►
From YouTube: 20220713 SIG Arch Prod Readiness
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
B
A
Everybody,
this
is
the
kubernetes
sig
architecture,
production
readiness,
sub
project
meeting
for
july,
13th
2022.,
yes,
so
david,
you
were
asking
about
how
the
end
of
well
it
was
fun.
Let
me
say
that
it
got
a
little
intense,
but
it
was
okay.
It
I
happened
to
have
several
other
deadlines
at
the
same
time,
which
was
awesome
so
it
made
for
some
late
nights,
but
we
managed
to
make
everything
you
know
part
of
the.
A
Got
through,
I
didn't
have
to
reject
anybody,
but
we
did
make
fixes
and
things
and
improvements
to
the
caps
and
the
plans
so
yeah
I
mean
it
was
shitty
for
me,
sorry
for
the
language
and
the
recording,
but
you
know
it
was.
It
was
survivable
and,
and
we
made
it
and
I
think
the
biggest
problem
was
there
were
issues
with
the
spreadsheet
that
there
weren't
proper
permissions
on
it
and
so
like
a
whole.
A
B
A
A
Zoycheck
too
right
he
was
he
was.
He
was
here
until
the
original
deadline,
and
then
they
shifted
it.
B
I
will
not
admit
to
planning
my
vacation
around.
That
would
just
be
a
sad
state
of
affairs,
but
but
I
think
then
that
means
next
release
and
the
release
after
we're
probably
going
to
have
to
look
and
re-subdivide
those
enhancements
that
you
had
to
review,
because
there
was
nobody
else
and
rebalance
them
as
opposed
to.
A
Out
you
guys
have
done
most
of
the
like
you
had.
I
had
done
nothing
before
then
so,
actually,
it's
not
that
it.
You
know
it's
just
sort
of
like
I
crammed
in
most
of
my
work,
plus
some
extra
in
the
last,
like
you
know
three
four
days
that
it
was
more
that
than
it
was
like
you,
I
hadn't
done
much.
I
had
kind
of
left
it
all
to
you
guys
for
the
first
few
weeks
so,
but
I
I'm
hoping
ben
elder.
I
don't
know
if
you
know
ben
david.
I
know
check
knows
him.
A
Yeah
yeah,
he's
interested
in
in
getting
involved
here
too
so
he's
on
vacation
right
now,
but
when
he
gets
back
we're
gonna
sit
down
and
chat,
and
so
he's
you
know
he's
we
can
start,
and
he
also
you
know
he
tested.
We
also
did
kind
right.
So
he's
very
he's
got
a
lot
of
breadth
and
kubernetes,
obviously
kind
it's
not
production
clusters,
but
it
you
know
it
gives
you
a
lot
of.
It
gives
him
a
lot
of
breadth
around
and
depth
around
kubernetes
in
general.
A
So
that's
looking
pretty
good
so
and
I
think
we
should
try
to
broaden
the
pool
a
little
bit
too
so
that
we
don't
end
up.
You
know
if
all
three
of
us
were
on
vacation,
we
would
have
had
a
real
problem,
and
you
know
redundancy
of
three
is
okay
for
machines,
but
maybe
not
as
good
for
people.
B
Sure
sure,
especially
if
those
people
might
be
aware
of
a
schedule
ahead
of
time
and
and
promise
their
wives.
B
Oh,
oh,
I
I
timed
mine
later
about
about
the
time
the
cube
schedule
came
out
so
anyway
that
sound.
That
sounds
good.
We
often
go
through
and
do
a
retro
of
how
we
thought
prr
went.
Do
we
want
to
do
that
briefly
right
here.
A
B
Sounded
like
you
said,
responses
were
good
for
your
comments.
I
found
that
to
be
the
case
as
well.
I
found
people
very
willing
this
release
to
consider
how
people
were
able
to
debug
their
features
at
scale,
and
I
think
that
you
know
it's
continuing
on
what
we've
seen
in
the
past,
but
I
was
very
pleased,
but.
C
A
Yeah
exactly
it's
encouraging
that
people
are,
and
that
means
this
thing
leads
are
probably
taking
it
more
seriously
right,
and
so
I
I
had
the
same
experience
the
things
and
and
where
they
weren't
the
people
were
responsible
like
where
the
where
they
didn't
it
was
more
like
they
didn't.
They
didn't
get
what
they
needed
to
do
as
opposed
to
they
were,
you
know,
resistant
or
something
to
doing
it.
So
I
didn't
have
any
trouble
at
all.
Folks
were
cooperative.
I
did.
