►
From YouTube: CNCF TOC Meeting 10.17.18
Description
Join us for KubeCon + CloudNativeCon in Barcelona May 20 - 23, Shanghai June 24 - 26, and San Diego November 18 - 21! Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy and all of the other CNCF-hosted projects.
A
A
D
E
A
G
B
D
D
D
C
D
D
A
A
We
have
a
short
presentation
from
Richard
I'm
over
metrics
and
I.
Think
that's
kind
of
it
for
today.
So
I
believe
that
we
can
go
straight
into
slide
seven
and
eight,
which
is
the
operating
principles.
You
can
see
they're
on
github
link
to
the
current
document,
which
Dan
has
cleaned
up
and
put
onto
github
and
I.
Think
I
can
just
send
straight
over
to
actually
okay.
E
E
We
want
to
make
sure
that
if
there's
needs
out
there,
especially
in
the
you
know
the
projects
that
that's
getting
raised
up
quickly
to
to
the
board
as
well,
if
there's
a
resource
need
in
particular
as
an
example-
and
you
know
there's
enough
of
us
now
around
the
table,
but
that
we
can
really
go
and
help
with
those
things.
But
but
in
terms
of
the
principles
you
know
we
met,
we
had
an
off-site
with
most
of
the
governing
board
present.
Obviously,
we've
added
new
members.
E
It
was
at
a
point
where
AWS
had
joined
Oracle
had
had
joined
in
Dell
had
joined,
so
they
were
able
to
be
included.
Alex
was
there
as
well,
too,
and,
and
we
had
a
really
good
session,
and
we
talked
about
the
things
that
were.
You
know
important
to
us.
The
strategy
going
forward
as
well
as
Alex,
did
a
great
job
of
representing
you
guys
with
the
the
principles,
discussion
and
and
I
think
that
all
of
the
board
members
walked
away
with
that
feeling.
You
know
pretty
comfortable
right.
E
E
So
you
know
as
cleaned
up
and
as
posted
I
think
that
what
what
you
guys
have
done
together
a
lot
with
Dan
and
others
is,
is
really
a
pretty
good
piece
of
work
and,
of
course
we
want
to
see
end-users.
That's
the
other
part
of
our
three-legged
stool,
now
interact
and
give
us
feedback
and
communicate
to
us.
You
know
real
requirements
and
their
experiences
so
that
we
as
a
group,
can
learn
from
that
and
move
forward
as
we
go
and
pick
other
projects
and
work
to
support
the
projects
that
are
underneath
us.
E
So
I
think
we've
got
a
good
structure.
We've
seen
this
used
in
lots
of
places
and
it
seems
to
work
pretty
well
one
of
the
things
that
it
needs
obviously
is
communication
among
us,
although
so
we
have,
you
know,
obviously
our
own
charter,
it
it's
been
defined.
We've
grown
to
about
20,
platinum
members.
Now
boy,
they're
gonna
actually
look
at
the
list
to
think
about
it.
So
Amazon
Cisco,
Cordell,
docker,
frigate
to
Google
Huawei,
IBM,
Intel,
Joe
and
mesosphere
Microsoft
Oracle,
pivotal
Red
Hat
sa
P,
Samsung
supernet
VMware
right,
you
know
it's.
E
It
gets
hard
to
to
put
that
list
out.
There
now
I
think
it's
it's
pretty
impressive,
that
we
have
so
many
heavyweights
in
the
industry
willing
to
put
money
and
time
into
this,
and
it
really,
you
know,
speaks
volumes
to
the
work
that
we've
done
on
the
project
side
of
things
right
so
and
and
that
project
side
has
obviously
been
influenced
by
what
you
see
in
the
TOC.
A
principals
document
and
I
think
we're
in
pretty
good
shape
there.
E
Obviously,
we
have
as
a
board
now
added
in
some
new
developer
seats
and
we'll
have
another
set
of
folks
there
for
through
the
community
coming
in
and
giving
us
suggestions,
ideas,
you
know,
being
part
of
the
board
helping
to
make
the
decisions,
so
that
will
that
will
help
us.
Obviously,
with
that
end
user
discussion.
E
You,
okay,
Chris,
I!
Think
you
sorry,
that's!
