►
From YouTube: Kubernetes Community 1 4 Retrospective 20161007
Description
We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.
https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY
This is the Special Edition 1.4 Retrospective --
https://docs.google.com/document/d/1R4zqrxgg8R2DyE_ozNx3WD07Uiil8cMiMkAqXy5aDnc/edit
A
B
So
one
thing
is
that
I'd
really
like
to
keep
this
to
an
hour
free
can
if
we
need
to
schedule
another
hour,
we'll
do
that.
We
had
to
do
that
last
time,
but
let's
shoot
for
an
hour.
So
I'll
be
brief.
Just
to
give
an
introduction,
I
like
to
start
out
these
meetings
with
a
quick
set
of
working
agreements
just
so
that
we're
all
on
the
same
page
about
what
behavior
we
expect
from
ourselves
and
others
participating.
B
Please
try
and
make
any
suggestions
you
write
below
in
the
deltas
or
things
that
could
have
been
better
sections
in
terms
of
what
could
be
improved
so
make
them
really
positive
I.
We
want
to
keep
respectful
communication
when
we're
talking
to
each
other
or
pointing
things
out
again.
We
don't
want
to
blame
and
we
really
need
to
respect
it
in
a
community
as
complex
and
interwoven
as
we
are.
The
problems
that
come
up
are
usually
system
problems
and
not
people
problems.
B
So,
let's
try
and
focus
on
making
the
system
better
and
not
focusing
on
the
people
at
hand.
Lastly,
please
respect
the
time
boxes
in
each
section.
I
have
some
time
allocated
and
I'm
gonna
be
keeping
a
stopwatch
and
I'm
going
to
be
relatively
ruthless
about
maintaining
our
time
frame.
Otherwise
this
will
go
on
for
too
long.
B
So
I
went
ahead
and
pulled
to
the
things
from
the
last
retrospective
1.32
look
at
what
we
said,
we're
going
to
do
as
a
community
and
how
we
did
on
that
I
sort
of
roughly
coded
these
as
red,
yellow,
green
there's.
A
couple
of
these
that
I
didn't
have
visibility
of
and
if
you
want
to
put
colors
on
those
and
sarah,
you
might
be
best
suited
to
do
that
since
you've
been
along
for
every
part
of
the
ride.
B
But,
generally
speaking
as
I
reviewed
the
last
retro,
we
get
a
remarkable
number
of
things
to
improve
this
process
compared
to
what
would
happen
in
1.3,
and
that
really
speaks
to
everybody's
commitment
to
continuous
improvement
and
I
am
so
impressed.
Seeing
this
this
is
this.
A
lot
of
moving
parts
do
to
be
able
to
improve.
Like
this
so
big
huge
hand
of
applause
to
this
community
for
being
outstanding,
so
I
don't
really
feel
like.
A
A
A
D
A
B
That
seems
like
okay,
so,
let's,
let's
go
down
to
what
went
well,
what
I'd
like
to
do
is:
have
people
just
go
ahead
and
enter
in
things
that
are
in
the
bullet
list
that
you
think
went
well
in
real
time
and
then
we'll
discuss
them
as
they
come
up?
If
we
need
to
I'm
going
to
try
and
focus
more
of
the
time
on
what
do
we
need
to
improve
and
what
we'll
do
differently
so
go
ahead
and
take
a
few
minutes
to
do
that
and
specifically
there's
a
section
of
this
is
kudos
to.
B
A
Kudos
is
also
one
of
the
spots
where
we
can
collect
information,
because
I
do
want
to
do
a
1.4
release
community
award,
just
as
we
did
14
a
1.3.
So
we
can
the
the
kudos
and
the
who
did
awesome.
Things
is
helpful
for
me
to
coalesce
that
I've
already
had
some
nominations
emailed
to
me
and
mentioned,
but
it's
another
good
spot.