A
No,
no
just
to
say
there.
There
is
some
friction.
I
think
around.
You
know
just
that
there
is
additional
process
that
there
is
an
additional
gait
and
that
for
certain
caps,
that
just
don't
necessarily
even
have
anything
to
do
with
pr
the
same
things
we've
seen
in
past
in
past
cycles,
where
there's
a
set
of
caps
that
we
don't
really
need
to
review,
but
because
of
the
tooling,
we
need
to
review
and
we
need
to
approve
anyway
and
that
still
causes
some
friction.
C
Yeah
one
thing
that
I
agree
with
that,
but
one
thing
that
I
wanted
to
also
add
is
that
I
think
we
we
get
to
the
point
where,
if
I'm
asking
for
something
the
response
is
not
do
we
really
need
that
like
it
is,
is
it
so
important,
but
really
people
caring
about
things
and
like
agreeing
that,
like
reliability
and
ensuring
that
we
can
monitor
that
and
like
we
know
the
scale
character,
understand
the
scale
characteristic,
and
so
on
that
those
are
important
things
and
they
are
just
like
acknowledging
the
fact
that
they
they
need.
C
B
This
release,
I
do
not
have
any
suggestions
for
improving
the
cap,
template,
which
I
think
is
is
new
for
me.
I'm
I'm
happy
to
say
that
I
don't
know
about
the
other
two
of
you
do
we
do
you
see
improvements,
we
need.
A
I
don't
see
specific
improvements.
We
need
that
I
can
think
of
in
the
prr
section.
I
think
there
is
some
we've
certainly
had
some
discussion
here
and
probably
I
should
have
gone
to
the
leads
meeting
and
talked
to
them
about
it,
but
I
just
didn't
have
time
this
week,
there's
been
like,
I
said
some
thoughts
around
like.
Should
we
separate
out?
A
A
Those
aren't
that
common,
there's,
an
intimidation
factor,
there's
a
there's,
a
people
come
and
they're
like,
oh,
my
god,
I
have
to
write
a
cap,
my
world's
gonna
end
right
and
the
longer
the
template
is
not
the
production
readiness
section
per
se,
but
just
the
template
in
general.
It
keeps
growing
and-
and
it
keeps
creating
some
perception
of
or
some
intimidation
factor,
and
so
you
know
there's
a
discussion
of
how
we
might
reduce
that
such
that
you
know
for
for
caps
that
prr
doesn't
apply.
A
Can
we
have
it
in
a
separate
file
so
that
they
don't
need
to
go
through
this
process?
But
I
don't.
I
don't
know
that
we
need
to
talk
about
it
more
here
beyond
beyond
me.
Just
saying
that
that's
a
that's
in
the
air
put
it
that
way.
C
It
seems
that,
like
people,
I
I
think
I
didn't
have
any
case
where
the
requests
that
I
had
for
people
to
change
some
things
were
because
they
didn't
understand
it's
like,
which
is
what
are
we,
which
is
which
used
to
be
the
case
in
the
past,
and
we
were
like
updating
the
template
to
qualify
clarify.
I
think
I
didn't
have
any
of
those
like
all.
The
requests
were
really
really
think
about
the
metric
a
little
bit
more
or
like
update
this
one,
because
I
like
it,
it's
not
really
the
case
and
so
on.
C
So
no,
I
think
I
don't
have
any
any
anything
that
I
I
so
or
at
least
I
didn't
note
anything
to
myself
that
we
should
change
something.
A
A
We
got
a
total
of
94
responses
to
our
2022
survey,
so
we
can
assume.
A
A
Basically,
if
you
know
getting
tim
hawkin,
to
put
it
on,
twitter
makes
a
big
difference,
and
that
was
the
difference
between
the
first
and
second
year
and
then
this.
B
A
Did
again,
but
I
probably
should
have
made
him,
do
it
more
so,
but
we
can
take
a
look
at
it.
Real,
quick
if
you
want,
I
just
don't
want
to.
I
want
to
be
careful
not
to
accidentally
show
people's
email
addresses,
which
I
did
the
first
year,
and
then
I
had
to
go
and
like
edit,
the
video
which
is,
like
you
know,
a
nightmare
just
because
I
don't
that's
not
something.
I
do.
Okay,
it's
just
the
very
last
question.
B
Yeah
alana,
like
cleverly,
took
her
sabbatical,
but
you
know
I
see
I
see
akil.
Hopefully
I
pronounced
that
right.
You
know
I'll
teach
him
to
show
up
at
a
meeting,
but
but
otherwise,
if
wojtek
can't
do
it,
I
guess
I'm
willing
to
go
again
next
year
in
a
row.