Okay,
so
obviously
we
we
as
a
group,
you
know-
are
more
than
just
that
so
with
our
member
companies,
our
TOC
chair
being
part
of
it,
it's
gotten
to
be
quite
a
large,
unwieldy
sort
of
board
in
many
ways
when
you
have
boards
in
this
large.
What
you
do
is
you
you
go
to
a
committee
structure,
so
more
and
more
we're
gonna
be
depending
on
subsets
of
the
board
to
actually
get
together
work
with
the
membership
work
with
you
guys
as
well
to
go.
E
Do
you
know
in-depth
analysis
or
help
with
certain
projects
or
do
things
like
financial
reporting
and
things
like
that?
So
that
way,
the
whole
board
doesn't
have
to
meet
and
tackle
a
specific
subject.
You
know,
as
a
group
will
will
be
you
know,
using
the
membership
as
best
we
can
to
cover
all
the
bases,
and
hopefully
do
that
very
effectively,
because
a
lot
of
people
worry
when
you
see
a
board
this
large
will
it
be
effective
and
I
think
the
committee
structure
will
help
us
go
and
do
that
next
next
slide.
E
E
Obviously,
as
I
said
presented,
that
I
think
the
most
important
thing
was
that
we
walked
away
as
a
as
a
group
from
that
that
discussion,
having
really
good
broad
consensus,
that
that
was
how
we
wanted
to
operate
and
that
working
with
the
TOC
and
working
with
Alex
to
clean
up
the
dock
and
and
talk
about
the
things
that
were
we're
going
on
at
the
time.
You
know
we,
as
the
board
could
get
behind
that
pretty
pretty
solidly.
The
document
I
think
is
in
really
good
shape.
E
Don't
think
we'll
have
any
issue
with
that:
okay,
excite
yeah,
so
our
budget
is
much
bigger
with
you
know:
twenty
Platinum's
we
as
a
group
can
afford
to
do
some
things
now
and
that's
good.
So
a
lot
of
the
feedback
we
got
from
Alexis
was
you
know?
How
could
we
go
and
help
the
projects
and
support
them
better?
Some
of
the
things
like
you
know
documentation
and
editors
and
things
like
that.
E
You
know
the
membership
can
provide,
but
then
you
know
other
things
infrastructure
and
what
have
you
you
know
we
can
fund
and
do
so
so
we've
got
some
money
now
and
hopefully
you
know
working
together
with
you.
Guys
will
continue
to
bring
in
really
high-quality
high-caliber
projects,
but
also
we
can
fund
and
fill
the
gaps
pretty
quickly
as
we
need
to
because
the
money
is
there
and
the
GB
wants
to
do
that
as
a
governing
board.
E
Pretty
much
you
know,
the
the
thing
that
we
can
do
is
approve
spending
right,
we're
as
a
nonprofit
required
to
have
all
of
our
spending
approved
through
the
board,
and
you
know
we're
there
to
just
go
and
and
do
and
help
that
wherever
we
can,
both
with
resources
and
dollars.
Of
course,
so
we're
also
been
working
to
you
know
make
sure
that
we
have.
You
know
the
projects
better
supported.
We
came
up
with
this
idea
of
the
service
desk
and
the
service
dashboard
and
Chris.
Why
don't
you
just
flick
to
the
next
slide
right?
E
That's
up
and
going
now
and
and
hopefully
people
have
have
looked
through
it
and
feel
comfortable
with
it.
You
know
the
idea
here
is
that
will
be
much
more
organized
now
about
how
we
go
and
hear
the
needs
of
people
and
and
Dan
and
Chris
are
really.
You
know
there
to
help
the
community
pretty
much
anything
that
the
community
has
been
asking
for.
I've
seen
that
they've
gone
and
and
just
done
right,
and
we
want
to
continue
to
foster
that
and
make
sure
that
that
that's
that's
what
we
see
from
the
organization.
We
don't
want.
E
Any
roadblocks
in
the
way
of
the
projects-
and
this
will
be
the
centralized
way
of
going
about
and
doing
that
so
you
know
we
get
to
the
end
and
you
want
to
talk
about
that
at
all.
I
think!
That's,
that's!
You
know
our
first
answer
and
if
there's
other,
so
we
need
to
do.
Obviously
we
can
come
together
and
talk
about
that
as
well
too.
So.
