►
From YouTube: Kubernetes SIG Testing 2019-04-02
Description
A
B
Totally,
if
you,
if
you
send
that
dock
out
to
community
suggesting
at
Google
Groups
comm
channel
I,
got
a
lot
okay.
A
A
It's
gonna,
look
like
a
lot
more
than
it
is
because
it's
gonna
be
a
lot
of
like
changing
import
statements
and
a
bunch
of
a
little
bit
just
like
a
bunch
of
little
plated
small
changes
in
a
bunch
of
little
areas
and
because
there's
so
many
of
them
I.
Just
you
know,
as
far
as
I
can
tell
everything
is
gonna
work
and
tests
pass
and
all
that
stuff,
but
I
just
and
I
I
just
know.
There's
gonna
be
little
things
that
that
could
potentially
go
wrong.
A
B
I
think
you'll
find
in
most
places,
even
though
there's
a
concrete
client
being
passed.
The
method
is
acting
like
some
interface
I'm
said
right
and
we
use
that
for
testing
up
anyway,
so
I
think
there
might
even
be
fewer
changes
than
what
you're
expecting
that
approach
does
some
good
were
you
able
to,
because
I
think
you
guys
were
saying
that
lab
was
gonna,
be
one
of
your
first
targets
have
a
chance
to
poke
around
and
see
like.
If,
there's
you
know,
big
differences
between
that's
new.
That
would
make
that.
A
Not
like
specifically
in
scallywag,
we
haven't
gotten
into
the
implementation
yet,
but
but
just
our
experience
from
Jenkins
X
is
that
the
differences
are
the
different
differences
tend
to
be
just
implementation.
Details
I
mean
they
all
kind
of
conceptually.
Do
the
same
things
and
I
kind
of
outline
this
a
little
bit
in
the
doc
that
you
know.
Basically,
they
all
follow
the
github
model.
A
Right
github
is
the
standard
that
everybody
is
trying
to
kind
of
be
more
or
less
it's
just
it
because
they
are
different
and
because
they
have
some
incentive
to
kind
of
get
you
to
use
their
thing
and
not
other
people's
thing.
The
API
calls
are
different,
but
I
mean
they're,
not
conceptually
they're,
not
different
enough
for
it
to
be
a
huge
problem.
It's
just
a
matter
of
translating
some
generic
data
model
to
the
individual
data
models
of
each
of
these
individual
providers.
A
So
if
you're
talking
about
get
louder,
if
you're
talking
about
bitbucket,
it's
pretty
straightforward
now,
if
you
get,
if
you,
if
you
stray
away
from
the
github
like
yet
providers,
you
start
talking
about,
you
know
like
I'm
tool
like
Gerrit,
which
just
does
code
review,
then
it
gets
a
little
bit
hairier.
But
if
you're
just
talking
about
kind
of
the
popular
get
providers
that
are
github
like
that,
involve
issue
tracking
and
code
review
and
kind
of
all
the
things
that
prowl
expects
to
be
there
for
it
to
work.
A
C
A
C
Level
I
think
that
sounds
really
good.
But
ideally
you
know
when
you
are
ready.
I
think
it
would
be
much
prefer
to
sort
of
have
some
sort
of
design
with
some
details
about
like
what
are
the
interfaces.
Things
are
going
to
move
and
sort
of.
You
know
a
document
for
us
to
talk
about
and
sort
of
something
more
concrete
to
look
at
and
make
sure
that
this
seems
sane
or
whatever
term.
I.
Definitely
think
that
this
seems.
C
You
know
complicated
enough
that
we
should
have
that
first
and
discuss
there
rather
than
just
you
know,
seeing
a
PR
that
you
know
changes
all
the
github
stuff,
but
or
in
general
that
that's
one
thing
which
scene
is
totally
fine.
It
seems
like
a
sane
and
good
approach.
I
think
there's
a
little
I
think.
C
Api
that
isn't
expressed
as
well
in
the
gitlab,
client
or
something
and
so
there's
a
way
where
we
want
to
specifically
use
you
know
some
part
of
the
github
API
or
some
part
of
the
get
live
API
to
make
it
look
like
a
you
know,
native
gitlab
or
github
application
at
some
point,
and
so
you
know
I'd
be
interested
in
seeing
you
know
how
you
are
thinking
about
that.
