►
From YouTube: Argo Contributor Experience Office Hour 25th Mar 2021
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
So
good
morning,
everyone
again
or
good
evening,
I,
if
I'm
not
mistaking
today's
meeting,
is
a
ux
ux
centric
meeting.
So
we
had
couple
agenda
items
from
from
the
last
week,
but
I
guess
it's
fair
to
start
from
ui
today,
because
we're
supposed
to
focus
on
ui.
B
B
All
right,
I'm
gonna,
stop
your
share
alex.
If
that's
okay,.
B
I
can
see
it
perfect,
all
right.
So,
for
those
of
you
that
don't
know,
we've
started
a
new
project
or
we've
started
a
new
big
feature
for
rollouts,
which
is
going
to
be
adding
an
api
server
and,
along
with
it,
a
ui
and
the
intent
behind
this
is
to
eventually
integrate
it
with.
B
So
this
is
what
we've
built
so
far,
so
this
is
running
an
api
server
and
rollouts
is
hosting
this,
so
this
will
also
be
a
sub
command
of
the
argo
rollouts
cube,
ctl
plug
out
plugin,
so
you'll
be
able
to
do
our
go
or
a
cube,
ctl
rollouts,
dashboard
and
see
this
locally.
So,
yes,
I've
got
a
few
rollouts
here.
This
is
just
an
overview
list
of
all
the
rollouts
that
are
in
this
namespace.
That's
just
the
demo
namespace.
D
Your
audio
is
in
in
a
loop.
I
think
you
need
to
yeah.
B
How
is
it
now
good,
okay,
sorry
about
that?
Let's
see
all
right,
can
you
guys
see
it
again.
B
Great
okay,
so
this
is
what
we're
hoping
to
integrate
into
cd.
So
I'm
going
to
resize
the
window
here.
So
in
cd,
you
guys
are
familiar,
there's
a
side
panel
that
will
slide
out
to
show
information
about
deployments
and
so
for
a
rollout.
This
will
just
be
an
additional
tab
that
you
can
click
on
and
see
information
about
the
rollout.
So
we
have
some
information
here.
B
So
just
the
summary,
so
strategy
step
set
weight
and
then
there's
some
actions
you
can
perform
here
and
these
actions
aren't
going
to
be
available
in
cd
because
we
have
to.
We
would
have
to
come
up
with
an
our
back
story
for
that,
but
for
the
local
ui,
when
you're
running
it
locally,
just
to
make
it
easier
for
users,
you
can
restart
and
you
can
see
the
pods
here
restart
you
can
roll
it
back
I'll!
Wait
for
this
to
finish
to
roll
back.
B
You
can
set
container
container
images,
so
this
will
suggest
you
a
list
of
all
container
images
that
you've
used
in
the
past
in
this
rollout-
and
you
can
see
your
steps
here.
So
this
one
has
somewhere
around
eight
steps,
and
it's
currently
waiting
for
me
to
promote
it
so
see.
B
We
can
go
ahead
and
promote
full.
So
now
what
this
is
going
to
do
is
it's
going
to
change
the
image
and
it
should
promote
it.
You
can
also
see
your
analysis
runs.
This
is
something
that's
a
work
in
progress.
There's
going
to
be
more
information
here
about
each
analysis
run
so,
namely
when
it
fails
why
it
fails,
and
typically
there
will
be
more
than
just
one
here,
but
it
does
max
out
at
I
believe
10..
B
B
Yeah
there
is,
this
is
the
night
mode.
B
Thank
you,
I'm
hoping
that
this
eventually
will
trickle
down
to
cd
as
a
whole.
So
a
quick
aside,
I
have
not
imported
the
argo
ui
npm
library
for
this
project
and
the
main
reason
for
that
was
a
lot
of
argo.
Ui
is
not
using
react
hooks,
and
so
I
wanted
to
refactor
them
anyway,
and
so
this
seemed
like
a
pretty
good
opportunity
to
use
as
a
testing
ground
for
new
components
that
could
eventually
make
it
into
argo
ui.
