►
From YouTube: 20210113 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
All
right,
hello,
everybody-
and
this
is
the
kubernetes
architecture-
production
readiness,
subproject
meeting
for
january
12th,
2021
happy
new
year,
everybody,
so
they
put
a
few
things
on
the
agenda
and
looks
like
we
have
a
new
new
new
member
here,
we're
a
small
team.
So
derek!
Do
you
want
to
introduce
yourself.
B
Yeah,
hey
I'm
derek.
I
was
kind
of
referred
over
here
from
jeremy.
He
touched
me
there.
There
might
be
some
might
be
able
to
contribute
a
little
bit
over
here.
I'm
the
bug
triage
lead
for
1.21,
as
I
shout
out
for
1.20
and
just
trying
to
get
more
involved
in
kubernetes
and
contribute
back
to
a
project
that
has
served
me
well.
C
A
Involved
here,
thank
you
great.
Well,
let's,
let's
get
started
and
if
anybody
else
has
agenda
items
please
you
know
bring
them
up
the
end.
But
the
first
thing
I
wanted
to
talk
about
was
the
discussion
yesterday
in
the
leads
meeting
and
we'll
have
a
similar
discussion
tomorrow,
david,
you
were
there
but
which
couldn't
make
it
and
derek
and
he
went
there
so
derek.
For
your
background.
A
We
enabled
enforcing
production
readiness
reviews
to
target
the
release
back
in
december,
but
there
was
some
objection
because
we
did
the
latest
consensus
period
in
december
and
a
lot
of
people
were
out
and
they
so
they
didn't
see
it.
So
we
we
reduced,
that's
fine
david,
so
we
reduced
the
rather
we
disabled,
the
enforcement
of
it-
and
I
didn't
get
to
it
yesterday,
but
I'll
I'm
gonna
create
a
pr.
A
So,
in
any
case,
there
was
concern
over
the
the
reviews,
delaying
the
kept
process
and
keeping
caps
out.
I
think
that
we
can
explain,
and
maybe
we
can
add
some
more
text.
A
Basically,
yes,
we
did
whitehead,
we
reviewed
many
of
these
and
and
the
reviews
aren't
that
you
know
intense-
it's
not
like
you're
reading
code,
but
in
any
case,
what
we
want
to
make
clear
a
few
more
things
we
want
to
clarify
in
the
documents
are
that
this
is
that
enhancement
freeze
time
this
approval
needs
to
be
needs
to
be
had,
which
means
we
need
to
be
pulled
in
you
know
more
than
the
day
of,
and
the
plan
right
now
is
put
a
pr
out,
that's
going
to
re-enable
the
enforcement
and
do
a
lazy
consensus
on
that.
A
D
It's
sort
of
arbitrary,
usually
for
short
things.
It's
three
days
for
longer
things.
It's
a
week,
okay,
like
sort
of
ballparkish.
A
I'm
gonna
say
I'm
gonna
put
out
the
pr
and
I'm
gonna,
I'm
gonna
start
the
lazy
consensus
period
after
the
meeting
tomorrow
for
three
days,
because
you
already
did
a
seven
day.
One
right
I
mean
I'm
sorry
if
you
weren't
around
but
like
you
got
through,
I
guess
it's
thursday,
it's
three
days
good
enough.
We
can
give.
We
can
give
it
to
until
monday
or
tuesday.
Maybe
yeah
okay,
if
that,
if
you
think
that
I
don't
want
to
upload
people,
I
just
want
to
make
sure
that
we
don't
just
lead
away.
E
Now
there
were
some
specific
action
items
that
were
requested
of
us
right
and
I
think
it
makes
sense
in
that
pr
to
clarify
those
points
as
well.
A
E
Sure
there
was
there
was
a
request
that
we
explain
why
it's
valuable
to
have
another
person.
Look
at
these
fields.
I
did
so
verbally
on
the
call,
but
I
think
formalizing.
It
is
fine.
There
was
a
request
to
explain
why
it
was
valuable
for
someone
who
wasn't
didn't
have
detailed
knowledge
of
a
particular
area
to
review
which
had
a
slightly
different
answer,
and
I
also
explained
that
on
a
call,
but
I
think
we
can
explain
that
as
well.
E
Right.
Cluster
admins
aren't
knowledgeable
in
every
aspect
of
the
product,
so
it's
actually
good
to
have
someone
who's
familiar
with
the
project
as
a
whole,
but
not
every
individual
piece.
