
in
flight
and
that
could
address
you
know
specifically
the
open
metrics
dependencies,
as
well
as
the
meteor.
A
J
A
The
honest
answer
is,
I
don't
know
because
we
never
talked
about
this
section.
We
are
in
in
territory
which
we
haven't
talked
about
before.
I
see
one
comment
by
peter
burton,
but
I
don't
know
what
he
means
precisely
if
you
cornelia
and
alina
as
members
of
toc,
tell
the
sick
that
we
can
simply
stop
doing
due
diligence
and
you're
happy
with
the
document
as
such.
A
We
can
stop
doing
this
this
due
diligence
work.
If
you
want
to.
If
you
want
to
seek
to
finish
it
will
be
the
next
call,
but
after
the
next
call
we
should
be
done.
I
I
don't
have
an
opinion
if
you
want
to
just
pull
it
into
toc
right
now
say
so
and
done
easy.
If
you
want
the
seek
to
to
to
find
consensus
on
all
the
points,
we
will
need
one
more
call.
I
don't
care
these
items.
A
A
L
So
one
of
the
questions
that
I
go
ahead
karina
yeah.
One
of
the
questions
that
I
would
have
is.
First
of
all,
I
think
that
if
we
hand
things
over
the
toc
now,
which
is
where
I'm
kind
of
leaning,
I
would
hope
that
sig
observability
would
continue
to
partner
with
the
toc
to
provide
feedback.
Should
we
have
questions
either
synchronously
or
asynchronously,
then
the
other
element
that
I'm
thinking
is.
L
I
would
like
to
put
this
to
the
open,
telemetry
team
and
say:
would
you
find
it
helpful
to
continue
to
do
these
things
or
how
do
you
feel
about
handing
this
over
to
the
toc
now
and
again
going
into
the
mode
of
elena,
and
I
will
begin
our
work.
This
context
is
super
helpful,
but
where
we're
missing
context,
because
we
weren't
able
to
go
through
the
entire
document,
all
the
way
that
we
circle
back
to
both
the
open
telemetry
team
and
the
sick
observability.
L
So
I
guess
I'm
asking
both
teams,
the
the
sig
observability
and
the
the
open
telemetry
folks
to
let
us
know
you
know
I
I
assume,
but
I
don't
want
to
assume
that
you're.
Okay,
if
we
take
this
and
then
circle
back
with
you
with
additional
questions
and
maybe
pull
the
team
together
to
discuss
something.
J
J
A
Okay,
so
just
to
also
reply
from
the
sick
point
of
view.
Yes,
of
course,
that's
part
of
what
to
see
you,
there
is
there
for
so
just
to
close
at
where
we
are.
We
will
take
one
more
minute
and
then
we
are
done
I'm.
I
wrote
this
with
my
open
telemetry
head
on,
which
is
which
I'm
now
pulling
off,
and
I'm
putting
my
chair
head
on
and
accepting
this
comment.