B
G
Awesome
is,
is
this
basically
so
like?
Are
there
plans
to
to
more
migrate,
the
the
aggro
uis
so
to
to
this
like
look
and
feel
as
well.
B
Not
that
we've
discussed
so
far,
but
I
guess
that's
a
discussion
that
we
could
have.
A
I
guess
maybe
initial
goal
should
be
to
make
you
know,
like
maybe
don't
have
very
prominent
differences.
It's.
I
personally
feel
like
it's.
Okay,
if
that
ui
is
slightly
different
and
you
know
it's
okay
to
integrate
it,
even
if
it's
not
really
much
argo
ui,
but
I
guess
the
end
goal
should
be
that
it's
100
the
same
and
we're
still
using
the
same
library
but,
as
remington
said,
it
gives
us
opportunity
to
play
this.
You
know
modern.
G
A
A
A
Yeah
and
what
remington
didn't
mention,
but
basically
it
was
easier
for
me
to
understand
what
we're
doing
so,
the
reason
to
build
that
ui
originally
was
you
know
we
have
argus
argo
loud
cli
and
that
cli
gives
a
lot
of
information.
A
A
Users
will
have
two
options:
they
can
keep
using
cli
or
they
can
launch
local
ui
and
see
that
you
know
in
the
in
the
browser
by
a
localhost
and
then
and
next
step
is
to
integrate
it
into
our
velocity
and
we
have
not.
I
didn't,
I
don't
think
we
had
time
to
prepare.
Basically,
there
is
still
a
plan
to
prepare
a
design
document
and
describe
how
exactly
we're
going
to
integrate
it
into
our
city.
B
I
think
the
other
interesting
thing
to
note
there
is
that
we
plan
on
making
that
extensible,
so
not
just
exclusive
coupling
between
rollouts
and
cd,
but
that
will
be
a
sort
of
api
that
others
can
use
to
add
their
own
uis,
their
own
ui
sections
to
our
go
cd.
So
a
good.
D
Yeah,
I'm
working
on
a
proposal
on
how
argos
cd,
basically
a
extensions
feature,
would
allow
third-party,
including
rollouts,
like
basically
treating
other
projects
and
provide
a
way
that
some
other
like
the
crds,
can
provide
like
a
view
that
can
be
embedded
into
the
argo
ui
without
changing
rocd.
D
So
I
mean
you
can
configure
this.
You
know
at
runtime,
basically
a
way
that
you
can
embed
third
party
eu
license
to
argo
cd
and
extend
it.
So
I'm
working
on
that
proposal.
So
look
for
that
in
probably
in
the
coming
weeks.
B
A
The
one
that
I
was
hoping
to
discuss
was
you
know:
shobik
wanted
to
propose
to
have
contributors,
community
meeting
planning.
H
A
A
If
we
have
five
with
six
volunteers,
then
it's
good
enough
and
basically
we
have
pool
of
moderators
and
then
we
can
just
automatically
switch
every
week,
and
I
guess
you
know-
maybe
we
don't
have
to
if
we
can
try
to
find
volunteers
right
now.
Who
wants
to
be
in
that
pool
and-
and
I
will
be-
you
know-
I'm
volunteering
for
sure-
and
I
guess
someone
from
entwin's
team
will
join
me.
A
E
A
Awesome,
so
how
about
that?
If
I,
if
I
basically
you
know,
create
a
list
of
everyone
who
did
it
so
far
and
then
feel
free
to
remove
yourself
if
you're
unable
to
do
that
and
yeah
and
then
we'll
go
with
that
list
sounds
good,
okay,
awesome
and
yeah,
and
we
we
need
to
find
someone
for
the
next
week,
because
I
think
we
basically
didn't
have
a
moderator
last
week.
A
Let's
go
thank
you
and
I
think
I
will
help
you
because
I
basically
it
was
my
fault
that
we
kind
of
skipped.
No,
no
worries,
no
worries.
Okay
sounds
good,
so
basically
I
will,
after
that
meeting,
will
prepare
released
and
have
it
in
the
same.