E
The
other
thing
that
we
as
a
group
need
to
protect
our
trademarks.
We've
started
a
conformance
program
with
kubernetes.
It's
the
first
one
to
really
take
up
the
mantle
to
go
and
do
this.
Kubernetes
has
become
pretty
important
to
the
world
and
having
interoperability
is
something
that
we
as
a
group
value
very
highly
and
therefore
you
know
working
with
the
team
in
the
community.
E
We've
gone
off
and
helped
to
sponsor
this
and
and
make
it
work
and
we'll
provide,
of
course,
the
infrastructure
and
whatever
it
takes
to
make
this
a
successful
project
for
the
for
the
community.
But
obviously
the
oversight
of
what
goes
into
the
witness'
certified
kubernetes
is
is
through
the
technical
team
and
the
sink
structure
that
we've
got
in
place
and
that's
the
way
I
think
we'd
like
to
go
and
continue
that
as
we
go
from
project
to
project
and
work
through
whether
we
we
do
a
certification
program
or
not.
E
But
so
far
the
feedback
has
been
really
good.
I
think
we've
seen
even
a
couple
of
architectural
issues
get
raised
and
and
decided
on
pretty
quickly,
as
folks
have
tried
to
run
the
conformance
tests.
You
know
performance
tests
are
never
enough.
You
continually
work
on
them.
I
expect
this
will
be
something
that
goes
on
for
quite
a
while,
and
you
know
I
think
it's
something
that
we'll
take
very
very
seriously
as
we
go
about
doing
this.
It's
important
to
police
and
manage
our
brands
transparency.
E
So
one
of
the
things
we'll
do
is
obviously
publish
an
annual
report
that
says
where
we're
spending
our
money.
As
I
mentioned
earlier
in
the
committee
structure,
we've
now
organised
a
finance
committee
so
that
you
know
now
that
we
have
a
lot
of
money
and
we're
trying
to
figure
out
how
to
spend
it
and
and
keep
oversight.
We
will
have
a
committee
that
goes
and
addresses
that
and
and
it's
there
to
make
sure
that
the
needs
of
projects
are
being
met.
E
I
know
alexis
has
been
working
with
dan
to
organize
those
so
that
at
the
face-to-face
as
an
example
in
in
Austin
here
yeah
Austin
we
can.
We
can
get
together
and
mingle
for
a
while
and
and
spend
some
time
talking
about.
What's
what's
going
on
and
and
just
make
sure
we're
interacting
and
you
know,
have
a
relationship
across
the
the
GB
and
the
TOC.
So
anything
you
guys
need
you
know
where
to
go
and
how
to
get
a
hold
of
us
as
well
too,
and
then
just
you
know
my
doors
open.
E
You
know
literally
figuratively
online,
just
ping
me
if
there's
anything
that
you
need
or
questions
you
have
or
anything
related
to
how
the
GB
operates.
Or
you
know
feedback
you
want
to
give
us
or
something
get
something
scheduled
on
an
agenda
or
anything
in
those
matters,
I'm
happy
to
go
and
help
and
follow
up.
So
that
was
that
was
my
little
spiel.
E
E
A
A
We
don't
always
have
representatives
of
the
individual
open-source
projects
in
the
CNC
F
joining
these
calls.
They're,
probably
not
aware
of
all
of
this
in
this
format.
What
is
your
plan
for
briefing
them
on
this
and
involving
them
and
getting
also,
you
know
a
health
check
from
those
projects
to
tell
the
TOC
how
things
are
going.
One
lower.
D
D
We've
also
served
we've
also
surveyed
our
maintainer
community
and
we're
going
to
be
publishing
that,
hopefully
you're
not
coming
to
UC,
meaning
in
terms
of
you
know
their
status.
Satisfaction
with
our
services
on
some
things.
What's
working
and
what's
not
so
we're
planning
to
do
that
on
is
probably
six
month
basis
twice
a
year
if.
D
A
F
A
E
Think
at
this
point
to
get
us
over
the
line-
it's
you
know
we'll
do
an
email
thread,
obviously
back
to
the
governing
board,
but
then
make
sure
that
we
can
bring
it
to
a
vote
in
the
12
5
session
so
that
when
we
all
get
together,
I
think
we're
gonna
have
dinner.