A
All
right:
well:
okay,
yeah,
well,
okay.
That
sounds
good
well
happy
to
have
you
join
and
we're
a
small
group
here,
but
we
try
to
be
friendly,
so
yeah
david.
I
I
definitely
am
not
not
going
to
have
time.
Maybe
we
make
ben
do
it
it's
a
trail
by
fire.
I
could
I'm
sure
he
doesn't
have
time
either,
but
he's
trying
to
make
it.
B
We
can
discuss,
we
can
discuss
after
the
meeting,
but
if
there
is
nobody
else
I
will.
I
will
go
through
it
round.
Two.
A
A
Oh,
I
see
that's
probably
too
much
detail,
but
okay
yeah.
So
those
that's
good.
I
don't
remember
again
when
we
do
the
real
analysis.
We
want
to
look
at
kind
of
what
it
was
last
year,
but
one
to
ten
clusters,
mostly
some
people
have
more
than
that
and
then
there's
just
a
couple
of
people.
This
is
these.
Are
the
you
know,
not
many
people
with
more
than
a
thousand
clusters
all
right
nodes
under
management?
A
B
A
B
A
Exactly
yeah
exactly
because
that
yeah
we'll
just
skip
that
for
now
all
right.
A
Most
people
on
have
121.
A
This
will
be
interesting
because
we
had
a
bunch
of
api
removals,
so
we'll
see
how
quickly
128
gets
adopted
our
next
year,
we'll
see
that's.
B
Last
year
we
had
one
big
spike
at
119
that
was
not
quite
as
high
and
then
on
either
side
we
had
118
was
almost
the
same
and
120
was
almost
the
same
and
this
year
it
is
a
significant
difference
between
121
and
122..
I
wonder
why.
C
B
A
A
All
right,
I
don't
know
david
if
you
have
the
other
one
up.
If
you
can
say
how
we
did,
I.
B
Do
so
comparing
it
in
production
is
nearly
identical
28.2
before
in
staging
it
used
to
be
43.6
about
the
same
in
development,
that's
dropped
quite
a
bit.
I
wonder
why
it's
dropped
quite
a
bit.
It
was
79.5
percent
last
year
and
only
67.9
this
year.
I
guess
that
would
feed
into
the
does.
It
feel
more
reliable.
A
B
It
was,
although
features
did
not
work
as
expected
was
20.
Last
year,
it's
only
8.7
this
year,
even
with
small
numbers.
We
would
expect
more
than
that.
B
B
Hang
on,
let
me
scroll
ahead
of
you
here:
yeah,
okay,
you're,
you're,
still
good,
just
making
sure
you
didn't
get
into.
A
A
B
But
previously
was
119
was
the
most
popular
and
the
rollback
was
about
the
same
and
then
120,
which
would
be
about
the
same
as
123
here
had
11
roll
back
three
people.
B
A
A
B
A
34
percent
of
them
did
it
in
production,
so
yeah
we
have
to
get
the
the
analysis.
Actually
I
wonder
if
we
did
that
properly
last
time
and
the
time
before,
like
did
we
do
we
take
the
right
denominator?
A
A
A
I
think
so
why
would
like
that
seems
like
an
unlikely
thing
to
cause
a
problem,
but
anyway
again
we
if
we
have
the
energy,
we
should
follow
up
with
these
folks,
trying
to
understand
it
or
or
send
it
to
the
node
team
and
say
hey:
can
you
follow
up
with
these
folks?
B
Actually,
you're
not
gonna,
believe
it,
but
it
was
10.3
percent.
Last
year.
B
More
people
were
yeah
more
are
now,
that's
the
opposite
of
what
I
would
expect.
A
B
I'm
scrolling
down
I'd
like
to
see,
there's
a
question:
what
tools
do
you
use
to
monitor
your
cluster?
Do
we
still
have
that
one
yeah
yeah
see
the
results
on
that
and
do
not
scroll
past
this
question.
Okay,.
A
Prometheus
80
wow,
I
guess
that's
winning
the
matrix.
What
else
is
there
cloud
provider
metrics
tools.
B
B
Persistently
high
and
now
it's
finally
dropping
okay,
I
mean.
A
B
A
Okay,
well,
that's
that's
a
that's
a
little
preview
for
folks,
but
we'll
we'll
work
on
the
the
official
report
and
try
and
put
something
together
for
some
time
in
the
next
month
or
two
realistically.
B
A
A
To
you
somehow,
so
we
have
to
remember
how
to
do
that.