B
The
of
course,
the
video
the
recorded
version
this
we
don't
did
to
seal
it,
the
cool
typing
in
place,
but
things
that
are
coming
up
so
far
follow
through
on
improvement.
That's
what
I
mentioned
before
exception
process
to
prevent
last-minute
code
churn
excellence,
a
hard
line
on
not
accepting
large
code
changes
late
in
the
process,
the
features,
repo
genius
and
release
notes
from
Aaron
yeah.
Thank
you
so
much
for
that
understanding
of
what
is
beta
versus
alpha
early.
B
B
All
right
and
then
let's
do
real,
quick
guide
boy.
All
these
people
have
been
super
stars
in
this
process
of
full
wittrock
has
been
amazing
and
managing
what
seems
like
a
million
different
cats
evar.
Thank
you
so
much
for
all
the
great
updates
that
you
do
every
week.
It
makes
following
the
process
so
much
easier:
David,
Oppenheimer,
yes
and
I'm,
going
to
pronounce
that
Anna
roots
name
right,
probably,
but
you
did.
B
Good
and
Aaron
awesome
work
as
always:
Justin
Santa,
Barbara,
yes
and
I
also
want
to
thank
Justin
for
work
on
cops,
because
that's
really
great
Emma
did
Jared
awesome,
Doc's
docs
for
the
win,
rocks
and
cig
docs
great
work.
Ok,
so
both
if
you
want
to
add
some
of
these
after
the
fact,
let's
go
ahead
and
do
that,
but
right
now,
let's
move
on
since
we're
our
time
window
to
what
could
have
gone
better.
B
B
E
A
And
that
has
to
do
with
improved,
tooling
and
improved
axe
accessibility
to
the
release
process,
so
that
I
think
is
a
I
think
that
is
a
combination
of
needing
to
get
volunteers
who
want
to
help
and
recognize
that
they
will
need
a
a
ride-along
passenger
from
google.
Who,
then
goes?
Oh
yeah,
you
can't
do
that
note
to
self.
We
have
to
fix
that.
A
Okay,
we
have
chosen
a
1.5
releases
are
if
I
remember
correctly,
and
we
have
handed
off
the
one
point
for
point:
point:
release
responsibilities
from
Philip:
he
is
now
free,
go
Philip,
and
so
maybe
what
we
can
do
with
1.5
is
get
someone
from
the
community
to
partner
with
the
1.5
release
are
and
do
it
in
sort
of
the
reverse.
This
time,
yeah.
F
I
mean
I'm
working
with
Zara
right
now.
His
name
is
subtly,
and
I
can
tell
you
that
if
anyone
is
interested
in
contributing
to
some
of
the
more
grungy
aspects
of
the
release,
please
feel
free
to
reach
out
to
him
he's
already
working
on
sort
of
who
owns
what
areas
for
one
not
five,
and
you
know
anything,
would
be
appreciated.
Okay,
Caleb.
G
B
B
Alright,
so
next
bullet
point
was
lots
of
on
lots
of
automation
to
build
project
management
features
on
top
of
github.
Will
it
ever
be
possible
to
use
dedicated
tool
for
project
management?
Patchy
foundation
is
Jim.
Cloud
Foundry
has
pivotal
tracker
put
this
in
there,
and
would
you
like
to
speak
to
it
briefly?
I.
F
Put
this
in
here
and
I've
been
super
appreciative
to
the
work
that
Igor
is
done,
but
I
noticed
that
I
wanted
a
lot
of
it
is
manual
and
a
lot
of
kind
of
our
tooling
with
the
plan.
Labels
and
pinging
people
are
things
that
we
I
have
seen
be
more
discoverable
I
mean
you
love
that
it's
just
github
underneath
so
from
that
point
is
rather
the
the
you
know.