After
the
TOC
and
the
episode
of
a
joint
gbto
sea
dinner,
we
can
mingle
and
talk
about
it,
but
by
then
I
would
expect
that
we'll
have
it
have
it
put
together
and
voted
on
and
agreed
to.
A
E
I
would
say
something
like
30
days
before
we
can
publish
it
out,
which
means
in
a
couple
of
weeks
we
put
it
up
and
ask
for
people's
comments
heading
into
the
the
GB
meeting.
That
would
give
us
an
inform
set
of
comments
as
well,
for
when
we
go
to
actually
discuss
and
vote
on
it.
So
I
think
that's
an
excellent
suggestion.
A
A
F
A
C
F
F
An
effort
to
spin
out
the
parameters,
exposition
format
into
its
own
project.
It
will
change
in
some
regards,
but
it's
basically
the
same
ish.
We
do
this
for
two
main
reasons:
a
it's
or
purely
purely
political
reasons,
because
it
makes
it
easier
for
projects
or
companies
to
adopt
something
when
there
is
not
an
already
established
name
attached
to
it,
which
might
or
might
not
be
seen
as
a
competitor,
pointing
Hayes
influx
data
which
really
made
a
huge
decision
by
actually
supporting
fermitas
format.
F
F
Is
we
really
want
to
try
to
push
a
a
certain
kind
of
thought
model
into
the
world,
for
people
to
realize
how
to
properly
for
some
value
of
properly
to
metrics
and
to
monitoring
and
to
basically
even
able
people
who
are
not
using
premises
to
still
use
the
benefits
of
the
format,
in
turn,
obviously
enabling
Prometheus
to
to
just
natively
access
more
data
sources.
So
it's
win-win,
but
these
are
basically
two
main
motivations
from
from
our
side,
a
this
detaching
from
an
already
established
name
and
B
to
get
this.
F
This
thought
model
out
into
the
world,
so
what
we
are
doing
is
we
have
bi-weekly
meetings
or
conference
calls
where
we
go
through
the
open
issues
as
we
see
them.
Sometimes
this
is
quick,
sometimes
as
it's
lower
partially.
As
a
result
for
of
that,
we
met
for
two
days
last
week
at
Google
London
and
went
through
everything
by
everything.
F
So
anyway,
there's
a
list
of
projects
and
companies
which
attended
you
can
probably
read
them
more
easily
than
I
can
read
them
out,
so
you
probably
have
already
so
the
next
slide
I
think
that's
the
first
time
I.
Actually
can't
accuse
for
thank
you.
So
basically,
there
were
a
lot
of
misunderstandings
because
turns
out
many
companies
and
projects
do
things
differently
internally,
and
so
the
the
underlying
assumptions
tend
to
be
different
and
it
was
a
lot
easier
to
iron,
those
out
when
we
were
sitting
face-to-face.
So
there's
still
a
lot
of
disagreements.
F
None
of
them
are
our
heart
blockers,
but
still
they
need
to
be
worked
out.
Basically,
we
are
currently
in
the
face.
We
wait
for
feedback
from
people
who
were
not
on
site
to
to
write
in
their
feedback
on
the
discussion
and
on
on
the
on
the
meeting
notes.
So
we
get
also
their
point
of
view
and
based
on
that
I'm
going
to
start
establishing
rough
consensus
on
all
the
things
which
which
can
be
closed
and
we
go
from
there.
F
F
That's
actually
a
next
slide,
if
you,
if
you
want
to
jump
because
we
expect
the
the
format
in
proto
just
to
make
it
explicit
what
actually
we
we
want
to
have
in
there.
What
is
required
blah
blah
blah?
Well,
there's
no
requirements
and
proto
III,
but
you
know
what
I
mean,
but
there
will
be
a
hard
requirement
for
a
text
format,
two
simple
reasons:
a
ease
of
adoption,
because
it
what
promises
exposition
format
currently
does
is
it
makes
it
really
really
really
easy
for
new
people
to
get
their
stuff
in
two
parameters.
