►
From YouTube: 2021-06-24 CAPZ Office Hours
Description
Cluster API Provider Azure office hours for 2021-06-24
Meeting agenda: http://bit.ly/k8s-capz-agenda
Join https://groups.google.com/forum/#!forum/kubernetes-sig-cluster-lifecycle if you don't have edit access to the meeting agenda.
A
Hello,
everyone,
my
name
is
david
justice
and
I
will
be
hosting
this
week's
meeting
for
cluster
api
provider
azure.
We
are
a
cncf
project,
which
means
everybody
here,
please
be
kind
to
each
other
and
let's,
let's
just
be
nice
and
have
fun
awesome.
A
So
this
week's
agenda
it
looks
like
do.
We
have
any
psas.
C
I
I
guess
my
open
discussion
thing
is
really
more
of
a
psa.
If
we
want
to
do
it
now,
because
I
don't
think
so
I'll
go
ahead.
Just
like
last
month,
unfortunately,
kubernetes
patches
came
out,
and
then
we
had
some
issues
in
the
pipeline.
The
windows
versions
from
last
thursday
are
published
and-
and
I
had
and
I
figured
I
think
we
figured
out
a
way
forward
to
fix
what
broke
so
probably
by
tomorrow.
They'll
all
be
done.
C
B
Yeah
and
for
cabsie,
particularly
also
they're,
they
have.
This,
like
cubidium,
revert
for
the
ephemeral
storage
requirement
which
actually
made
the
multi
control
plane
scenarios
flaky
on
120
plus,
so
that
should
fix
it
in
120,
121.
C
A
All
right,
just
quick
recap:
they
will
be
out.
When
are
you
thinking,
they'll
be
published.
C
Assuming
the
basically,
we
have
to
revert
to
the
last
known
working
version,
and
then
I
have
to
start
publishing
again
and
assuming
that
all
goes
smoothly.
You
know
it
it
can
take
up
to
a
day
for
it
all
to
make
it
out.
So
hopefully,.
A
I'll
do
it?
Oh
sorry,
nader!
I
could
see
you
thanks.
Man
awesome
all
right,
cool,
so
cecile.
I
think
you're
next
up.
B
Yeah
so
also
sort
of
a
psa,
I
guess,
but
as
we're
getting
closer
to
the
release,
we
started
working
on
the
release
notes
for
cappy
with
the
captain
maintainers.
I
think
we're
hoping
to
get
that
ready
by
end
of
today,
but
I
in
the
meantime,
since
this
is
a
really
big
release,
there's
hundreds
of
commits
in
it.
B
I
thought
we
should
get
a
head
start
on
our
release,
notes
and
try
to
maybe
highlight
what
our
users
should
pay
attention
to
like
what
are
the
things
in
the
release
that
we
care
about
showing
since,
like,
I
think
it's
going
to
be
really
hard
to
digest.
The
entire
t
of
the
release
notes
like
if
you're,
a
user
and
you're
trying
to
go
through
every
feature
in
bug
fix
that
merged.
B
So
I
started
a
dock
which,
for
now
just
has
like
a
really
rough
draft
with,
like
the
generated
release,
notes
basically-
and
I'm
not
even
sure
that
includes
everything
that
has
two
so
far,
because
I
was
still
trying
to
figure
out
how
to
get
it
from
zero
four
zero,
but
without
stuff
that
went
into
the
patches
since
we
branched
off.
At
some
point,
after
releasing
from
master
branch
for
a
while
so
yeah,
but
everything-
that's
just
to
say,
I'd
love
some
help
on
that.
B
B
But
if
you
have
other
suggestions
of
things
that
are
in
the
release
that
should
be
called
out
there
please
add
to
it
and
then
let's
write
like
a
one
to
two
sentence
summary
for
each
one
to
describe
like
what
it
does
and
why
it's?
Why
it's
a
big
deal
so
yeah,
that's
the
dock!
