►
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
A
B
C
C
C
It
was
taking
in
communities
there's
it's
not
quite
an
API
object,
but
it's
like
an
object
that
defines
the
interface
client
binary,
that
the
existing
interface
is
specific
or
it's
called
the
credential
provider,
and
it
provides
credentials
for
like
like
HTTP
authentication,
but
we,
my
proposal,
was
to
sort
of
revamp
this
for
to
add.
Daugher
credential
support.
C
C
There's
a
number
of
tools
out
there
that
do
a
very
similar
thing,
so
they're,
it's
called
like
the
docker
credential
helper
protocol
and
they're
their
binaries
that
are
meant
to
be
executed
by,
for
example,
the
the
docker
CLI
which
go
and
do
something
in
return
credentials
back
to
be
sent
to
the
docker
registry.
So
that
was
suggested
to
me
by
someone
here
at
Amazon,
so
I'm
considering
it's
it's!
Basically.
If
you
look
at
that,
auth
config
struct,
that's
in
the
cap,
it's
basically
a
subset
of
that.
C
B
C
Nothing
there's
nothing
docker
specific
about
anything
and
yeah
I'll.
Make
sure
that
there
isn't
there's
the
the
docker
credential
provider
or
docker
credential
helper
protocol
thing
that
I
was
talking
about
I'll,
make
sure
that
that
doesn't
have
anything
specific
to
docker.
But
it
shouldn't
because
the
there's
two
ways
to
authenticate
to
docker
registries,
there's
the
username
like
the
basic
auth
and
then
there's
a
newer,
jot,
Roth
sort
of
flow
and
the
in
the
docker
registry,
HTTP,
API,
v2
and.
C
C
Would
be
a
concern
to
me,
yeah
I'll,
look
into
that
and
there's
tools
like
there's
a
there's:
an
implementation
for
various
cloud
providers
that
have
image
registries
and,
like
I,
know,
there's
a
there's,
a
GC,
p1
or
GC
r,
there's
a
ec
r.
So
yeah-
I
don't
know,
I
think
it's
it's
definitely.
It
was
definitely
like
proposed
and
the
initial
implementation
was
all
done
by
docker.
So.
D
C
D
C
It
would
really
just
be
I
mean
it's
the
interface
between
the
caller
and
the
binary,
that's
being
exact
in
my
proposal.
It's
the
auth
config
struct.
It
has
you
name,
password
email,
which
is
a
field,
that's
being
deprecated
identity,
token
registry
token,
and
this
other
specification
is
basically
a
subset
where
it's
like
username
secrets,
one
other
fields
and
when
you
use
a
token,
you
just
put
token
as
the
username
and
then
the
token
contents
are
the
secret.
C
B
B
A
Thanks
yeah,
so
there
isn't
a
next
that
I
put
that
I
plan
for,
but
there's
a
few
topics
we
need
to
discuss
for
the
current
provider.
The
cap
deadline
for
118
is
two
weeks
right.
There's
known
that,
do
you
need
do
you
feel
like
you?
Have
everything
like
you
feel
unblocked
and
good
to,
but
keep
writing
the
capping.
You
have
the
right
people
you
know,
or
do
you
still
need
support
from
us
to
ping
the
right
people
and
get
the
right
books
looking
at
this
I
think.
C
B
C
A
And
then
I
think
like
as
a
last
resort.
If
none
of
them
can
the
proof
we
can
ask
Jordan
they
get
just
because
I
know
he's
been
keeping
an
eye
on
the
provider
extract
and
stuff.
So
if
we
kind
of
tell
them
like
hey,
we
need
it.
We
need
this
approved,
so
we
can
get
clapping
fighters
out,
I'm,
sure
he'll.
A
B
My
memory
was
that
we
were
we
were,
we
had
just
done
proof
of
common,
we
were,
we
were
doing
something
on
the
proof
of
concept
and
I
thought
we
were
just
about
ready
to
finalize
the
sort
of
is
it
so
we
wanted
the
group
of
concept
in
a
reasonable
shape
for
the
cap,
but
I
feel
like
we're.