F
I've
helped
a
lot
of
people
and
a
lot
of
different
projects
get
their
stuff
into
Prometheus
and
having
a
very,
very
low
barrier
of
entry
has
always
proven
to
be
one
of
the
key
factors
into
therefore
getting
people
into
the
ecosystem
early,
and
then
they
realize
why
this
is
good,
but
at
first
they
just
need
to
be
able
to
start,
and
the
other
thing
is
easier
with
debugging,
because
obviously
just
using
kerlor
or
your
web
browser
is
easier
than
having
to
install
any
tools
to
do
actual
debugging
or
development.
Work.
F
Then
there's
a
point
which
might
be
surprising,
which
is
why
I
marked
it
in
red.
He
went
the
generic
term.
Events
will
become
first
class
optional,
as
in
it
will
be
designed
into
the
spec.
It's
not
quite
sure
if
this
will
be
the
first
version
of
the
spec
or
if
there
will
be
an
appendix
appendix
or
the
second
version,
or
maybe
it's
in
the
first
version,
that's
still
not
really
really
sure,
but
the
gist
is.
F
We
will
have
support
native
support
for
events
in
the
general
case,
those
obviously
split
out
into
interlocks
and
traceless
and
all
other
kinds
of
stuff,
but
these
are
the
two
most
important
ones.
Obviously
this
also
touches
upon
other
projects
which
are
within
C
and
C
F,
and
we
didn't
talk
to
any
of
those.
Yet
if
you're
in
both
those
feel
free
to
to
invite
them
to
talk
to
us
and
or
tell
me
who
I
should
talk
to
other
than
the
obvious
ones,
because
we
realized
we.
F
F
And
basically
the
plan
is
if
you've
got,
for
example,
Prometheus,
which
only
understands
metrics,
then
can
simply
disregard
or
not
even
even
grab
that
kind
of
theta
and
just
grab
what
it,
what
it
understands
and
other
projects
or
other
software
which
might
understand
a
different
subset
or
a
totally
different
set,
might
grab
different
types
of
data.
But
it
basically
all
follows
one
single
spec
which,
which
is
defined
to
work
with
itself
yeah.
F
Also,
there
will
be
a
test
suite
and
this
will
probably
become
the
official
bar
to
pass
to
actually
be
called
compliant
to
to
your
open
matrix,
because
this
takes
all
of
the
politics
and
discussions
are
off
stuff
and
also
it
takes
out
quite
a
bit
of
interpretation
out
of
the
spec,
because
there's
a
reference
implementation
either
you
pass
or
you
don't.
And
finally,
we
are
still
committed
to
to
write
in
your
RFC
because
with
traditional
vendors,
it's
just
easier
to
toss
them
an
RFC
in
a
tender
and
tell
them.
F
F
There
was
the
open
question
about
what
CNCs
can
actually
do
to
help
open,
metrics
there's
three
points
which
we
identified.
First,
once
we
start
the
int
F
process,
one
of
the
things
which
is
invaluable
and
will
help
us
speed
this
process
up.
Quite
a
lot
is
for
people
who
are
already
involved
on
the
ITF
mailing
lists
to
just
voice
their
support,
not
just
without
constraint.
F
Obviously,
everyone
is
free
to
just
waste
their
their
opinion,
but
if
there's
a
hundred
new
people
who
who
yell
about
a
new
thing,
this
might
be
perceived
as
trying
to
basically
astroturf
things.
I'd
prefer
for
people
to
be
already
involved
within
ITF.
So
if
you
or
your
member
projects
or
companies,
know
people
within
the
ITF
now's
the
time
to
tell
me
about
this,
I
know
I
already
made
this
call
once,
but
did
that
go
anywhere
or.
F
D
F
Is
something
okay,
and
so
whatever
mechanisms
you
have
to,
for
example,
contact
your
Platinum
members
or
all
members
or
whoever
there's
a
few
establish
them,
for
example,
Cisco
and
quite
a
few
others
who
have
people
who
are
active
within
ITF
feeling
around
and
just
getting
names
of
people
and
contact
information
coming.
Ok,
this
effort
exists
retrieve
it
through
it.
Do
you
agree?
Do
you
disagree?
Can
you
support
this
cause?
F
F
The
second
thing,
which
is
probably
even
more
generic.
If
there
is
other
projects
and
companies
who
could
benefit
from
this,
they
should
be
made
aware:
I'd
like
two
white,
getting
into
a
situation
where
there's
tons
of
bikes,
shutting
and
just
people
heaping
on
new
requirement
requirements
and
we've
got
feature
creep,
I
know
tomorrow.