And
then
there
was
a
request
to
explain
how
people
get
added
to
this
list
and
I
think
we'd
be
happy
to
see
people
added
to
this
list.
E
But
I
think
we
need
to
be
able
to
describe
the
prerequisites
of
a
a
fairly
deep
knowledge
of
cube
at
sort
of
the
general
level.
Not
a
networking
expert,
not
a
storage
expert,
but
someone
who
has
a
working
knowledge
of
the
entire
system
and
an
sre
would
be
perfect
to
really
review
these
questions.
A
So
that's.
My
next
agenda
item
is
to
talk
about
that
ladder,
some
more
so,
let's
we
can
talk
about
that
and
then
get
back
to
it.
So
were
there
any
other
things
we
we
we
wanted
to
address.
Oh
the
bottom
lack
bottleneck.
How
is
it
not
a
bottleneck?
Well.
E
The
questions
are
very
crisp.
What
we're
looking
for
ahead
of
time
over
time,
we've
improved
those
questions,
so
the
answers
have
gotten
a
lot
better
than
the
first
time
we
started
looking
at
them.
I
think
one
at
least
for
my
part,
119
was
pretty
rough,
but
in
120
the
answers
to
the
questions
were
really
getting
it.
F
A
F
D
Oh,
I
was
just
asking:
how
do
you
think
the
pr
review
went
last
time,
especially
with
the
people
that
would
try
and
get
the
enhancements
in
at,
like
the
last
possible
moment?.
E
So
I
think
it
went
pretty
well,
although
I
did
discover
after
the
fact
that
the
pr
load
was
not
very
well
balanced,
wojtek,
I
think
had
as
many
as.
G
E
And
I
combined,
but
about
a
week
or
two
before
the
freeze,
we
started
running
this
tool
that
I
think
was
the
enhancement
sub
project
built
that
allowed
us
to
query
metadata
and
say
hey.
These
are
the
ones
that
are
assigned
to
me.
These
are
the
ones
that
are
targeted
and
need
review
and
we
were
able
to
go,
and
you
know
open
them
up
see
what
was
there
find
the
pr's
that
were
present
and
move
from
there?
E
Usually,
I
would
say
about
half
of
them
all
the
questions,
all
the
answers
that
I
looked
at.
They
were
all
clear
at
the
time
I
read
it
and
then
the
others
it
was
back
and
forth,
but
still
fairly
minor.
Does
that
match
what
you
guys
had
as
well.
E
Cool,
thank
you.
I
was
just
sort
of
curious
yeah,
the
the
first
time
we
tried.
We
tried
a
dry
run
of
this
in
119
and
at
least
my
experience
in
119
is
that
our
questions
weren't
weren't
crisp
enough
to
get
us
the
answers
that
we
wanted.
So
we
went
back
and
we
updated
the
questions
with
example,
answers
in
many
cases
to
help
and
explained
better
why
we
wanted
the
question
answered.
F
Okay,
I
have
one
of
my
to
do
list
like
to
add
some
sample
answers
to
two
more
questions,
so
I
will
try
to
have
it
done
later
this
week
or
early
next
week,
because
I
think
there
are
common
patterns
that
I
see
that
is,
I'm
expecting
a
little
bit
different
answer
that
I'm
getting
so
it,
and
this
is
a
common
pattern.
So
I
I
I'm
going
to
open
a
pr
with
the
sample
answer.
D
E
E
C
E
D
C
So
maybe
what
what
this
is
growing
in
scope
here?
Hopefully
today
I
can
get
this
done.
I'll
get
a
draft
of
this.
The
pr
will
one
re-enable,
the
the.
A
Enforcement,
which
is
like
two
lines
and
then
I
think
they
might
be
different
repo,
we'll
see
if
it's
all
the
same
repo,
then
it'll
also
add
some
more
documentation
like
I've
taken
in
the
notes
here
and
that
ideally
like
david.
If
you
could
get
me
that
link
to
the
metric
you're
adding
we
could
even
have
a
sort
of
commonly
missing
items.
You
know
you
know
metrics,
maybe
there's.
No,
that's
the
only
entry
right
now
but
like
we
can.
We
can
include
a
link,
then
how
did
it
out
how
to
add
a
metric?