F
The
on
ramp
is
pretty
low
there,
but
when
we
have
you
know
three
or
four,
or
you
know,
and
repose
now,
with
specially
with
the
incubator
work,
makes
it
really
hard
to
get
kind
of
a
global
view
arm
over
the
project
as
a
whole
and
I'm
just
wondering
if
there
ever
is
a
time
where
we
could
use.
You
know
some
dedicated
kind
of
product
management
too
late
in.
B
F
Good
sounds
great.
Does
it
next
item
you
Sarah
is,
I
think,
is
brian.
The
one
Brian
grant
has
been
trying
to
look
into
this
area
too.
Would
he
be
a
good
person
for
monster?
I
didn't
hear
who
was
interested
in
this
for
him
to
hook
up
what
those
are
a
cig
that
is
looking
into
these
areas.
I
know
I've
heard
this
come
up
a
bunch
and
I
know
personally,
I
get
a
little
frustrated
with
it
too.
So.
A
I
think
the
answer
is
always
most
recently
been
looking
at
it.
That
I
know
for
sure
is
david,
oren,
chick
and
we've
tried.
I
know
the
brian
is
also
tangentially
related
and
generally
the
contributor
experience
working
group
is
probable.
We
the
best
place
to
put
this
end
or
the
PM
group,
because
we
have
defined
p.m.
to
be
definitely
product
but
to
product
to
product
managed.
F
B
E
So
I'll
go
through
these
quickly,
so
the
first
one
when
the
p0
flakes
is.
We
have
p0
flakes
open
until
the
like
weeks
before
we
needed
to
cut
the
release
and
the
Peter
flakes
had
zero
attention.
I
get
all
throughout
the
release.
Cycle
and
they've
been
open
for
months
in
pain,
fp0
level,
so
we
could
have.
We
could
have
saved
there's
a
lot
of
stress
I,
just
staying
on
top
of
those.
E
The
next
one
is
still
open
issues
left
in
the
milestone.
This
is
a
similar
vein.
We
track
issue
who's
a
we
do
it
burned
down
through,
what's
in
the
milestone,
to
try
and
figure
out
how
much
progress
we're
making
towards
mm
actually
releasing
on
time
and
there's
a
lot
of
issues
that
are
just
left
open
there
they're,
like
feature
requests
many
weeks
past
feature,
freeze
that
just
never
even
got
started
so
closing
those
out,
maybe
a
feature
freeze
time
like
just
doing
a
total
bankruptcy.
There
might
be
a
good
idea,
upgrade
test,
upgrade
version.
E
F
Phil
yeah
I
had
a
question
on
this.
To
how
well
understood
is
it
in
the
community
that
the
nuances
of
virgin
skew
that
we
only
I
think
we
only
go
back
two
versions
of
that?
You
know
this
is
something
that
is.
You
know
we
view
as
important.
As
you
know,
adding
features
ability
to
you
know,
have
this
queue
system
still
working.
Do
you
feel
like
it's
well
understood?
No,
it
wasn't.
G
A
F
My
point
of
this
is
only
that
you
know
yet,
maybe
even
before
the
doc.
Just
agreeing
with
you
maybe
include
this
in
the
doc
update
item,
but
I
kind
of
can't
tell
if
there's
widespread
agreement
on
what
we're
trying
to
do,
and
it
is
that's
cool
I.
Just
don't
know
if
there's
widespread
agreement
on
it.
If
that
makes
sense,.
E
F
E
But
there's
there's
many
types
of
upgrade
testing
we
do
in
version.
Skew
is
like
one
subset
of
that
and
there's
yeah
there's
a
whole
bunch
of
different
types.
It's
highly
nuanced
and
putting
together
a
plan.
Inscription
of
that
would
definitely
be
helpful,
specifically
I.
Think
the
low-hanging
fruit
here
is
just
have
those
tests
day
passing
and
keep
them
running
against
skew
master
head
verse.
The
last
versions
we're
going
to
test
that
should
be
simple
to
do,
and
it's
pretty
straightforward.
