►
From YouTube: Argo Contributors Office Hours 7th Oct 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
welcome
to
the
contributors
meeting
today
we
have
an
agenda,
so
the
idea
would
be
to
start
with
it
with
a
triage
section,
then
we're
gonna
go
and
I'm
gonna
be
presenting
two
coming
features:
one
in
rollouts
and
one
in
argo
cd.
A
So
those
are
two
prs
that
I'm
being
working
with
then
alex
is
going
to
be
giving
some
highlights
for
the
hacktoberfest
then
doing
some
then
presenting
the
roadmap
for
argo
cd
2.3,
and
there
is
one
last
minute
topic
here:
service
site
apply
that
hong
will
be
presenting.
So
let's
start
right
now,
so
the
triage.
B
C
Kind
of
the
same
observation
we
didn't
have
much
of
discussions
and
I
actually
there
were
no
discussions.
You
know
no
new
questions
and
discussions
until
yesterday
and
I
just
noticed
that
today
there
were
three
questions
asked
and
I'm
still
going
to.
I
know
answers
at
least
to
two
of
them,
and
then
there
was
a
lot
of
activity
in
github
issues
and
kind
of
I
didn't
there
were.
There
was
one
issue
that
looked
really
bad.
It
sounded
bad
that
the
person
complained
that
we
broke
sync
functionality,
but
I
could
not
reproduce
it.
C
So
it's
like
good
and
bad
news.
I
suspect
that
sync
unnecessary
waiting
for
three
minutes,
but
this
what
person
person
claims
that
argus
dcli
executes
the
same
command,
unnecessary
waits
for
at
least
three
minutes,
but
it's
not
reproducible.
C
On
my
environment
and
looks
like
there
is
something
special
about
the
person's
environment
and
that
was
the
biggest
the
most
notable
change.
Everything
else
was,
you
know
minor
complaints
and
feature,
and
you
know,
suggestions,
that's
it
so
and
yeah,
and
I
guess
the
update
for
the
next
discussion.
Moderator
is
to
you
know,
don't
answer
three
most
recent
unanswered
questions.
I
will
try
to
answer
today.
I
just
didn't
notice
the
questions.
First.
D
Yeah,
actually,
I
went
through
the
discussion
this
week,
like
I
found
it
a
bit
weird
because,
as
compared
to
the
last
one,
so
this
time
it
was
more
related
to
their
configuration,
the
laptop
configuration
or
setting
up
how
argosy
is
so
was
not
sure
exactly
how
I
could
reply
to
them.
So
was
I
wanted
to
have
a
chat
with
either
alex
or
jan
regarding
that,
like,
if
there's
just
a
configuration
like
they're,
just
asking
for
some
help
in
configuration,
how
would
be
the
best
way
to
answer
those
questions.
C
D
Just
discuss
with
you
like
I'll
start
a
thread
in
slack
after
this
and
regarding
charging.
I
got
one
issue
that
was
regarding
helm
values.yaml,
so
one
of
them
were
requesting
for
an
enhancement
if
the
values.tamil
could
even
be
added
through
rest
api.
I
was
not
sure
if
we
are
interested
in
working
on
that
I
could
try.
I
forgot
adding
I
forgot
noting
the
number,
but
surely
I
could
add
that
in
the
dock
right
now.
B
C
C
B
Okay,
it's
a
leo,
actually
call
you
do
you
want
a
quota.
Sorry
say
that
again,
I
I
think
ishita
has
a
request.
Can
you
go
to
scroller
and
in
the
topics
it
seems
here.
A
All
right
can
we
move
on
to
the
next
topic.
A
All
right,
so
the
next
topic
is
going
to
be
two
quick
presentations
for
upcoming
features,
one
inaugural
rollouts,
one
in
argo,
cd.
B
A
Sure
no
problem,
okay,
so
the
feature
that
I'm
going
to
be
presenting
is
the
this
one
that
was
originally
opened
by
alex.
So
the
idea
is
to
support
the
new
ingress
that
was
moved
from
the
extension
package
to
the
networking
networking
package
since
cube
119.