So
we
need
to
constrain
the
total
amount
of
people
who
voice
an
opinion
in
a
too
early
phase
to
some
extent.
F
Yet
they
should
be
aware
that
this
is
coming
and
there
is
a
lot
of
companies
which
which
have
similar
stuff
in-house
or
just
could
benefit
from
this,
so
low
level
pushing
out
the
information
that
this
exists,
and
this
will
become
a
thing
would
also
be
helpful
and
the
last
thing
we
are
in
the
process
of
deciding
on
our
code
of
conduct
for
obvious
reasons.
What
we
do
not
have
is
any
non
males.
F
We
have
non-white
males,
but
we
don't
have
any
non
males
and
speaking
from
experience
from
other
projects,
I
would
usually
prefer
to
have
at
least
one
female
or
non
male,
or
what
have
you
on
the
list
of
people
to
to
poke
when
anything
happens
with
regards
to
the
code
of
conduct,
as
we
don't
have
any
and
I
know,
you've
got
females
on
staff
or
also
in
the
talk
and
in
other
places.
I
would
hugely
prefer
to
have
a
wider
scope.
F
F
F
Okay,
perfect
yeah,
so
the
last
question
which
I
have
from
my
side,
if
is
there,
is
an
actual
need
for
a
face-to-face
in
use
and
I
saw
that
you
also
had
face
to
face
or
an
offside
for
or
for
the
talk.
If
there
is
a
need
to
tell
me
if
there
is
no
need
just
as
well,
and
that's
it
from
my
side.
Basically
Austin
use.
Okay,.
A
A
F
From
my
side,
where
the
original
requirement
came
from
thus
developed
from
the
face
to
face
meeting,
basically,
what
Google
does
is
they
need
to
attach?
He
went
to
their
metrics,
specifically,
they
have
traces
which
they
need
to
append
to
specific
pockets
in
their
metrics.
So
they
just
know:
I've
got
iron-on
latency
X,
and
this
is
a
trace
of
something
which
had
a
Nathan
T
of
X
or
just
what
have
you.
F
F
The
only
thing
which,
with
the
open
metrics
head
on
we
need
to
do,
is
to
make
sure
that
it
fits
the
same
data
models,
and
this
is
basically
where,
where
I'm
coming
from,
because
as
we've
seen
with
with
premises,
format,
labels
which
allow
the
user
to
define
arbitrary
n,
dimensional
matrices
and
just
slice
and
dice
the
data.
However,
they
need
it.
That's
one
of
the
main
powers
of
what
Prometheus
and
Boardman
and
all
those
things
are
having
or
doing
and
I'd
like
to
also
have
this
in
the
event
ecosystem
and
when
you
think
about
it.
F
Basically,
the
difference
between
a
metric,
and/or,
a
time
series
and
an
event
in
its
purest
form.
It's
basically
that
the
one
has
its
label
set
and
the
value
and
numeric
value,
at
least
in
promises
world,
and
it
has
an
optional
timestamp
which
you
can
set,
or
you
can
just
have
it
as
this
is
the
time
of
whatever
happened
and
then
have
it
implicitly
so
and
actually
went
is
slightly
different.
F
J
Yeah
I
wanted
to
just
add
a
voice
to
that
I
think
it
would
be
really
valuable
to
be
able
to
figure
out
whether
one
could
D
couple
those
two
without
losing
the
benefit
of
being
able
to
associate
events
with
metrics,
but
have
them
able
to
develop
in
different
and
independent
directions,
rather
than
be
tied
at
the
hip.
I.
F
Don't
think
it
needs
to
be
tied
at
the
hip,
what
what
the
one
thing
which
Google
would
need
and
which
is
not
currently
possible
with
the
premises
exposition
format
is
to
have
the
information
that
this
one
specific
trace.
Id
is
linked
to
that
particular
bucket
in
this
particular
time.
So,
basically,
what
what
it's
relatively
easy
to
do
is
to
just
define
a
label
set
for
metrics
and
a
label
set
for
for
events,
and
then
you
cross
relate
basically
by
searching
through
both
those
indices.
F
That's
relatively
easy,
but
they've
got
issues
with
scale,
so
they
want
to
do
basically
just
toss.
