►
From YouTube: Kyma Prow Migration WG meeting 20181005
Description
Meeting notes: https://docs.google.com/document/d/1ljEAoCBJXlxx_ATPyvKZ1KoyFOSIBzEAOkN-2H-HhUY/edit
A
A
Okay,
so
at
the
beginning
will
present
the
goal
of
this
working
group.
It
will
be
presented
by
by
me
next
Salima
Noctis
present
current
status
of
the
immigration
to
pro.
The
next
point
is
some
organizational
topics,
so
we
need
to
elect
meters
and
the
final
point
you
will
be
some
open
stop
for
any
questions
and
our
answers
and
today
meeting
meeting
notes
will
be
written
by
Chris
Angela
Christian.
Are
you
ready.
A
Some
of
us
know
that
previous
approach
does
not
scale
well,
and
we
want
to
somehow
resolve
that
problems.
How
we
want
to
measure
if
we
succeed,
so
we
have
a
deadline
and
defined
for
this
working
group
that
migration
needs
to
be
done,
but
at
the
end
of
easier,
and
we
need
to
get
positive
feedback
from
FEMA
developers
and
there's
also
defined
assessment
frequency
monthly.
So
it
means
that
after
every
month
we
should
check
if
we
are
on
tracked
and
what
is
the
attitude
to
the
Pro?
And
next
is
next
point.
A
Here-
is
the
exit
criteria,
so
in
what
what
situation
this
group
should
stop
working.
So
first,
we
will
be
if
the
prod
does
not
cover
all
our
needs
and
the
second.
If
the
group
just
do
not
give
enough
support
for
achieving
the
goals
presented
just
before
that's
anyone
want
to
add
something
to
twist.
Due
to
this
point,
to
the
goal
for
this
working
group.
C
Just
maybe
to
clarify
so
when
we
say
that
all
components
migrated
to
new
CI.
Of
course,
we
don't
assume
that
it's
really
one
to
one
migration,
but
we
I
guess
take
into
account
that
we
should
also
take
this
opportunity
that
all
of
us
are
focused
and
we
should
already
improve
the
things
not
just
copy
the
pipeline,
but
maybe
improve
the
pipeline
as
well.
Right
I
give
you
an
example
which
I'm
interested
with
the
topic
of
selective
testing
so,
for
example,
that
we
already
work
on
the
pipeline,
where
it's
enabled
by
default.
C
C
E
D
D
E
Think
that
we
should
use
common
sense
here,
so
I
would
like
to
have
some
first
iteration
as
soon
as
possible,
of
course,
which
it
depends
on
the
on
the
effort
here.
So
if
this
changes
workers
for
examples
that
will
cost
us
two
months
of
waiting,
refactoring
or
whatever
I
would
skip
that.
So
if
it
is
straightforward
in
implementation
in
pro,
we
should
do
that
here.
So
so
that's
for
me
the
common
sense,
if
we
should
think
so,
we
should
not
move
Bradley
all
these
jobs.
E
B
B
A
B
C
C
There's
also,
this
important
topic-
that's
lemon
mentioned
when
he
first
came
up
with
this
idea
of
the
working
group
that,
for
example,
we
should
also
already
with
this
group,
because
again
we
were
focused-
maybe
choose
already
some
basic
set
up
that
will
have
for
using
boats
and
the
plugins
the
process
and
maybe
some
basic
ones.
We
should
also
enable
them
like
the
okay
to
test,
which
we
don't
have
so,
and
it's
kind
of
must-have
right,
I,
guess,
because
we
don't
have
want
to
have
a
current
situation.
We
have
with
the
CI
right.
E
C
E
Agree
with
you
that
we
should
just
collect
tickets
and
then
set
the
priorities
doing
her
some
some
kind
of
planning
next
I
do
not
know
we
should.
We
should
do
the
planning
and
prioritize
them
here.
So
we
should
have
some
basic
topics
like
yeah
really
run
tests
for
four
components
for
pull,
request
deployment
for
on
a
GE,
and
then
we
can
add
additional
tasks
like
boats
whatever,
and
this
will
be,
then
we
will
set
priorities
accordingly.
E
I
would
maybe
set
up
separate
or
already
reused
this
timeslot
for
for
the
planning.
So
maybe
my
I
think
that
we
will
just
we
want
to
have
tickets
prepares
already
before
the
planner,
so
to
not
create
them
during
the
planning,
and
then
we
can
have
just
a
quick
run
with
setting
priorities.