You
know
working
document
so
again,
this
first
and
and
then
I'm
going
to
use
just
alphabetical
order.
A
F
Yeah,
I
was
just
curious
like
through
which
channels
you're
getting
feedback
on
argo
cd
2.0.
I
was
specifically
thinking
about
the
uui
for
purposes
of
like
what
more
improvements
to
make,
but
it
could
just
also
be
in
general.
A
I
guess
it's
basically
a
brainstorming
session
like
how
do
we
collect
so
historically,
when
you
kind
of
like
every
feature,
we
would
build
for
a
release
already
had
someone
waiting
for
the
teacher,
and
we
cannot
know
who
to
ask.
I
think
2.0
is
kind
of
different
that
we
don't.
You
know
we
improved
logs.
We
don't
have
a
specific
person
who
is
waiting
for
better,
looks
you
so
you're
rightly
like.
We
need
some
way
to
collect
feedback
from.
I
guess
multiple
users
from.
A
A
G
G
A
So
far,
every
time
we
improve
something
we'll
always
break
something
at
the
same
time,
and
you
know
as
long
as
we
fix
it,
it's
fine.
So
I'm
really,
you
know
I'm
trying
to
find
a
way.
You
know
how
could
we,
I
think
these
things
are
called
regressions.
Basically,
if
we
have
any
regression,
this
should
be
cached
first
and
I
kind
of
struggle
to
find
any
way.
How
to
you
know
what
could
we
do
to
improve
the
process
of
catching
those
yeah?
A
A
And
maybe
did
you
have
any
other
thoughts
about
you
know
like
did
you
want?
Do
you
think
it
makes
sense
to
maybe
ask
specifically
for
feedback
in
some
kind
of
in
media
and,
like
I
don't
know,
we
could
create
a
discussion
and
ask
everyone
to
explicitly
tell
us
what
they
like
and
don't
like
about
argoshi
diy
changes.
F
Yeah,
I
think,
probably
making
a
proposal
to
like
gain
community
feedback
in
some
way.
Just
so,
we
know
like
how
like
what
goals
to
focus
on
for
next
releases
and
where
our
attention
should
be
focused
on
so
maybe
I'll.
Just
like
make
a
proposal
in
the
issue.
Queue
to
like
make
a
way
to
solicit
feedback.
A
A
It
sounds
good
to
me
as
well.
Anyone
else
you
know
do
stop
asking
if
you
think
it's
bad
idea,
but
basically
it
sounds
like
we
want
to
maybe
create
some
some
for,
like
github
discussion
or
issue
specifically
to
collect
positive
negatives,
about
the
ui
changes,
and
then
you
know
tweet
it
in
our
twitter
and
ask
everyone
to
write
that
feedback.
Oh
maybe
maybe.
F
A
I
Okay,
how
the
the
cataclysm
scenarios
could
help
getting
feedback
from
from
customers
or
maybe
who
are
new
to
the
project?
I
already
see
a
katakura
scenario
for
probably
alex
you
created
it
before.
Is
this
maintained.
A
Oh,
yes,
I
it
was
basically
I
created
it
and
the
hope
was
that
we
have
some
super
easy
way
to
test,
and
basically
I
didn't
continue
because
it
was
like
extremely
slow.
Basically,
if
I
was
the
user,
I
might
not
use
it
because
I
would
just
prefer
to
you
know
just
start
my
own
mini
cube.
It
was
quicker.
I
I
don't
know
if
I
know
that
category
was
acquired
by
another
company,
I
forgot,
and
hopefully
they
improve
the
infrastructure.
A
A
And
for
those
who
don't
know
what
katakure
is
it's
a
it's
like
a
playground
and
you
can
package
kubernetes
application
into
that
playground
and
you
can
kind
of
script
user
experience
so
people
just
someone,
can
click
the
url
and
get
the
mean
cube
cluster
or
some
kind
of
kubernetes
cluster,
and
then
you
know
just
keep
clicking
buttons
and
basically
start
your
application
and
play
with
it
and
it
works.
It's
just
taking.