A
So
from
119
to
122.
We
had
the
ingress
present
in
in
the
two
different
packages,
so
but
from
122,
which
is
the
latest
version.
Ingress
is
not
available
anymore
from
the
extensions
package,
which
means
rollouts
is
incompatible
with
the
newer
version
of
kubernetes.
So
this
this
issue
is
to
make
that
to
bring
that
compatibility
back
to
our
rollout.
So
it
will
know
how
to
how
to
interact
with
ingress
resources
from
the
networking
package.
A
A
So
the
approach
that
we,
I
we
had
a
brief
discussion
here,
jesse
alex
and
myself
and
the
direction
that
we
that
I
I
I
went
was
basically
implementing
a
raptor
in
rollouts.
A
So
that
will
decide
if
rollout
should
deal
with
ingress
coming
from
the
extensions
package
or
from
the
networking
package
and
based
on
that
configuration,
which
is
provided
by
by
an
argument
to
the
argo
rollouts
command
actually
to
the
controller.
A
A
So
that's
pretty
much
the
the
feature,
so
the
pr
is
out.
It's
it's
a
little
bit
big
because
the
in
roll
outs,
we
we
leverage
the
network
resource
for
different
purposes.
So
there
are
some
traffic
routers
that
depend
on
this
resource,
there's
also
a
controller,
a
small
ingress
controller
that
will
be
watching
the
the
ingress
and
reporting
it
back
to
the
the
rollout
controller
if
he
noticed
any
any
changes
in
in
the
state.
A
So
there
are
quite
a
few
places
where
this
this
is
well,
where
the
the
the
rollout,
the
argo
rollouts
project,
is
making
usage
of
ingress,
and
that's
why
this
pr
is
a
little
bit
big.
So
the
way
I
decided
to
approach
this
is
to
centralize
everything
in
a
wrapper,
so
there
is
a
wrapper
file.
This
is
a
new
file
that
will
have
this
strap.
A
That
actually
has
the
two
two
versions,
one
from
the
extensions
and
the
other
one
from
the
networking
one
and
there's
also
some
wrapped
methods
that
I'm
then
replacing
everywhere
in
the
code
where
we
depend
on
those
operations,
for
example,
for
patching
an
ingress.
A
So
that's
pretty
much
the
strategy
I
decided.
I
I
try
to
centralize
as
much
as
much
as
I
could,
because
this
is
a
kind
of
a
intermediate
stage
for
rollouts,
because
at
the
end
of
the
day,
once
we
move
once
kubernetes
move
on
the
extensions
are
not
gonna,
be
there
anymore
and
we're
gonna
have
to
to
clean
up
the
code
and
remove
this
from
from
the
code
base.
A
So
that's
pretty
much
it
any
questions.
Anyone.
F
A
119
just
came
out,
but
I
wouldn't
I
I
wouldn't
be
surprised
that
a
lot
of
clients
of
argo
rollouts
are
still
using
180
kubernetes
118
119,
so
this
is
to
so
from
the
original
ticket.
What
alex
wrote
was
that
it's
maybe
too
early
to
drop
support
for
cube
versions
older
than
119.
A
So
at
this
point
we
we
think
it's.
It
would
be
better
to
for
our
rollouts
to
support
multiple
versions
and
once
clients
have
the
chance
to
migrate
and
more
and
more
users
are
on
newer
kubernetes
clusters.
Then
we
drop
support
from
from
the
other
one,
basically
so
that
that
that
was
the.
F
Oh
yeah,
maybe
I'm
digressing
yeah,
maybe
digress.
We
can
take
this
offline.
I
think
in
general.
We
should,
if,
if
it's
good
enough
for
two
bananas,
it
should
generally
be
good
enough
for
us.
I
would
think
if
I
I
know
that
a
lot
of
people
run
a
lot
of
older
versions,
but
it
in
general.
I
would.
I
would
like
to
see
us
kind
of
stick
to
whatever
kubernetes
doing
and
not
do
work
to
support
non-supported
kubernetes
versions,
but
we
can
take
that
offline.