E
A
E
E
The
second
item
is
actually
too
closely
related
to
the
first
action
item
we
have
there,
which
is
there's,
there's
also
manual,
upgrade
testing,
which
is
separate
from
the
automated
upgrade
testing
its
resource
intensive,
and
it
limits
how
mitt,
how
often
we
can
run
it
oftentimes.
We
can't
actually
run
it
in
the
final
version
that
we
cut,
because
we're
carry
paging
changes
up
to
the
Larry
very
last
moment.
E
Too,
I
think
the
the
plan
there
would
be
fully
defined.
What
upgrade
testing
is
what
needs
to
be
done
and
the
different
pieces,
and
some
of
that
can
be
automated
so
that
actually
I'm
there
is
make
it
the
stuff
that
can
be
automated.
Make
it
easily
automatable
then
find
the
folks
who
should
be
responsible
for
automating
or
did
define
how
to
automate
it
and
then
to
find
all
the
pieces
that
we
need
to
automate
and
then
I'm
the
owners
for
those
and
have
them
do
it.
B
E
B
E
So
one
thing
I
noticed
was
there
wasn't
capacity
reserved
and
the
sub
sub
teams?
I
talked
to
and
folks
I
talked
to
for
taking
on
random
tasks
related
to
the
release.
This
man,
when
I,
was
trying
to
find
help
if
there
is
constantly
discussion
going
on
of
saying,
like
oh
but
I'm
working
on
this
other
thing,
and
so
now,
I
need
to
talk
to
this
person
to
see
what
some
more
important
and
just
really
made
my
finding
folks
to
help
with
task
made
it
a
lot
more
hard.
B
E
Yeah
I
mean
I,
think
it's
that
it
takes
a
crew
of
people
to
get
this
done
to
some
degree
and
we
allocate
one,
and
we
should
define
ahead
of
time
like
where
where's
Dover
from
going
to
go
to
when
we
need
more
more
folks
on
board
and
it
doesn't
take
a
crew
the
entire
time,
it's
really
as
it
gets
closer
and
closer
to
the
end.
Yeah
more
tasks
come
up
at
that
time.
We.
E
Okay
and
then
there's
just
one
more,
which
is
a
the
definition
around
like
feature
complete
mostly
it
was
worked
really
well,
there's
a
couple
instances
where
features
were
considered
feature
complete,
but
they
weren't
fully
functional
and
they
still
need
a
test
written
for
them,
and
so
this
actually
really
dragged
on
for
a
long
period
of
time
and
I.
Think
I
think
we
need
to
have
a
real,
clear
definition
that
at
feature
complete
time.
There
should
not
be
open,
known
issues
with
features,
and
we
should
have.
E
B
F
A
F
That
sounds
good.
That's
not
even
really
exciting
to
me.
The
idea
that
you
could
be
running
tests
at
scale
does
clown
foundry
run
on
the
cloud
provider.
Do
they
have
their
own
hardware
yeah,
so
cloud
foundry
runs
continuous
tests
at
minion,
so
it's
broken
up
into
a
different,
a
lot
different,
but
I
had
a
project
level.
It
runs
on
AWS
as
you're
a
couple
different
flavors
of
OpenStack.
You
know
vsphere
loud
and
there's
a
couple.
A
F
F
F
So
the
thousand
no
test
is
done
with
Amazon.
If
that's
not
done
continuously,
it's
done
kind
of
at
a
more
like
I'm
on
a
release,
kind
of
kids
or
if
there's
particular
improvements
targeted.
Actually
I
can
link
in
the
little
tracker
to
show
you
how
they're
working
on
that,
because
it
about
to
run
them
again
because
it
Jason
I.
A
F
Want
a
rat
hole,
but
if
you
could
send
me
a
link
to
scaling
hungry
that
would
be
cool
and
I
can
try
to
forward
that
to
Philip
gee.
Who
is
the
I,
don't
know
if
he's
on
this,
this
call
or
not
Philip.