You
know
like
10
minutes,
if
not
longer,
just
to
start
our
ocd.
C
A
A
A
To
admit
that
that
you
know
for
me
last
week
this
week
is
basically
I
just
started
working.
I
was
pretty
much
offline,
you
know
since
since
monday,
and
I
saw
that
I
think
we're
still
getting
bug
fixes,
and
basically
I
saw
that.
May
sharma
and
jonathan
they'll
fix
several
bugs
already
and
I'm
basically,
I
have
exact
same
question
as
you
and
I
I
plan
to
help
as
much
as
I
can
starting
today
and
go.
A
A
Basically,
it's
running
for
on
a
big
portion
of
it's
not
really
dog
food,
it's
kind
of
production
into
it
clusters
and
no
complaints.
So
far,
it's
been
extremely.
H
G
That
sounds
very
awesome,
and
did
you
have
the
chance
to
you
know
this
this
one
long-standing
issue
where
we
not
where
we
were
not
able
to
figure
out
the
root
cause
the
tls
to
the
to
the
repo
server
has
this?
Has
this
become
better
with
go
116?
I
guess
it's
time
to
check.
A
H
A
Everywhere
then
we
can
it's.
I
would
wait
another
week,
you
know
just
yeah
yeah
and
then
they
could
close.
That
issue.
That's
a
good
thing
and
I
wanted
to
say
that,
even
though
we
don't
have
complaints
it,
basically
it
kind
of
it's
a
good
proof
enough
that
we
don't
have
really
bad
regressions,
but
you
know
we
still
mainly
have
to
test
new
features,
especially
because
android
users,
they
kind
of
maybe
don't
even
know
that,
yet
that
there
is
new
log
viewer,
so
they
have
not
tested
it.
G
D
Yeah
and
the
performance
actually
improved
the
reconciliation
performance
inc
improved.
So
we
have
some
performance
improvements
that
basically
it
was
achieved
through
the
some
additional
caching
that
we
do
for
to
to
do
diff
avoidance-
and,
I
think,
possibly
get
some
get
avoidant
stuff
too,
that
improved.
A
I
think
we
didn't
notice
one
issue
and
I
kind
of
I
didn't.
I
don't
think
I
got
enough
confidence
to
create
issue
for
that,
and
I
just
documented
it
in
release
notes.
So
basically
radius
upgrade
took
more
time
than
usual
and
I
guess
it's
because
we
moved
from
you
know,
version
five
to
version
six
and
there
was
a
kind
of
incident
because
of
mistake,
but
basically
one
radius
stuck
and
I
had
to
recreate
it
only
in
one
cases
out
of
actually
28.
A
D
A
D
I
think
maybe
the
last
thing
to
mention
about
2-0
is
that
so
we
were
running
1.8
for
you
know
a
while,
and
then
we
had
this
incident
where,
as
our
platform
team
was
upgrading
the
kubernetes
clusters,
all
of
our
argo
cd
instances
suddenly
got
this
long.
10
to
15
minute
hung
time
out
to
the
api
servers,
the
master
kubernetes
masters
and
basically
for
for
10
to
15
minutes
racing.
No
one
could
deploy
that
into
it,
because
the
argo
cds
were
not
be
able
to
talk
to
kubernetes
masters.
D
We
when
we
investigated
this,
we
tried
to
reproduce
this
with
the
198
and
2.0
side
by
side
and
found
that
we
could
reproduce
this
at
will
with
one
argo
cd
1.8,
but
when,
with
the
2.0
sitting
side
by
side
with
it
even
sitting
on
the
same
kubernetes
node,
we
weren't
able
2.0
was
not
susceptible
to
that
same
hung
connection
to
the
api
server.
D
To
this
day
we
still
don't
know
the
root
cause
of
it,
but
our
suspicion
is
that
the
upgrade
in
the
kubernetes
client
libraries
might
have
improved
the
situation.
So
that
was
one
2.0
I
guess
fix.
So
this.