A
I
think
the
thing
I've
seen
also
is
often
a
a
discussion
about
when
they
should
have
a
metric
versus
an
event
versus
a
log
and
who
those
are
geared
towards
like.
A
E
E
G
C
D
Sorry
go
ahead.
They
drafted
two
schedules,
one
based
on
like
idea
how
long
121
was
going
to
be
if
it
was
going
to
be
a
quarter.
You
know
four
releases
a
year
or
three
releases
a
year.
It
looks
like
the
current
one
scheduled
one
is
set
for
february
9th.
B
C
Okay,
then,
the
latter.
A
Yeah
like,
if
you
think
about
how
the
api
review
team
works
right,
I
mean
somebody
comes
in
and
they
do.
They
watch
somebody
doing
api
reviews
for
a
while.
Then
they
do
api
reviews,
but
you
know
it
takes
a
while
to
become
somebody
who
can
actually
approve
the
apis
as
opposed
to
just
sort
of
doing
an
initial
pass.
Do
we
want
that
kind
of
structure
we're
such
a
small
team?
So
I
don't
know.
E
So
for
for
code
and
stuff
you
end
up
seeing
other
people's
prs
and
the
same
name
pops
up
enough
right.
So
you
notice
like
okay,
this
guy
knows
what
he's
doing
and
he's
contributed
a
lot
of
code
in
this
area
he
comes
and
he
asks
me.
I
want
to
be
a
reviewer
or
approver.
Then
we
go
through
the
like.
You
can
be
a
reviewer,
let's
start
there,
and
then
we
get
to
be
an
improvement.
E
A
E
Actually,
I'm
curious,
so
derek
you're
you're
a
new
face
here.
What
is
your
view
on
what
the
goal
of
the
pr
section
of
a
cap
is
for.
B
So
I
haven't
had
a
chance
to
go.
Look
at
that
documentation
yet,
but
without
that
background
knowledge,
I
guess
I
would
say
that
I'd
be
looking
for
that.
The
code
is
not
gonna
break
things.
Unexpectedly,
it's
gonna
have
same
defaults,
you
know
and
as
a
somebody
running
clusters,
it's
going
to
kind
of
back
to
the
same
defaults
work
as
expected
right.
B
E
Okay,
so
a
lot
of
the
focus
on
the
question.
That's
very
close,
a
lot
of
the
focus
on
the
questions
that
we
have
is
around.
How
do
we
know
that
people
are
using
it?
How
do
we
know
that
it
is
running?
How
do
we
know
that
it
is
broken
when
it
breaks?
Can
we
turn
it
off
after
we
turn
it
off?
Can
we
turn
it
back
on
and
does
it
leave
my
cluster
in
a
clean
state?
E
B
Kind
of
okay,
so
we
run
like
multi-tenancy
clusters
at
vmware
and
in
the
past,
I've
ran
dedicated
clusters
for
as
a
devops
engineer,
and
I
mean
a
lot
of
deploying
clusters
and
managing
workloads
on
them
and
the
main
things
that
I
want
out
of
that
are
that
the
different
components
are
working
right
and
that
if
there
is
an
issue,
I'm
able
to
look
at
logs
and
get
metrics
on
it
and
see
what's
wrong.
E
B
E
Review
is
supposed
to
be
a
spot
so
that,
if,
if
david,
who
is
looking
at
a
fleet
of
clusters,
looks
and
says
something
something
looks
squirrely
or
I'm
getting
reports
of
something
weird.
How
can
I
confirm
what
that
problem
may
be,
and
how
can
I
recover
my.
E
A
Yeah,
how
do
I
kind
of
yeah?
How
do
I
determine
because
we're
talking
kubernetes
development,
we're
focusing
so
it's
for
a
given
kubernetes
feature
like
dance?
How
can
I
tell
people
are
using
it
and
if
they
are
using
it,
how
can
I
tell
that
it's
working
properly
not
broken
if
it
is
broken?
B
So,
for
for,
like
answering
the
broken
part,
I
mean:
are
you
thinking
like
in
a
test
testing
capacity?
No.
A
A
A
E
Yeah,
so
we
need
people
who
have
experience
doing
that
right
like
and
when,
let's
see
we
have
the
results
of
that
survey
right
where,
basically,
we
found
that
if
you
run
enough
clusters,
everything
has
failed
at
some
point.
I
don't
know
if
you
recall.