If.
C
A
sort
of
exists:
you
can
only
write
to
the
you,
know,
scallywag
interface
and
if
scallywag
doesn't
support
something,
then
you're,
Sol
or
maybe
there's
some
other
way
to
you
know
turn
this
into
the
concrete
or
get
the
github
clients
it
from
I.
Don't
know
exactly
what
the
answers
are,
but
I
think
would
that
be
another
interesting
thing
other
than
just
the
specific
details
about
how
it
looks,
but
how
to
handle
that
problem
of.
Maybe
there
are
some
instances
where
there's
more.
A
A
So
I
didn't
get
a
chance
to
send
it
out
before
the
call
but
I'll
send
it
out
afterwards
and
I
get
into
this
in
a
little
more
detail.
But
basically
the
idea
is
that
you
know
there
will
be
some
can
face
some
entry
in
the
config
file.
That
says
you
scallywag,
you
know
a
little
flag
and
the
config
file.
It
says
you
scallywag,
true
or
false.
If
it's
false,
it's
just
gonna
go
try
to
use
the
github
client
and
it's
gonna
work.
A
Essentially,
the
idea
is
that
for
folks
that
are
currently
using
prowl
and
they're
totally
happy
with
the
way
it
works,
they'll
just
be
able
to
keep
using
it
exactly
the
same
way
and
there
should
be
no
surprises.
No
loss
of
functionality,
I
mean
the
the
the
problem
you're
talking
about
is
just
is
in
the
nature
of
these
things.
Right,
you
create
a
generic
interface.
A
There's
gonna
be
some
some
loss,
and
you
know
because
in
those
minut
details
that
differ
so
I
think
that
I
think
that
you
know
it
is
that
the
cleanest
and
most
elegant
solution-
I,
don't
I,
don't
know,
but
I
think
that
it
makes
sense
to
me
just
because
the
one
thing
I,
the
one
thing
I
don't
want
to
do-
is
to
affect
anybody
that
is
currently
using
prowl
and
it's
totally
happy
with
it.
It's
totally
happy
happy
using
github.
Really
it
should
just
this
should
be
like
they
should
need.
C
C
A
Is
that
hook
will
pass
hook,
will
just
get
a
web
hook
and
then
it'll
send
it
to
scallywag
and
scallywag
will
say.
Ok,
this
is
a
you
know.
Scallywag
will
figure
out
what
needs
to
happen
where
it
needs
to
go
or
scallywag
will
send
back.
Some
sort
of
generic
thing
that
hook
can
understand
and
hook
doesn't
have
to
understand
specific.
A
You
know
github
based
events,
so
that
it
does
sort
of
involved
this
event,
and
this
is
a
part
that
needs
to
be
fleshed
out
a
little
bit
more,
but
basically
a
hook
gets
it's
a
web
hook
from
someplace
doesn't
know
or
care
where
it
came
from,
but
just
know
that
came
true
came
in
from
somewhere
and
then
sends
it
onto
scallywag
scallywag
figures
out
exactly
what
you
know.
Scallywag
knows
how
to
handle
some
predetermine
the
implement,
depending
on
the
implementation
of
scallywag
it'll,
know
how
to
handle
certain
types
of
web
hooks.
A
A
A
Ideally,
it
would
be
a
single
deployment
would
be
able
to
handle
both
and
maybe
even
a
single
instance
of
scallywag
or
possibly
multiple
instances
of
scallywag.
Now,
if
you're
talking
about
like
with
Dennis
so
you're
talking
about
like
from
different
projects
but
then
the
same
or
being
able
to
configure
things
differently,.
C
You
know
right
now
you
know
prowl,
like
sort
of
all
of
the
various
repos
they're
across
different
orgs.