A
A
A
G
A
G
A
H
A
A
How
is
everyone
feeling
and
okay
one
thing
we
didn't
do
correct
me
if
I'm
wrong,
but
I
we
were
supposed
to
do
introduction
when
every
time
someone
new
joins
and
I
totally
forgot
to
introduce
harry
actually
yeah.
I
don't
think
I
did
it
so
he
is
new
to
our
team
at
introverts.
I
One
other
topic
that
comes
to
my
mind:
it's
like
ocd,
best
practices
and
it's
something
I
think
someone
mentioned
that
they
saw
a
doc
somewhere
or
maybe
there's
a
discussion
sometimes
before,
but
I
can't
find
like
we
have
an
updated
version
of
our
ocd
best
practices.
A
One
topic:
we
were
we
kind
of
started
discussing,
but
nothing
concrete.
Yet
we
I
think
we
want
to
come
up
with
best
practices
to
describe
argo
cg.
I
mean
not
argo
cdi
deployment
to
positive
folder
structure.
It's
totally
missing,
but
this
is
one
thing
I
would
want
to
work
on
and
I
so
devin
are
you
you
you.
You
know
you're
kind
of
you're
looking
for
someone
who
also
interested
in
working
on
best
practices,
yeah.
I
And
I
I
had
a
chat
with
yann
about
somewhat
unrelated
topic,
but
touching
on
the
best
practices.
Part
is
like
experienced
folks
like
you
who
have
worked
with
customers
and
they're,
sharing
their
pain
points
of
what
worked
for
them
and,
more
importantly,
what
didn't
work
and
then
combining
sort
of
a
document
of
these
war
wounds
that
okay.
I
A
Yeah,
I
think
it's
good,
basically
yeah
and
I'm
going
to
tell
that,
but
basically
code
fresh
wants
to
co.
You
know
work
on
that
too,
and
they
want
to
codify
it.
They
want
to
build
some
kind
of
tool
that
can
help
people
to
use
argosy
in
opinionated
way.
You
know
in
in
some
way
that
we
think
the
best
way
yeah
and
I
that's
why
we
recently
talked
about
it
as
well,
but
we
don't
really
have
a
good
plan
yet
at
least
I
have
not.
A
I
have
never
seen
a
single
document
that
describes
you
know,
for
example,
a
git
repository
how
how
folders
and
that
repository
should
be
laid
out.
I
think
it's
super
useful.
I
personally,
I
I
remember
a
lot
of
questions
from
users
who
kind
of
asking
exactly
that
question
yeah.
A
Maybe
I
don't
know,
I
think
we
it
makes
sense
to
think
of
this
code
fresh
and
I
think
they
need
help.
A
I
Yeah
makes
sense
yeah
if
you
have
any
contacts
I'll
be
happy
to
initiate
a
conversation.
Maybe
crowd
crowd,
source
a
document
and
then
maybe
take
the
conversation
from
there.
Okay.
A
Yeah,
if
you
want
to
help
them
at
least
I
can
you
know,
start
the
chat,
and
you
know
I
guess
in
just
our
cdf
channel
and
invite
three
people
who
was
active
and
I
I
will
mention
I
will
try
to
invite
them
into
the
into
the
next
meeting.
So
if
something
happened
during
the
week,
we
will
have
it,
you
know
we
can
discuss
it
in
the
teaching
awesome.
Thank
you.
Thank
you.
D
In
terms
of
like
operationalizing
our
ghost
cd,
like
you
know,
like
operated
that,
like
you
know,
when
do
you
sharding
and
how
should
you
you
know,
do
configuration?
D
Are
you
asking
about
kind
of
best
practices
on
using
argos
cd
for
like
like
well
cofresh
is
giving
an
opinion
weight
of
almost
like
configuration
management
of
of
clusters
like
bootstrapping
clusters,
using
a
tool
so
which
one
or
maybe
both
are
you
asking
about.
I
I
What
are
the
some
important
things
to
remember
about
full
versus
pull
or
push
so
this
set
of
practices
that
someone
could
use
either
or
but
it's
just
using
one
set
of
approaches
worked
for
many,
and
for
these
these
reasons.