A
A
All
right,
so,
let's
move
on
to
alex,
where
he's
going
to
be
talking
about
oktoberfest
with
ulx.
Yes,
so
let
me.
C
C
It
took
some
time
to
apply
that
label
to
good
first
issues
that
we
have
and
yeah,
and
I
promoted
it
into
it
at
least,
because
I
know
that
intuit
users
encouraged
intuit
developers,
encouraged
to
participate,
and
I
know
that
few
people
already
reach
out
and
they
work
in
compares,
and
I
guess
the
cold
auction
here
is
to
don't
hesitate
and
you
know
promote
it
in
your
organizations
as
well.
So
I'm
going
to
put
that
link
in
the
chat.
G
And
alex
I
just
quit
question
here
for
workflows
and
events.
We
I
think
we
only
highlighted
like
half
a
dozen
or
a
dozen
issues,
I
noticed
you've
got
30
issues.
G
The
reason
I
did
that
as
there
are
always
some
people
who
submit
hacktoberfest
documentation
changes
because
they
want
whatever
t-shirt
is
available
for
it.
That's
fine!
We
love,
you
know,
love
all
contributions,
but
it's
you
know
it's
obviously
quite
useful
for
us
to
be
kind
of
focused
on
those
ones.
Now
you
might
be
in
a
different
position
that
you
have
like
a
lot
more
contributors
and
it
sounds
like
you
probably
do.
G
I
just
wanted
to
kind
of
like
wonder
if
you
wanted
to
be
kind
of
quite
focused
or
more
focused
on
on
those
items.
I
honestly.
C
I
didn't
think
about
it.
I
was
just
thinking
the
more
the
better
and
I
was
just
trying
to
label
as
many
issues
as
possible.
My
only
criteria
was
criteria
was
that
it
was
supposed
to
be
straightforward
issue,
at
least
like
I'm,
not
ready
to
start
design
discussion
about
30
issues,
but
you
know
if
the
change
is
clear
and
basically
you
know
the
contributor
should
understand
from
the
issue
description.
What
to
do
that?
That's
why
I
didn't
hesitate
to
just
apply
that
label
and
yeah.
C
So
your
point
is
basically
maybe
it's
better
to
you
know.
Okay,
I
feel
like
I'm.
My
judgment
is
subjective,
but
I
was
thinking
that
the
most
important
stuff
for
for
current
release
it's
already
in
milestone,
and
I
could
not
find
a
single
issue
that
I
would
suggest
as
a
hacktoberfest,
because
all
of
them
are
like
you
know,
hardly
reproducible
bugs
and
it's
impossible
to
expect
from
external
contributor
to
to
work
on
on
these
for
us,
and
so
I
had
to
switch
to
to
simpler
issues
which
I'm
not
focused
it.
C
E
Just
a
question
because
I'm
not
I'm
not
very
familiar
with
oktoberfest,
I
know
last
year
I
was
already
struggling
with
understanding
it,
but
basically
so
this
is
like
a
public
event
where
people
can
go
to
open
source
projects
right
and
and
pick
an
issue
labeled
hacktoberfest,
and
if
they
implement
it,
they
get
a
reward
right
from.
Like
the
t-shirt
alex
mentioned.
I
think
I
think
same
level
of
understanding.
E
Yeah,
just
I
have
a
question
so
basically,
when
we're
labeling
those
issues
with
hacktoberfest,
we
are
committing
to
like
if
we
receive
a
pr
for
that,
we
commit
to
review
and
eventually
merge
it
right.
So
do
we
so
hacktoberfest
in
the
name.
It
says
it's
probably
scope
for
october
right.
So
I'm
just
wondering
about
the
time
constraints
here.
When
does
the
pr
have
to
be
merged
in
october,
or
does
it
have
to
be
submitted
in
october
and
maybe
merged
on
november,
2nd
or
november
3rd
or.