Are
you
there
he's
in
Poland,
so
the
time
in
network
out
on,
like
you
forwarded
to
me
and
I'll,
forwarded
to
him
and
I,
think
that
would
be
something
to
be
very
actually.
A
Also
have
access
to
the
cluster
at
the
cognitive
foundation,
but
it
would
require
us
to
put
some
sort
of
substrate
under
it.
Be
that
OpenStack
or
bare
metal
on
you
know
being
able
to
fit
under
metal.
But
we
have
access
to
hardware
on
top
of
the
Google
Cloud
and
then
and
and
of
course
we
can
purchase
just
as
anyone
else
can
amazon
time.
B
I
B
F
Cover
it
at
all,
I
have
one
question:
okay:
is
there
like
it's
hard
to
tell
what
we're
doing
the
burned-down
meetings?
If
you
know
we're
in
the
center
of
things,
is
there
like
an
army
of
people
who
are
trying
to
contribute
to
the
releases
willing
to
do
grungy
work
and
just
don't
know
how
to
get
involved?
Or
at
this
point,
are
we
doing
a
good
job
catching
people
in,
and
there
was
also
there's
an
email
list.
I
don't
know
if
people
know
about.
I
think
it's
called
the
burned-down
milestone
list
or
something
like
that.
F
If
someone
could
paste
that
down,
because
I
always
forget
the
wording
and
gmail
tabs
it
for
me
because,
like
that's
a
we
started
using
that
as
a
central
Communication
place
on
it
and
so
a
I
know,
people
knew
what
that
list,
because
it
was
very
helpful
and
too
I
didn't
know.
If
you
know
their
people
at
home,
you
know
going
darn.
I
really
want
to
help
this
release.
You
know
I
it's
hard
to
tell
I'd
say
so.
B
B
E
E
It
wasn't
there
and
had
inductive
issue
and
then
so
we
had
to
go,
find
someone
who
knew
how
to
be
there
like
comment
on
the
issue
or
found
someone
you
could
give
us
stayed
on
it
I'm,
not
sure
I'm,
not
sure
a
better
process
there.
It
might
be
like
just
to
paint
bought
that
pains
these
things
setting
expectations
around
how
often
they
get
updated
is.
A
A
I
think
we
need
to
figure
that
out,
because
the
burn
down
the
milestone
burned
down
mailing
list,
which
I'm
still
looking
for
the
email
address
for
only
has
45
members
and
its
people,
who
are
very
interested
in
and
engaged
with,
what
the
release
cycle
and
that.
But
it
isn't
necessarily
be
engineer
who
has
the
issue?
That's
blocking.
B
A
E
F
Okay,
I
think
just
knowing
about
that
list,
which
I
think
a
lot
of
people
didn't
know
even
in
google,
it
is
helpful
and
just
maybe
sending
that
out
for
a
if
you
want
to
help.
This
is
a
you
know,
a
good
list
to
go
as
a
central
spot,
because
otherwise
you
have
all
the
cigs
and
then
order
you
have
to
show
up
at
a
meeting
and
with
time
zones
that
doesn't
always
work.
F
Piece
so
I
guess
what
I
was
curious
about.
Was
you
know?
I
was
you
know
it
I
wanted
to
contribute,
and
I
did
not
know
about
that
list
like
for
me,
once
I
found
out
about
that
list,
I
used
it
a
lot
for
one
dot
for
and
I
started
communicating
on
that
list
as
a
public
way
of
sharing.
You
lean,
and
you
know
it's
the
things
that
maybe
you
don't
want
to
spam
to
all
of
cooper.
Nities
dev,
but
for
people
are
interested
in
the
release,
specifics
and
then
to
be
gritty.
F
It's
a
good
place
to
send
data
and
went
out
for
actually
I.
Don't
even
think
it
was
open
for
everyone.