Basically
there
and
one
way
or
the
other.
The
capped
needs
to
land.
A
B
When
the
good
news
is
that
I
think,
we've
got
enough
to
show
what
we
want.
I
mean
that
the
proof
of
concept
basically
works,
I
think
we
need
to
get
the
cap
merged,
and
then
we
get
another
six
weeks
for
the
code
and
the
proof
of
concept
I
think
is
in
we'll
need
to
do
some
work.
But
we
have
some
time
on
that.
A
A
B
B
B
C
B
B
That's
the
absolute
worst
case
scenario
I've
seen,
but
it's
bad
enough
that
I
do
not
want
to
be
in
a
world
where
I'm,
saying
yeah,
you
can't
create
new
nodes
for
three
hours
like
auto-scaling.
Doesn't
work
for
three
hours
and
just
I
mean
I
may
be
a
misunderstanding.
Andrew
suggestion,
but
I
think
that
is
a
consequence
of
Andrews
suggestion.
B
A
D
What
what
number
do
you
consider
large?
Are
we
talking
a
thousand
three
thousand
five
thousand.
Give
me
a
ball,
something
on
that
order.
Yeah
I,
don't
think
I'm
allowed
to
say,
but
something
on
that
order.
Okay,
so
over
a
thousand,
then
okay,
I'm,
just
wondering,
because
if
we
want
to
test
ours,
I
want
to
know
there
might
be
something
we
haven't
thought
of
and.
A
C
C
B
C
B
A
Yeah
I,
agree
and
I
think
we
should
be
clear
with
like
with
what
the
cluster
and
Edmunds
SLO
SLA
is
like
I.
Just
don't
want
to
be
in
situation
where,
like
all
users
convince
themselves
that
they
need
this
migration
without
knowing
that,
if
they
just
did
the
swap
like
I
said
their
downtime
is
like
negligible,
but
they're
doing
all
this
work
for
no
reason
so
like
we
should
really
make
it
clear
somewhere
that
like
use
this,
if
you
have
the
3000
no
cluster,
and
you
have
very
very
strict
requirements
on
reliability,
then.
B
Absolutely
I
mean
I,
think
yeah,
absolutely
I
think
we
should
probably
have
that
in
the
cap.
I'm
saying
you
know,
this
is
designed
for
these
scenarios,
starting
with
H
a
and
one
or
two
other
things,
possibly
even
having
it.
In
the
I
mean
it
sounds
like
we're.
Gonna
have
a
couple
of
flags
on
the
controller
manager,
and
we
should
probably
reference
this.
Even
there.
A
C
So,
just
to
kind
of,
because
I
think
I
I
wasn't
around
around
when
we
were
initially
designing
this
like
did.
We
did
we
build
this
off
of
of
some
amount
of
user
requirements
or
are
we
like?
Is
there
an
opportunity
here
before
we
spend
the
time
implementing
this,
to
talk
to
some
users
who
have
these
large
clusters
and
long.
B
B
A
A
A
A
B
Yeah,
mostly
just
updates
on
where
we
are
so,
the
good
news
is,
we
got
the
kept
merged
last
release,
so
the
beta
requirements
are
known
and
are
done.
I
have
two
engineers
Jeffery
Yang
and
joshu,
who
are
both
working
on
this
frantically.
So
we
are
right
now
still
planning
on
meeting
the
beta
requirements
for
this
coming
release.
B
And
we've
actually
added
some.
Our
own
internal
security
review
has
caused
us
to
add
one
or
two
new
features
that
everyone
seems
fine
with.
So
as
an
example,
the
one
of
the
things
that
should
be
merging
in
the
next
day
or
two
is
will
probably
allow
for
API
server
talking
to
the
local
proxy
server
over
UDS.