This
information
of
this
specific
trace
ID
is
attached
to
that
particular
point
in
time
of
that
time.
Series-
and
this
is
currently
not
possible.
So
it's
a
very
lease
which
will
need
to
accommodate
within
the
open
metric
space,
because
else
Google
standpoint
is,
they
can't
do
anything
with
the
whole
standard
and
tails
and
they
have
to
do
something
else,
which
would
not
be
the
best
of
outcomes.
F
So
this
is
the
minimum
which
we
need.
Everything
else
is
basically
on
top,
but
definition
wise
standard
twice.
I
suspect
the
work
is
already
mostly
done,
because
it's
I
don't
see.
That's
very,
very
involved,
at
least
on
on
the
format
side
for
everything
else,
we'll
see
where
it
goes,
but,
as
you
said,
it
doesn't
need
to
be,
it
doesn't
need
to
be
in
the
same
thing.
It
just
has
to
work
together
together,
really
really
well,
except
for
the
straight
IDs
stuff,
which
needs
to
come
through
open
matrix
else.
Google
will
not
support
open
matrix.
A
A
G
Just
briefly
that,
if
it's
an
extension
essentially
so,
we've
got
the
core
of,
it
is
really
focused
on
the
metrics
and
we
have
the
extension,
which
is
you
know
here
is
the
standard
that
you
would
use.
If
you
wanted
to
implement
events,
I
think
that's.
You
know,
I
think
that
makes
sense,
especially
given
that
Google
has
such
a
strong
use
case
for
it.
A
If
we
go
down
the
route
that
I
figured
Richard
was
describing,
which
I'm
not
sure,
he's
necessarily
completely
sure
about.
Yet
we
can
end
up.
You
know
pre-empting
decisions
that
that
we
get.
You
know
we
have
to
undo
later
so
I'm
very,
very
firmly
in
favor
of
having
things
as
a
loosely
coupled
as
possible
and
having
metrics
people
do
metrics
and
events.
People
do
events
and
other
things
and
look
at
the
correlation
problem
and
Association
problems,
that's
even
technically
possible,
but
that
would
be
my
preference.
A
F
D
F
That's
basically,
it
that's
the
whole
use
case
at
their
scale,
even
though
they
really
have
probably
quite
of
experience
with
search,
they
prefer
to
just
store
both
at
the
same
time
or
this
correlation
is
hard
correlation.
Instead
of
a
soft
correlation,
which
you
have
to
recreate
every
time,
you
do
an
analysis.
F
F
Matter
if
it's
a
full
trace
or
a
trace
ID
or
if
it's
a
log
line,
because
as
soon
as
I'm
able
to
put
the
single
event
into
the
format,
I
basically
already
have
a
format
which
can
support
everything
else.
We
are
not
talking
about
efficiency,
blah
blah
blah,
but
at
least
in
theory
I
can
I
can
put
as
many
events
as
I
want
through
exactly
as
mechanism,
and
people
will
abuse
this
or
they'll
just
shoehorn
in
their
own
extension.
A
Okay,
well,
the
timing.
I
think
we've
covered
the
instruction
and
thank
you
very
much
Richard
that
was
great,
so
I
think
I
can
go
back
to
the
slides
and
it's
jump
ahead
to
this
is
kind
of
preempted
by
the
previous
topic.
But
you
know
we
now
have
a
number
of
different
ideas
kicking
around
concerned
with
you
know,
metrics
events,
some
people
have
asked
about
logging
and
the
remote
out
three
or
four
other
things
where
I
think
similar
kind
of
efforts
might
be
fruitful.
A
Does
anybody
else
want
to
identify
areas
where
we
should
be
looking
at
a
specific
project
right
now,
trying
to
encourage
activities
similar
to
want
Richard
has
successfully
pulled
together
really
in
the
last
nine
months,
but
working
with
the
Prometheus
community
and
bringing
it
out
to
the
CNC
F
and
the
most
recent
face-to-face,
of
course
involved.
Others
like
in
flux,
TV
in
Google
and
we've
works,
so
one
of
what
do
people
think
about
other
values
of
asterisk
here?
If
anything.
F
So
I
don't
have
anything
in
regard
to
of
specific
projects,
but
one
thing
which
at
least
to
me
seems
to
be
missing
in
the
whole
ecosystem,
as
of
right
now
is
something
which
actually
maintains
what
the
state
of
the
whole
system.