So
what
is
the
most
important
thing?
We
will
work
next
week
next.
F
A
F
G
F
Yeah
some
colleagues
and
I
here
try
to
came
up
with
some
tasks.
First
of
all,
sorry,
oh
sorry,
for
my
voice.
F
It's
it's
on
the
Google
Docs.
So
I
will
share
the
link
on
the
chest
and
then
I
can
share
it
on
the
select
channel
as
well.
So
basically,
we
came
up
with
some
tasks
from
infrastructure
part
to
documentation
and
security.
So
this
might
be
an
input
for
for
us
to
create
the
issues.
I
even
started
operating
some
issues
here
on
the
testing
Ferrara.
You
can
also
review
them
and
and
then
maybe
we
can
using
them.
We
can
even
prioritize
and
decide
what
to
work
on.
F
G
F
Perfect,
so
what
I
have
done
so
far?
The
first
I
investigated
some
some
of
the
plugins
and
I
came
up
with
this
investigation
results.
Here
is
a
pull
request.
You
can
take
a
look.
I
even
chose
some
of
them
for
the
first
installation,
so
maybe
you
can
directly
relieve
them
from
this
PR
and
give
me
some
feedback.
You
know,
please
include
this
one,
and
you
know
exclude
this
one,
something
like
that
and
then
I
defines
the
first
set
of
configurations
for
those
plugins,
but
this
definitely
needs
more.
Like
extensive
investigation.
F
Later
on
and
I
have
installed
a
pro
cluster
on
google
kubernetes
engine,
we
have
one
here
with
three
notes
and
twelve
virtual
CPUs.
So
it's
pretty
enough
for
us
at
the
moment
and
if
you
need
to
do
anything,
just
let
me
know
I
can
share
the
cube
config
with
you
as
well
and
I
think
we
can
even
ask
for
one
more
cluster,
maybe
for
testing
purposes
and
I
integrated.
F
My
pro
set
up
there
with
the
chemo,
but
we
have
on
get
up
and
also
I,
said
testing
for
the
poll
and
also
one
repo
here
from
chemo
incubator.
For
this
pro,
so
it's
already
started
listening
and
even
students
taken
and
over
here
they
started
to
work
with
proud.
They
wrote
a
scripts,
it's
still
in
going
so
maybe
in
this
next
weeks
they
can
even
share
sound
effect
with
us.
You
know
how
the
prow
is
behaving
and
if
it's
solves
their
problems
and
I
have
created
an
installation
scripts.
F
It's
also
in
one
of
the
PRS
here
you
can
just
use
it
to
quickly
set
up
a
prowl
installation
on
google
kubernetes
engine.
There
is
one
thing
here
which
should
be
discussed
again
in
this
group:
the
usage
of
starter,
oh
I'm,
using
it
at
the
moment,
but
last
week,
I
read
one
of
the
comments
from
kubernetes
contributors
saying
that
we
shouldn't
be
using
the
starter.
Eml.
It's
just
for
testing
purposes
is
that
we
should
reuse
the
deployment
files
for
H
these
prowl
components
and
also
we
need
to
customize
them
for
our
needs.
F
So
that's
another
topic.
Maybe
we
need
to
discuss
later
on
and
I
came
up
with
a
basic
feedback.
You
can
also
find
it
in
one
of
the
PRS
here
in
this
report.
It
says
:
to
link
and
also
not
Jes,
to
link,
so
almost
everything
that
we
need
for
building
our
component
schema
components,
and
also
one
controversial
thing
here-
is
the
usage
of
the
Korean
Decker.
F
F
So
that's
why
it
looks
like
the
only
option
is
to
use
the
crane
deck,
but
maybe
we
can
discuss
about
it
later
as
well.
If
you
have
any
other
idea
and
what
I'm
doing
at
the
moment,
I'm
trying
to
come
up
with
one
basic
bill,
chip
I
actually
started
implementing
something
very
basic,
like
cloning
to
repo
and
running
make
CI
for
the
components.
F
E
E
F
E
F
E
C
C
F
Was
just
for
the
installation?
Actually,
you
know
we
need
to
configure
some
plugins
for
the
first
installation,
at
least,
for
example,
trigger
plug-in
is
triggering
the
jobs.
So
it's
it's
a
must-have.