E
Yeah
that
was
kind
of
humorous.
The
way
that
result
came
in,
but
we
we
would.
I
think
my
preference
would
be
to
get
people
who
have
experience
dealing
with
the
large
numbers
of
clusters
either
they
directly
maintain
or
that
they
field
reports
from
right.
So
there's
there
are
on-prem
distributions
as
well,
and
not
all
of
them
have
you
know,
aggregate
alerting,
some
of
them
do
or
aggregated
metrics.
Some
of
them
do,
but
not
all
of
them.
A
A
Something's
failing
and
you
need
to
to
know,
you
know
how
to
stop
the
the
bleeding
and
that's
where
hopefully
hoping
that
the
pr
answers
to
the
pr
questions
provide
that
guidance.
E
A
I
think
the
only
way
the
only
way
we
can
really
do
it
is
the
value
is,
is
have
people
have
a
trial
period
where
they
evaluate
x,
prrs
xprs,
and
if
they
look
like
they
have
the
right
background.
You
know
I
mean
there.
Yes,
there's
some.
A
A
But
and
then
you
can
propose
yourself
to
be
on
the
on
the
list
right
and
from
there.
We.
The
vote
is
that
they
need
to
get
somebody
who
has
permissions
on
that.
Whatever
the
owner's
file
is
for
that.
That
list.
B
And
that
sounds
fair
to
me
as
someone
from
the
outside,
not
having
a
ton
of
background
on
the
specifics
of
the
processes,
but
I
mean
that
sounds
insane.
I
mean
it's
a
it's
hard
to
hire
for
sre
right,
it's
kind
of
the
same
similar
boat.
There
might
look
good
on
paper,
but
they
might
not
have
the
real
practical
experience
or
mindset,
or
maybe
they
don't
look
good
on
paper,
but
they
do.
A
A
G
B
A
And
then
we
just
make
that
decision.
So
is
that
if
that.
D
I
would
say,
like
you,
don't
necessarily
have
to
pick
a
number
right
now.
Either
you
can
sort
of
you
know,
feel
it
out.
One
other
idea
like
if
you,
if
there's
other
people
that
want
to
be
pr
reviewers
or
other
people
that
you
you
know
have
in
mind.
You
know
it
can
be
a
little
fuzzy
right
now
and
sort
of
the
first
group
of
people
to
graduate
can
probably
flesh
out
some
of
the
needed
skills,
or
you
know
needed
background
more
than
what
you
know.
Y'all
might
know
as
sort
of
seasoned
prr.
A
G
Black
and
after
showing.
D
A
So
I'll
look
at
we,
we
can
probably
just
pick
any
pick
any
pr
that
comes
up
in
our
tooling,
where
others
of
who's
assigned
to
and
grab
it
and
just
say
in
their
comment
in
the
review.
Just
have
a
comment
that
says:
I'm
you
know
I'm
shadowing
the
pr
process-
and
this
is
my
review
just
so
that
those
of
us
who
are
actually
doing
the
approvals,
let's
start
to
recognize
their
names.
E
Yeah,
that's
good
recognition.
Is
it's
gonna
be
helpful
for
me.
A
Okay,
cool,
I
think
that's
good
enough
for
for
now
we
can
refine
it
over
a
course
of
releases.
The
only
other
thing
I
wanted
to
remind
people.
C
Was
okay?
We
have
enhancement,
freeze
coming
up
now
in
a
month
or
so,
and
so
we've
got
to
get
especially
we're
going
to.
F
F
B
Sorry,
how
would
somebody
apply
to
be
a
shadow?
Do
we
want
to
document
that
too,
or.
E
I
would
accept
you
just
running
the
tool.
Actually,
we
definitely
need
to
make
sure
we
get
the
the
link
for
that
the
tool
and
like
the
suggested
command
for
people
to
run
but
yeah
like
run
the
tool
pick
one
out,
do
a
review
on
it
and
then
bug
it.
Bug
us
in
the
channel
and
yeah
you
do
that
a
dozen
times
we
will
be
both
grateful
and
recognize
you.
G
C
A
Let's
yeah,
let's
get
start
getting
some
reviews,
then
I'll
I'll
try
to
get
this
pr
all
this
written
out
today,
although
it's
it's
like,
I
said,
expanded
and
scoped
so
but
I'll
get
it
as
quickly
as
I
can.