Everyone
here
should
have
access
already,
but
let
me
know
if
you
don't.
A
Awesome
that
that's
a
great
idea,
thank
you
for
taking
action
on
that.
B
A
Awesome
anything
else
that
we
have
to
cover.
B
I
would
say:
let's
do
a
quick
release
check-in
just
to
make
sure
nothing
has
popped
up
since
last
time.
That
were
that
we
should
be
aware
all.
A
B
B
A
B
B
B
A
All
right
so
refactor
remaining
services
following
established
pattern.
This
is
something
that
would
be
a
non-breaking
change,
probably
a
breaking
change
at
the
actual
api
like
go
api
layer,
but
nothing
externally
surfacing.
B
You
want
to
actually
we
would
lose
the
history
if
we
do
that,
but.
B
Yeah,
let's
just
leave
it
with
the
other
one,
since
it's
it
has
like
the
examples
and
the
yeah.
This
is
a
cool
feature,
though,
in
general.
A
Okay,
anybody
remember
exp.
What
was
it
managed.
A
B
You
can
just
leave
it
like
this
yeah,
okay,.
A
There,
okay,
so
awesome:
let's
move
to
smilestone05x
all
in
favor,
all
right
yay!
Let's
do
it.
A
Okay,
I
think
that
is
all
it
looks
like
we
have
a
clear
backlog
for
zero
five,
zero,
all
right.
B
So
the
one
thing
that
we
should
probably
check
in
there
is
the
azure
cluster
identity
templates.
Things
like
we
should
come
to
a
conclusion
on
that
later.
Does
that
make
sense.
D
Yeah,
so
the
one
we
have
right
now
that
is
changing
this
tests,
I'm
just
having
some
trouble
with
adding
adding
updating
the
templates
for
like
the
copy
test,
because
the
cluster,
usually
what
we're
doing
in
the
other
ones,
is
that
we
create
the
secret,
but
in
the
name
space
right.
D
So
hopefully
I
get
that
done
and
then
the
other
issue
that
I
will
the
other
thing
I
have
to
do
that
I
will
create
a
different
issue
for
is
updating
the
actual
templates
and
then
adding
a
deprecation
warning
saying
that
this
is
going
away.
I
guess
in
the
next
release
or
something
that
environment
variables
are
going
away.
You
have
to
use
the
actual
cluster
identity.
B
I
think
we
do
right
and
then
the
other
pr
is
nine,
which
is
making
the
variables
optional
in
the
manager
that
one's
on
hold
it's
open.
D
Yeah
I
was
gonna.
Do
that,
like
I'm
working
on
it
today
and
tomorrow,
I'm
hoping
to
be
done
by
tomorrow.
Are
we
expecting
that
our
release
is
gonna,
be
up
like
it's
not.
B
D
Yeah,
I
think
I'm
planning
to
like
I've
been
working
on
it
since
yesterday's.
Hopefully
it
will
will
be
done
in
the
next
couple
of
days.
A
Yes,
awesome
yeah
somebody
could
probably
take
on
the
deprecation
notice.
You
know
in
parallel,
we
can
probably
spread
some
of
those
tasks
out.
A
I
I
think
it
is
calico's
fault.
Do
we
have
any
indication
of
speaking
of
calculate?
Do
we
have
an
indication
when
the
next
release
of
calco
will
land.
D
It
should
be
in
the
next
couple
weeks,
according
to
the
schedule,
they
usually
do
too
early
to
the
quarter.
So
the
last
one
was
in
like
mid
april
or
something
so
roughly
around
like
the
next
couple
of
weeks,
but
like
if
I
went
on
their
slack
and
they
don't
like
have
a
day
before.
D
D
Right
actually
create
two
issues:
one
for
the
duplication
and
one
for
adding
like
the
identity
reference
to
the
templates,
so
that,
if
somebody
wants
to
pick
one
of
them,
while
I'm
still
working
on
that.