That's
why
I,
actually
edit
it
and
I
think
you,
for
example.
Okay
to
test
is
also
a
must,
but
of
course
we
can.
F
E
A
E
F
For
that,
for
example,
there
are
some
other
things
we
need
to
think
about.
For
example,
one
thing
is
secret
management's,
you
know
we
have
secrets
for
icky
top
and
for
GCR
at
the
moment.
We
need
to
find
a
way
to
use
them.
You
know
from
proud.
This
is,
for
example,
one
topic.
That's
we
need
to
take
over
right
now
and
another
thing
is
threat.
Modeling
I,
don't
know
what
we
like.
A
A
So
next
week
in
our
agenda
is
leaders
election.
So
let's
have
a
brief
look
at
the
team,
a
project
community
repository
when
we
have
defined
this
role
and
group
leaders.
So,
as
you
can
read
here,
working
group
must
nominate
at
least
two
group
leaders
and
group
leaders
make
sure
that
the
group
topics
are
relevant
to
the
kheema.
A
They
facilitate
the
group
meetings
maintain
communicate.
They
meet
meeting
minutes.
Group
leaders
also
record
tasks
identified
by
the
group,
and
they
are
also
responsible
for
smooth
communication
with
other
six
working
groups
and
the
Mekons
fail.
So
do
we
have
any
volunteers
to
be
such
group
leader?
Don't
be
shy.
A
E
We
talked
about
the
community
repository
so
to
query,
to
make
it
clear
what
to
do
for
them
before
the
next
meeting
and
what
we
can
and
how
to
start
right
now
we
have
because
we
will
have
maybe
planning,
but
when
you
we
should
select
topics,
we
can
start
right.
Now
we
have
a
couple
of
people
that
can
do
the
work
before
next
meeting
here.
I
think.
E
E
B
E
D
A
C
A
C
A
D
B
B
G
C
B
B
B
F
Yeah,
that's
true,
of
course,
there's
one
plugin.
Maybe
we
can
enable
it
as
well
I
think
it
was
called
config
upload
or
something
like
that.
So
it's
we
will
listen,
one
of
the
repositories
for
config
changes
and
whenever
it
changed
to
the
configuration
file
is
March,
then
it
will
automatically
pick
up.
B
B
B
E
F
G
H
B
B
G
F
B
Think
for
notes,
okay,
because
in
the
beginning
at
least,
we
can
monitor
disruptions
when
we
will
be
safe
further
in
the
implementation,
we
may
think
of
persistence,
the
deluxe,
but
for
now
I
think.
Yet
we
don't
need
it
as
from
the
beginning.
Yes,
because
even
if
the
test
will
fail,
then
we
have
this
job.
What
do
to
look
into
it?
Yes,
for
some
time
at
least
for
now
it
be
fine.
F
Would
it
be
okay
to
use
tarter
mo
file
for
now
for
you
as
well,
because
at
least
it
can
pick
up
the
changes
and
it
can
run
the
jobs?
So
so
I
don't
know
what
it's
missing
key
moments,
but
kubernetes
guys
says
we
can
only
use
it
for
testing.
B
E
B
E
G
A
A
B
E
Agree,
we
shouldn't
do
all
this
work
that
is
known
how
to
do
that.
We
need
to
do
this
investigation
and
give
guide
and
guideline
how
to
do
that.
Yes,
so
one
component
or
two
components
that
may
be
two
components
to
have
parallel
job
or
whatever
I
took
one,
maybe
not
enough,
but
two
components
and
the
rest
will
be-
will
follow
the
pattern
here.
B
H
Would
be
also
nice
if
we
could
build
a
component.
That's,
for
example,
build
several
docker
images.
There
are
several
cases
of
those
which
seems
maybe
more
complicated.
We
don't
have
to
go
for
the
most
complicated
to
complicated
one,
but
at
least
find
one
that
the
builds
to
docker
images,
for
example,
to
be
sure
that
we
can
do
it
easily
yeah
for
the
others.
Yes,.
D
B
B
B
B
D
C
B
C
B
H
E
H
E
H
E
G
F
F
A
Okay,
so
that's
all
yes,
so
thank
you
for
today's
meetings.
Next
meeting
will
be
organized
by
the
leaders
on
the
next
Friday
also
the
leader,
but
maybe
now
no
I
will
not
feel
organizing
it
I
hope.
Okay,
so
thank
you
for
today.