Well,
however,
your
system
looks
should
look
like,
basically,
if
you're
familiar
with
the
term
CMDB,
something
where
I
can
put
in
the
data
and
the
country.
F
How
I
am
how
much
stuff
should
look
so
there's
EDD,
which
is
really
good
at
taking
this
information
once
it's
been
put
into
key
value
stores
and
just
giving
a
pair
sorry
and
just
pushing
it
out
to
two
machines.
But
the
thing
is:
there
is
no
central
brain
which
actually
manages
what
services
should
exist
and
how
many
of
X
I
should
have
what
IP
addresses
have
I
be
assigned?
How
many
data
centers
do
I
have
for
some
of
this
information.
F
F
A
J
One
suggestion
they
seem
to
be
in
emerging
set
of
cases
where
people
want
to
do
things
very
similar
to
what
is
happening
in
the
CNC
F
around
kinetics,
but
in
non-traditional
data
center
environments
like
edge
computing
and
IOT,
and
all
those
kind
of
things.
Now
we
do
already
have
you
know,
groups
of
those
people
in
the
Linux
Foundation,
but
I
think
there's
a
bit
of
a
vacuum
as
to
where
exactly
goofin
Eddie's
in
the
CN
CF.
What
its
approach
to
that
space
is
engine
IOT
might
be
interesting
to
just
explore
that
space.
A
little.
A
Okay
I
mean
it
feels
a
little
bit.
It's
a
half
in
half
out
because
it's
a
potentially
containerized
but
I'm
not
sure
for
affray,
represents
kind
of
dynamic
infrastructural
microservices,
for
example,
and
I.
Think
this
is
that
kind
of
thing.
It's
more
of
a
good
topic
for
dinner
at
the
CN
CF
dinner
times
in
in
Austin.
Perhaps
you
know
when
we're
all
getting
together
face
to
face
and
to
see
when
people
sort
of
feel
things
are
going.
F
A
Well,
but
yes,
I
hear
you
on
that:
okay,
so
good,
I'm,
gonna
declare
it
into
that
section.
Thank
you
very
much
Richard.
So
let
us
wrap
up
quickly.
We
have
a
few
more
similar
items
just
to
run
through
projects,
so
we've
got
the
haven't,
made
any
changes
to
the
priorities
in
the
backlog.
We
have
a
notary
vote
that
is
almost
closed.
A
I
believe
Chris,
you've
already
notified
people
on
the
list
that
you're
gonna
hold
a
vote
open
for
a
few
more
days,
yep
and
we're
landing
and
an
announcement,
and
then
not
just
my
future,
as
shown
on
the
slide
that
probably
next
week
we
have
some
working
group
updates.
Not
today,
though,
we
can
do
another
time,
then
we
have
the
dev
stats,
which
is
something
nothing.
Damn
I
like
to
spend
a
moment
telling
us
about
yeah.
D
I
think
denge
dropped
off,
but
essentially
we've
been
funding
an
effort,
prove
statistics,
collection
within
the
kubernetes
project
and
being
able
to
see
kind
of
stats
across
SIG's
within
SIG's
and
so
on.
So
we're
asking
folks
to
test
drive,
dev
stats,
cate-cate,
e8s,
io
and
and
kind
of
give
us
any
feedback.
But
it's
it's
better
than
kind
of
the
paternity
of
stuff
that
we
had
and
we've
been
working
closely
with
the
kubernetes
team
to
ensure
that
this
is
something
for
them.
So
we're
just
just
trying
to
raise
up
as
feasibility
of
this
work.
A
Go
ready,
move
on
yeah
like
this
yeah,
that's
good
move
on
and
I
think
that
just
brings
us
to
the
standard
reminders
about
the
events
coming
up.
I
think
the
main
news
item
here
is
watch
out
for
Thanksgiving.
We
have
a
meeting,
that's
moved
from
the
21st
of
November
to
the
14th,
so
we
have
two
meetings
in
November,
November,
the
7th
and
November,
the
14th
and
I.
Don't
know
you
have
an
agenda
yet
14th.
We
have
spiffy
earnest
here,
knocking
on
the
door
for
the
7th,
more
information
from
service
and
events.