A
All
right
do
we
have
so.
It
sounds
like
we're.
We're
focused
pretty
pretty
well
to
end
of
next
week
for
the
release.
That's
fantastic.
A
This
is
nearly
a
year
or
so
of
of
work
to
get
this
next.
You
know
minor
release,
so
this
is.
This
is
significant,
just
great
job
for
everybody
driving
it
to
this
point
and
let's
get
this
out
and
hopefully
not
take
so
long
for
the
next
one.
C
Yeah
it'll
be
really
quick,
all
right
cool
can
I
sure
screen
or
do
you
have
to
give
me
that
power.
C
Wait
a
second,
let
me
make
sure
yeah
all
right,
so
I've
I'll
make
it
a
little
wider.
So
I've
had
this
pr
out
there
for
a
little
while,
but
you
know
it's
not
very
well
timed
with
respect
to
the
release
that
we're
doing
so.
I
don't
expect
we
need
to
merge
it
anytime
soon,
but
essentially
this
replaces
the
jaeger,
collector
and
and
the
stuff
we're
doing
for
tracing
with
the
open
telemetry
component
tree.
C
So
it's
more
of
an
under
the
covers
change
and
it's
been
and
the
significant
thing
it
does
is
it
lets
you
do
traces
from
controller
manager
and
send
them
out
to
azure
app
insights
seamlessly,
which
is
a
nice
place
to
view
them.
C
It's
been
sitting
out
here
for
a
while,
so
I
kind
of
went
nuts
on
it
yesterday
and
I
have
another
pr
that
eventually
I'll
put
out
there,
that
adds
jaeger
back
in
because
it
seemed
kind
of
like
fender
lock-in
or
not
quite
kosher.
If
the
only
place
you
can
view
traces
is
in
azure
where
you
might
have
to
pay
for
it,
so
I
put
jaeger
back
in
there
and
so,
let's
hope
this
works.
C
So
at
the
beginning,
the
meaning
literally
all
the
prep
I
did-
was
make
tilt
up
and
I
got
lucky
with
calico,
so
it
actually
came
all
the
way
up.
For
some
reason,
ephemeral
failed,
but
I
don't
care
so
I
just
started
up
windows,
so
there
would
be
some
actual
traces
to
look
at
you
can
see.
Now
we
have
an
open,
telomere,
telemetry,
collector
agent
and
an
open
telemetry
collector
resource.
C
We
still
have
a
jaeger
all-in-one,
although
it's
kind
of
it's
just
for
viewing
traces
as
we'll
see
in
a
second
prometheus
hasn't
really
been
touched,
although
under
the
covers,
all
the
guts
have
been
upgraded
to
what
is
now
open,
telemetry
1.00
rc1,
which
is
great,
which
means
most
of
their
apis,
have
started
to
stabilize,
because
this
has
been
a
real.
C
It's
it's
the
definition
of
an
unstable
api.
So,
but
the
nice
thing
you
can
do
now
is
click
on
jager
all
in
one,
for
example,
and
you
get
this
link
here
view
traces
that
opens
up
the
jager
all
in
one
component
and
you
can
just
query
four
traces
and
we
should
have
some
by
now,
because
we've
been
actually
deploying
stuff-
and
here
we
go.
Let's
find
something.
That's
interesting
that
took
more
than
a
second,
let's
say.
C
This
is
just
running
on
it
in
memory
it
doesn't
have
a
real
backing
store,
like
obviously
jager
really
should,
but
for
development
purposes.
This
is
they're
out
of
the
box.
I
don't
think
it's
using
a
lot
of
memory
and
it's
super
convenient
for
doing
this
kind
of
stuff,
but
then
we're
also
as
long
as
you
go
ahead
and
put
in
tilt
settings
a
key
to
an
azure
app
inside
sportspace.
C
You
send
them
here
as
well.