C
F
C
Yeah
and
I,
but
I
kind
of
I
I'm
not
hoping
to
get
30
public,
at
least
at
least
I
didn't
expect
to
get
like
for
dprs
and
maybe
like
I
was
hoping
to
get
like
five
and
it
felt
like
okay,
but
I
I
agree
it's
like
just
my
guessing
like
not
necessary.
C
I
would
maybe
I
I
I
feel
like
it's
like,
maybe
when
it
concerned.
Maybe
what
should
we
wait?
You
know
like.
Basically,
we
can
go
for
kubecon
came
back
next
week
and
if
we
already
have
like
half
of
this,
you
know
prs
for
half
of
this
issues.
Maybe
we
can
slow
down
and.
E
Yeah
yeah,
I
just
I
I
was
just
thinking
I
mean
because
this
this
is
a
commitment
right
and
if
someone
sends
a
good
pr
and
we
don't
come
around
to
review
it
and
merge
it,
then
yeah
that
probably
would
be
very
unfair
yeah.
Maybe
a
contributor
right,
yeah.
C
F
C
Work
like
do,
I
know
what
should
be
done
here
or
not,
and
most
of
these
things
are
kind
of
either
very,
very
simple
feature
or
kind
of
a
bug
fix
that
is
limited
to
a
small.
It's
like
not
trying.
C
But
I
I
would,
you
know,
really
appreciate
like
second
pair
of
eyes,
if
someone
go
through
these
cracktoberfest
issues-
and
you
know
if
you
can
find
something
you
think
is
like
dangerously
big
and
might
require
a
lot
of
effort
to
review
then
just
feel
free
to
understand.
C
C
It
was
an
enhancement
to
you
know,
improve
a
large
visualization
of
flash
applications
and
I
felt
like
it's
easy
to
implement
it,
but
it
might
be.
You
know
if
someone
else
noticed
that
issue
and
try
to
to
do
it
like.
Basically,
we
might
end
up
not
merging
it,
and
so
it's
like
selfish
to
ask
someone
to
just
send
us.
E
Yeah,
but
I
think
that's
in
I,
I
agree
with
you
that
probably
not
every
issue
labeled
with
hacktoberfest
will
receive
a
pr,
but
I
think
it
may
be
a
good.
You
know
a
broad
selection
of
of
things
that
people
could
pick
from
so
yeah.
I
think
I
think
it's
fine
to
have
to
have
those
but
yeah
I
will.
I
will
also
go
through
them
to
see
if
they
may
be
too
big
for
for
that
yeah.
C
A
Next
topic
will
be
argo
cd
roadmap
2.3
also
with
you
alex.
C
All
right,
okay,
so
I
basically
can
keep
sharing.
So
the
idea
here
was
that
there
is
a
pull
request
open.
It
was
just
to
refresh
our
roadmap
and
I
think
I
was
just
trying
to
put
on
the
roadmap
things
that
we
already
discussed
and
what
I
remembered-
and
I
think
that
what
I
wanted
is
to
get
feedback
here
and
basically
I
probably
missed
something.
Maybe
I
didn't.
I
forgot
some
features
that
we
were
planning
to
work.
C
Oh
okay,
and
there
is
no
time
constraint
that
all
the
changes
which
I
did
for
next
release
and
beyond
so
basically
like
I'm
just
looking
for
things
that
we're
definitely
going
to
do
and
not
necessarily
is
going
to
happen
in
next
quarter,
maybe
quarter
after
next
quarter,
but
something
certain-
and
this
is
what
I
come
up
with
so
one
is.
C
C
We
spoke
a
lot
about
merging
application,
sets
argus
identifications
and
image
updater
into
argo
cd
and
then
even
like,
even
if
some
of
these
projects
not
yet
ready
to
get
merged,
I'm
pretty
sure
it
will
happen
kind
of
soon
like
within
several
months.
I
think
that
upset,
maybe
will
be
merged
very
very
soon.
C
So
I
added
those
three
and
this
one.
I
don't
know
if
it
deserves
to
be
here
or
not,
but
I
I'm
judging
based
on
like
at
least
android
users
care
about
this
feature
a
lot.
