►
From YouTube: Kubernetes SIG Instrumentation 20191031
Description
Meeting notes: https://docs.google.com/document/d/17emKiwJeqfrCsv0NZ2FtyDbenXGtTNCsDEiLbPa7x7Y/edit
A
Okay,
hello,
everyone:
this
is
the
sick,
instrumentation
meeting
happy
Halloween,
it's
October
31st.
We
actually
only
have
a
relatively
short
agenda
today,
which
is
basically
reiterating
everything
that
we
are
doing
for
one
that
17.
If
every
what
anyone
has
any
other
items
they
would
like
to
discuss,
you
can
still
put
them
on
the
agenda
so.
A
His
handle
is
rainbow
mango
on
github
he's
working
through
a
bunch
of
the
work
for
the
stability
metric
stability
framework
to
allow
switching
deprecated
metrics
that
would
technically
by
the
rules
of
the
metrics
stability
framework,
already
be
hidden
by
default
to
be
able
to
enable
those
again
and
yeah.
These
are
the
poll
requests
related
to
that
any
comments
that
people
want
to.
A
B
The
only
one
that
I've
seen
is
the
API
server
one.
There
was
a
bunch
of
discussion
on
there
that
happened.
That,
like
I,
mean
I
was
a
little
I
guess
concerned.
Just
that,
like
design
discussion,
what's
happening
in
a
PRI
supposed
to
Anika
I
think
that
we
need
some
changes
happening.
One
of
the
things
that
came
out
of
the
discussion
there
was
the
suggestion
that
the
show
hidden,
metrics
flag
takes
a
parameter
of
the
like
latest
major
version
of
kubernetes.
B
B
A
So
I
think
all
of
those
are
everything
you
said
makes
sense,
but
I
do
agree
that
they
should
be
separate
caps.
I,
don't
feel
like
they
necessarily
belong
in
the
metric
stability
framework
per
se.
Specifically,
the
the
later
one
I
think
is
like
it's
a
great
feature.
I
think
I
would
love
to
have
that,
but
it
doesn't
seem
like
it
needs
to
be
a
part
of
this
effort.
I
think.
B
C
B
Idea
being
you
want
to
show
hidden
metrics
in
the
116
release?
So
if
you
accidentally,
like
you
know
you
upgrade
to
117,
you
don't
change
any
of
your
command
line.
Flags
and
you
don't
like
I,
guess,
keep
seeing
hidden
metrics
that
you
don't
want
to
see
your
deprecated
metrics
that
you
don't
want
to
see
there
was.
There
was
a
concern
raised
by
someone
from
one
of
the
other
things
about
that,
and
so
that
I
I'm
the
question
of
like.
B
Should
we
be
updating
a
cap
to
reflect
that
there's
also
been
some
back-and-forth
as
to
whether
the
showing
hidden
metrics
should
be
a
feature
flag
or
a
CLI
flag,
because
currently,
like
its
it's,
just
a
binary,
on/off
thing
for
the
entire
binary.
But
the
keps
says
that
you
should
be
able
to
specify
specific
metrics
that
you
want
to
show
or
hide
it's
not
an
all-or-nothing
thing,
so
I'm
not
sure
if
we
should
be
doing
it
as
a
feature
flag
for
the
sort
of
like
interim
case,
because
it's
an
all-or-nothing
thing
right
now.
A
A
Ok,
so
yeah
I,
don't
know
my
general
feeling
is
that
I
feel
similar
to
what
you
said.
I
feel
like
we,
a
bunch
of
new
things
came
up
that
we
haven't
thought
of
or
haven't
thought
through
entirely
previously.
I
do
feel
that
we
need
to
discuss
these
things
more
inform
of
the
cap
to
make
sure
that
we're
all
aligned
on
what
we're
doing.
B
Yeah
I
mean
currently,
this
stuff
is
showing
as
work-in-progress
and
the
thing
is
I
actually
liked
when
I
was
looking
into
this
I
picked.
This
up,
like
I,
was
looking
to
do
a
very
similar
thing
for
the
116
release
with
a
feature
flag,
I
sort
of
ran
out
of
time.
But
the
thing
is
we
currently
don't
have
any
metrics
marked
as
like
ready
for
deprecation.
So
there's
no
like
strict
need
to
have
this
flag
in
yet
I.
Don't
think.
There's
a
rush.
I
mean
that
was
the
conclusion
of
the
116
release.
B
I,
don't
know
if
anything's
changed
for
the
117
release,
but
given
that
these
discussions
are
coming
up
like
this,
PR
is
still
in
draft
state.
I
know
that
I
don't
know
his
name,
but
rainbow
mango
was
talking
about
getting
this
in,
for
the
116
are
sorry
to
117
release,
but
what
were
passed
enhancements
reason.
This
is
an
enhancement,
so
I'm
not
really
sure
I
think
that
Han
is
under
the
impression
that
we
can
continue
with
development
sort
of
targeted
for
the
next
release
without
like
having
to
get
it
into
117.
B
So
we
can
continue
to
have
these
discussions,
but
I
am
under
the
impression
that
rainbow
mango
may
be
trying
to
get
the
stuff
in
for
117
I.
Don't
know
that
that
makes
sense
without
a
kept
update.
So
I
just
wanted
to
see
where
everyone's
ad
on
that,
and
maybe
what
next
steps
we
should
take
if
I
mean
I,
think
that
the
caps
need
to
be
updated
before
we
before
we
proceed.
But.
A
B
A
A
A
A
Track
essentially
for
the
contributor
summit,
where
there
were
the
chairs
and
whiteboards
and
stuff
like
that,
set
up
everywhere.
So
what
I
was
recommended
was
that
we
designate
a
couple
of
people
who,
in
the
morning
basically
squat
a
couple
of
seats
and
the
whiteboard
at
the
contributor
summit,
and
then
we
gather
there
and
do
the
discussions
that
we
wanted
to
do
in
person.
A
I
mean
I'm
going
to
be
there
anyways
and
I.
Guess
a
bunch
of
my
team
as
well.
So
I
think
we
could
take
that
on
and
then
maybe
we
can
communicate
via
slack
where
we
are
but
I
guess
the
the
venue
for
the
contributor
summit
isn't
actually
that's
huge,
so
I
would
expect
that
we
should
be
able
to
find
each
other
anyways.
So
I
think
that
sounds
good
yeah.
A
B
B
A
A
No
specific
timing
for
anything,
it's
more,
it's
pretty
freeform,
so
I
would
suggest
that
we
like
try
to
meet
after
lunch
or
something
like
that.
That's
what
I
have
been
suggested,
but
I'm
totally
open
to
anything.
We
can
also
just
do
it
in
the
morning
and
then,
let's
trade
on
in
the
morning,
I
will
be
jetlag.
Okay,
I
guess.
A
lot
of
us
at
least
folks
in
this
hall
I
think
are
largely
from
Europe
I'm,
not
mistaken.
Yeah.
B
A
A
B
A
B
Yeah
I
somebody
had
already
quit
on
the
agenda.
I
might
have
even
been
me
for
that
meeting.
Skip
in
preparation
for
cube
con,
a
question
marks
I'll,
maybe
remove
the
question.
Mark
and
market
is
cancelled
and
then
I'll
put
an
agenda,
maybe
on
here
for
our
in-person
meeting,
which
will
be
on
the
18th
yeah.