►
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
I
found
this
improved
json
parsing
speeds.
This
is
for
the
azure
cpi.
B
Yeah
exactly
so,
this
role
is
currently
out
on
downstream
landscapes
on
our
side.
That
would.
A
B
Would
see
how
stable
it
is
and
potentially
the
improvement.
A
A
B
A
Yeah
I
mean
it's
not
I.
I
understand
your
your
concerns,
it's
more
like
that.
I
know
that
our
our
customers
want
like
long
stability
right.
So
this
is
a
put
it's
a
a
change
that,
on
the
surface,
looks
really
simple
and
small,
but
it
could
be
like
weird
ways
in
which
this
breaks
yeah,
which
is
something
I'd
rather
not
roll
out
to
customers.
A
B
Yeah,
it's
totally
fine
for
me,
yeah
it
just
it
always
depends.
I
guess,
what's
currently.
B
B
A
C
B
A
fork
where
we
currently
put
it
into
production,
we
learn
from
it,
and
it's
no
problem
to
to
bring
this
upstream,
then
behind
a
flag.
C
A
C
A
B
So
with
this
so
in
bosch
we
use
the
fork
of
nets
right.
Is
that
still
true.
A
Yeah
we
have
a
private
for
or
not
private,
but
we
have
a
fork
that
has
some
like
special
logic
to
do:
authentication
based
on
certificate
names
or
something.
Yes,
that's
how
that
like
go
here.
A
So
that's
this
whole
client
id
gets
interpolated
with
something
from
the
certificates,
the
common
name
or
something
there.
That's
where
it
gets.
The
client
id
form.
A
So
all
that
doesn't
exist
upstream.
So
what
we
do
instead
now
with
this
up
with
this
br,
is
that
we
we
write
out
a
net
config
file
dynamically.
B
C
A
And
this
will
also
in
the
future,
maybe
potentially
we
have
been
thinking
about.
Like
short,
this
is
more
like
so
in
like
design
or
like
ideation,
but
we
have
had
requests
of
like
wanting
to
be
able
to
the
the
net
credentials
that
we
inject
through
the
metadata.
A
To
be
able
to
invalidate
those
after
initial
use
right,
so
you
basically
bootstrap
a
node
and
then
immediately
rotated
nets,
credentials
through
navs,
so
that
you,
you
just
the
secrets
in
the
metadata
servers-
are
not
valid
anymore.
Okay,
yeah
that
I
mean.
That
is
a
thing
that
often
comes
up
with
security
audits.
A
Yes,
I
mean
the
the
scope
of
things
that
you
can
do
with
those
credentials
is
pretty
limited,
but
still
it's
better.
If
we
can
save,
they
are
only
valid
for,
like
the
initial
minute
of
the
vm's
life
cycle,
yes,
yeah,
absolutely
being
able
to
sync
these
credentials
have
more
to
to
be
able
to
have
more
control
of
that
right.
We
would
be
able
to
like
if
we
write
out
those
those
credentials
all
the
time
that
file
it's
not
credentials.
A
It's
like,
I
think
it's
identities
or
something
that
that
are
in
that
file,
but
it
doesn't
matter.
If
we
we
sync
those
all
the
time,
then
we
can
also
start
invalidating
them
easier
more
easily.
That
would
be
like
further
down
the
road,
but
this
this
work
opens
that
up
in
the
future.
A
B
A
B
Yeah,
we
can
definitely
provide
then
findings
at
scale.
B
B
If
it's
still
a
thing,
I
could
look
at
it.
I'm
surely
I
also
have
a
lot
of
gaps,
but
I
mean
certain
proposals
I
might
have
so
I
don't
want
to
be
in
the
critical
path,
but
after
those
three
weeks,
maybe
if
it's
still
open,
I
can
have
a
look.
Okay.
B
B
B
C
C
We
already
used
this
now.
This
fix
in
sap
right.
B
C
B
B
B
A
I
will
ask
around:
let's
see
who's
assigned
to
work
on
this.
A
C
A
Okay,
migrate
to
jammies
done
so.
C
C
Yeah,
it
needs
to
be
re-piped
in
there,
but
I
asked
it
in
the
channel
and
they
yeah
is
it?
Does
it
fix
anything?
Does
it
do
anything?
No!
No,
but
we
bump.
C
B
C
B
C
B
B
Apple
had
autoscaler
also,
apparently
works
and
yeah.
This
one
is
just
I
added
a
little
bit
more
context,
so
create
m
doesn't
seem
to
work
for
heavy
stem
cells
on
aws
it.
It
apparently
requires
to
run
from
an
ec2
instance,
but
at
least
from
our
jump
box-
ecu
to
instance.
It
also
did
not
work,
so
I
don't
think
it's
an
issue
for
us
now.
We
will
work
around
it
by
creating
bosch,
light
stem
cell
up
front
but
yeah,
I'm
not
sure
for
whom
this
is
working
really
so.
A
But
can
you
use
an
and
have
a
heavy
stem
cell
when
you
upload
it
to
a
director.
C
So
it
doesn't
matter,
it's
not
os
related.
It's
it's
been
an
issue
for
a
long
time
and
just
no
one
looks
at
it
because
everyone
is
using
light
stem
cell,
because
why
would
you
use.
A
A
full
stem
cell
yeah
it
has
to
do
with
like
the
way
you
need
to
convert
or
upload
a
heavy
stem
cell
right.
It
needs
to
create
an
ebs
volume
and
then
map
it
to
the
machine
and
then
copy
the
data
from
the
heavy
stem
cell
over.
C
Yeah
and
it
tries
to
connect
to
the
metadata
service
for
that
and
from
your
local.
That's
an
internal
ip
address,
so
once
you're
doing
that
from
your
local
machine,
it
just
doesn't
work.
C
A
C
B
B
A
So
there
is
this,
I
think,
there's
a
what's
it
here
copy
something.
B
Yeah,
probably
yeah,
I
have
to
drop
off
yeah
yeah
so,
but
for
that
issue
I
mean
that
was
our
scenario,
and
maybe,
if
it's
not
an
issue
based
on
specific
dogs,
we
can
close
it.
Otherwise,
someone
who
stumbles
over
the
same
issue
has
a
pointer
where
they
can
look.
B
A
A
C
C
A
A
C
C
C
C
C
C
C
Yeah,
I
will
release
the
the
stem
cells
with
the
the
last
the
kernel
cves
that
we
got
yesterday,
but
it
was
so
busy.
They
were.
C
There
were
yeah
so
once,
and
I
thought
oh
now,
it's
done
and
then
two
hours
later
again,
multiple
cps
for
the
same
thing.
Probably
so
I
just
waited
until
some
packages
get
upgraded
properly.