It's
a
little
hard
to
find
because
there's
a
million
things
here
related
to
metrics
and
we're
not
populating
metrics
we're
only
populating
traces.
But
if
you
click
on
transaction
search
say
show
me:
what's
happened
in
the
last
30
minutes.
You
get
essentially
a
similar
view
to
what
we're
looking
at
in
jager.
Let's
find
a
meaty
one
here
to
poke
at
come
on.
C
That's
it
pretty
much.
The
overall
take-home
story,
I
guess,
is
you
know
we
can
it's
a
big
win
for
us
to
have
open
telemetry,
which
we
already
did,
but
now
now
we're
kind
of
all
in
on
the
otlp,
their
recommended
collectors
and
stuff
like
that,
and
the
nice
thing
is
so
we
can
collect
everything
in
one
way
and
then
fan
it
out
and
send
it
to
whoever
we
want
in
this
case
we're
sending
it
to
jaeger,
even
though
we're
not
collecting
traces
in
the
jaeger
jaeger's
preferred
format.
In
the
first
place.
C
A
I
totally
love.
I
got
to
see
how
you
did
that
link
the
viewed
traces.
C
Oh,
it's
it's
pretty
trivial,
it's
a
one-liner
and
tilt
just
does
it,
for
you
say
kate's
case
resource,
and
then
you
declare
the
port
forward
in
a
certain
way
and
it
gives
it
a
title
like
that.
That
is
awesome.
You
made
that
so
much
easier.
C
C
So
it's
really
only
intended
for
development
and
it's
in
the
it's
in
the
readme.
That's
in
the
pr,
but
essentially,
if
you
add
azure
instrumentation
key
and
you
set
it
to.
C
You
go
to
overview
for
your
app
insights
and
you
look
over
here
in
the
upper
right.
It's
this
thing.
So
if
you
set
azure
underscore
information,
instrumentation
underscore
key
to
that
value
and
then
you
do
make
tilt
up
it'll
just
work.
D
C
And
if
you
don't
set
it
to
that
right
now
it
actually
you
get
a
warning
until
which
might
be
a
little
heavy-handed,
especially
now
that
we
have
jaeger
in
there
as
a
default.
So
I
might,
we
might
think
about
changing
that,
because
a
warning
shows
up.
As
you
know,
it's
kind
of
in
your
face
here
and
tilt.
C
So
I
guess
my
overall
question
is
what
I'm
running,
what
I'm
showing
you
here
amounts
to
an
xxl,
it's
a
huge
ton
of
code
change,
and
so
I
probably
prefer
to
stage
it
the
way
it
is
now
and
have
this
be
a
follow-on
pr,
but
I
also
don't
want
to
distract
from
the
release
stuff.
So
just
look
at
it
when
you
have
time,
if
you
can.
A
That
was
really
awesome
now
I
think
we
have
a
little
bit
of
work
to
do
to
you
know,
make
that
production
ready
and
make
it
easy
for
folks
to
just
turn
this
on.
A
We
also
have
an
issue
open
outside
of
here
in
azure
cloud
provider
to
also
instrument
that
so
start
to
provide
more
observability
across
all
of
these
all
the
azure
related
stuff,
so
yeah
anything
else.
Any
other
questions.
C
A
That
was
a
fantastic
demo
and
I
please
encourage
anyone
who
would
like
to
do
a
demo
in
the
future.
Please
do
we
love
seeing
it
or
I
do
at
least
I
don't
want
to
speak
for
anybody,
but
that
was
a
great
job
man.
Thank
you
all,
right
with
that
said,
we
have
one
more
week
to
sprint
towards
the
finish
line
and
get
zero
five
zero
out.
I
want
to
thank
everybody
for
their
effort
so
far
and
look
forward
to
getting
this
new
release
out
and
then
talking
about
it
with
everybody.
A
Awesome
all
right
great,
thank
you
so
much
everyone-
and
I
hope
you
have
a
wonderful
week
and
we'll
talk
to
you
soon.