You
know
like
there's
kubernetes
SIG's
as
well
as
kubernetes
and
a
bunch
of
different
repos
and
all
of
those
all
of
those
are
handled
by
proud,
Katie,
oh,
and
so,
if
they're,
you
know,
the
question
is:
if
we
then
wanted
to,
you
know
start
supporting
the
kubernetes,
gitlab
org
or
whatever
would
be
Wanna
now
we
would
be
idea
be
that
we
support
that
in
proud
case
tayo,
or
would
we
then
create
like
a
proud
get
lab
KC?
A
A
B
How
are
you
supposed
to
handle
like
github
isms
if
different
providers
provide
similar
things
like
I'm
thinking,
the
LG
TM
plug-in,
and
you
can
configure
it
so
that
it
behaves
in
a
specific
way
if
you
leave
like
it
could
help
improving
review
if
gitlab
has
a
different
verbiage
for
that
same
feature,
are
you
expecting
people
to
like?
Well,
the
config
allow
both
names
to
make
it
more
collab,
native
or
we'll
have
to
learn.
D
A
My
hope
is
that
the
core
functionality
should
it
should
all
kind
of
be
the
same.
You
know,
but
but
who
knows
there
may
be
it
may
come
down
to
something
like
we're
we're
just
like
okay?
Well,
you
know,
if
you
use,
you
know,
if
you,
we
just
know
that
if
you
use
gitlab
for
ask
alleyway
here's
a
couple
of
plugins
that
aren't
gonna
work,
the
way
you're
expecting
them
to
or
they
might
not
work
at
all
or
something
like
that,
they
may
just
need
it.
That
may
be
part
of
it.
B
If
we're
all
influencing
the
same
interface-
and
you
know
maybe
just
like
throwing
not
yet
implemented
errors
or
whatever
at
runtime,
maybe
instead
of
check
config,
it
would
be
nice
to
validate
that
I've
chosen.
This
sort
of
plugins
and
I've
chosen
this
provided
for
scallywag
and
that
actually
will
work
with
my
oh
right.
B
B
C
A
C
D
D
C
The
answer
is
yes,
except
that
I
don't
know
that
there's
any
we
don't.
Certainly
you
know,
yeah
I,
don't
know
that
there
we
don't
have
anyone
who
we
know
I
feel
like
there's
been
some
demand
in
there
and
some
attempts
and
I
think
there
are
maybe
some
bugs
and
they
may
or
may
not
have
been
resolved,
but
ideally
it
would
work
and
we
would
certainly
want
it
to
work
and
I
feel,
like
other
than
you
know,
devil
being
the
details
of
like
this
particular
weird
thing.
Doesn't
work
right
like
I,
think
Chloe.
C
When
beaten,
there's
me
fiction
theme,
but
for
the
most
part
I
think
it
should
work.
I
would
be.
I
would
not
be
at
all
surprised
if
it's
like,
oh
hey
this,
you
know,
I
think
there
probably
thinks
that
don't
work,
that's
obviously
not
the
Cunha
kubernetes
does
not
use
github
enterprise.
So
that
isn't
our
core
used
case,
and
so
we'll
need
someone
else.
D
A
Yeah
and
we've
done
some,
so
we
on
the
Jenkins
X
project,
we've
definitely
done
some
work
on
using
github
enterprise
with
prowl
and
I.
Think
it's
been
a
while
since
I've
looked
at
it,
but
it
last
time
I
checked,
it
basically
did
work
again.
It's
been,
it's
been
a
minute
now
so
I
haven't
I
can't
say
that
with
much
more
confidence
than
what
other
folks
have
said,
but
yeah,
but
it
github
Enterprise
is
close
enough
to
there.
There
was
one
snag
we
had
where
we
were
on
an
older.
We
had
an
instance
of
github
Enterprise.
A
B
C
C
Was
just
a
basil
job
to
make
that
PR?
That
was
birth?
No,
the
opposite!
The
emerge
the
PR
to
do
the
role
to
to
the
rollout
and
then
the
rollout
job,
which
also
uses
basil
failed.
So
we
think
that
I
thought
that
we
were
at
whatever
we
thought
that
we
were
Tuesday's
version.
Then
I.
Look
today,
it's
like
Oh!
Actually
you
are
on
Thursday,
okay,.