It
wasn't
a
member
to
post
that
got
fixed
very
quickly,
so
I
don't
know
if
we
can
use
that
more
aggressively
to
let
people
know
if
you're
interested
in
helping
you
know
this
list
is
a
good
place
where
there's
a
lot
of
traffic
and
where
people
can,
you
know,
volunteer
and
I.
Don't
know
it
seems
like
a
nice
central
point
of
communication,
I
see
what
you're
saying
so
just
reusing
that
list.
F
Because,
where
do
you
go?
If
you
don't
happen
to
be
sitting
next
to
you
and
you
don't,
you
can't
make
the
burndown
meeting
for
one
reason
or
another.
We
don't
even
know
when
the
bird
now
meeting
is
like
how
do
you
get
involved?
So
I
thought
you
know
that
that
might
be
a
good
entry
point
yeah.
That
sounds
great.
B
A
E
A
E
B
B
B
A
G
A
G
A
G
A
A
G
G
G
Mean
I
it
it's
clearly
in
better
shape
than
it
was
last
release,
so
there
may
not
be.
There
may
not
be
a
concrete
thing
that
needs
to
be
done
there.
The
the
work
that
we
you
need
to
do
is
to
it
is
to
get
away
from
polling
and
start
and
either
implement
web
hooks
or
some
other
mechanisms.
We
we
can
scale
larger
because
we're
running
out
of
github
tokens
so.
G
A
A
B
A
B
B
Okay
and
so
I'd
say
that
it
sounds
like
these
I'm
going
to
go
ahead
and
just
go
these
as
green
to
because
it
sounds
like
those
mobile
and
legit
better
we're
ahead
of
our
time,
which
is
great
because
it
looks
like
we're
going
to
wrap
up
in
time.
Is
there
anything
else
that
needs
to
be
discussed
or
parking
lot
items?
It
looks
like
there's
one
to-do
item
about
issue
enters
attending.
Byrnedm
means
that
whole
thing
that
we
talked
about,
who
wants
to
take
ownership
of
that?
That's
for
anybody.
A
So
I
have
project
management
tracking
or
not,
as
well
as
upgrade
testing
discussion
and
then
doc
update,
which
includes
versions
you
as
a
specific
case.
So
those
are
two
topics
for
the
community
meeting.
Then
there
is
also
sending
out
the
burden
down
address
list
and
use
it
mean
what
it
is
expected
to
be
used
for
is
another
community
meeting
topic
most.
F
B
A
I
would
like
to
say
that
for
the
first
time
since
I
started
at
Google
last
week
we
had
community
leaders
for
the
community
meeting
and
while
the
agenda
was
all
prepped
ahead
of
time,
it
went
flawlessly
there's
a
recording,
the
recording
was
published,
it
went
it
went
up
just
as
we
expected
it
to
so.
The
answer
is:
I
am
trying
to
make
myself
fungible
as
well.
It's
a
slower
process.
A
I
also
have
another
part-time
community
person
that
is
working
with
me,
Cameron,
who
is
starting
to
learn
the
community
and
is
taking
off
loading
and
turning
into
more
repeatable
processes,
a
lot
of
the
stuff
that
I
have
been
doing
so?
Okay,
you
had
to
try
to
make
everyone,
if
not
replaceable,
at
least
representable
excellent.
F
A
A
B
B
A
I
have
to
say
that
this
has
to
have
been
a
much
more
smooth
release,
because
the
last
retrospective
took
us
two
hours
to
discuss
all
the
good,
the
bad,
the
ugly
and
the
what
we
need
to
do
better-
and
we
just
did
this
in
45
minutes-
I'm
sure
that
there
will
be
updates
and
additions
to
the
doc
as
we
share
it
out
to
Cooper
name's
dev,
but
but
I
think
it's
well
worth
a
pat
on
our
collective
backs
for
improving
this.
So
much
so.
Thank
you
all
for
your
efforts
on
on
our
project
together.