So
we've
got
to
a
state
where
we
have
every
big
applications
managed
by
rbo
cd,
and
there
was
ideas,
discussed
kind
of
offline
about
how
to
improve
it,
and
one
of
these
ideas
was
to
have
a
compact
resource
tree.
I
know
that
alex
collins.
You
were
describing
this
idea.
C
I
think
we
discussed
it
like
ability
to
come
kind
of
to
to
collapse,
so
collapse,
kind
of,
say,
similar
resources
into
a
group
of
resources
so
and.
G
G
I
guess
two
scenarios
in
that
case
and
if
some
of
those
pods
are
unhealthy,
you
probably
want
to
be
able
to
expect
you
want
to
expand
them
and
investigate
the
unhealthy
pods.
But
if
they're
all
happy
and
healthy
and
in
sync
you
don't
need
to
to
look
at
them.
You
can
kind
of
shrink
them
down
into
a
single
resource.
I
can't
think
of
any
other
resources
that
are
replicated
like
that
yeah.
So.
C
That's,
I
guess
we,
you
know
the
discussion
can
happen
in
in
the
issue
itself
and
I
I
some
like.
I
felt
it's
important,
that's
why
I
just
edit
it
edit
it
here
yeah-
and
maybe
it
was.
This
item
was
added
when
I
started
getting
run
off.
Ideas
like
I
didn't
know
what
else
to
add,
and
but
there
are
two
more
important
kind
of
items.
One
was
multi-tenancy
improvements
and
we
kind
of
we
have
like
two
activities
in
progress.
C
One
was
one
is
a
proposal
open
proposal
to
support
applications
outside
of
argosy
namespace
and
second
is
ability
to
create
applications
using.
C
E
Yeah
I
mean
I
mean
this.
This
can
also
be
like
if,
when
we
go
ahead
right,
probably
there
may
be
some
more
items
to
put
on
v23
and
biond.
So
this
is.
This
is
not
stone
carved
right,
and
I
agree
that
yeah,
it's
it's
good
to
give
a
broad
overview
about
what's
happening.
C
C
A
All
right,
let
me
stop
sharing
okay.
Next
topic
hung
server
side
apply
with
you,.
H
I
think
this
coming
from
this
annoying
feature
is
called
manage
field,
because
every
time
we
see
the
manifest
we
see
this,
I
think
the
whole
purpose
of
have
this
is
trying
to
support
in
the
server-side
api.
I
don't
have
a
strong
proposal
yet,
but
I
think
this
is.
H
We
definitely
notice
that
the
flux
already
did
that
in
their
flux,
v2-
and
we
are
thinking
about,
evaluate
this
feature
and
to
see
whether
we
should
like
include
it.
One
thing
I
do
notice
that
is
like
in
argo
city
currently
for
the
manifest
applied
we're
using
the
coup
cdl.
We
need
to
bundle
a
couple
cdl
version.
Basically,
we
need
to
fork
a
process
and
to
do
the
apply,
which
will
do
the
three-milli
merge
for
us.
H
I
believe
we
can
just
use
an
api
to
give
in
all
the
manifest
to
the
to
the
net.
Is
api
server
and
on
in
the
server
side
it
will
do
the
three-way
merge.
I
think
we
are
banning
phase
to
keeping
our
code
simpler,
but
I
do
agree.
I
just
think
it's
there's
no
like
rush
here,
but
we
should.
I
think
we
should
keep
keeping
our
eye
on
this
feature
and
factor
that
into
our
roadmap.
H
I
do
know
that
I
feel,
because
there
are
a
lot
of
the
requirement
regarding
the
to
using
this
feature.
If
I
look
at
the
flux
server
side
applied,
I
do
notice
this
is.
It
seems
to
be
able
to
use
this
feature
due
to
the
managed
field
regression
you
have
to
be
using
our
latest
hot
patch
of
a
particular
kubernetes
distribution.
H
So
what
I'm
proposing
is
we
should
trying
to
actually
has
a
feature
flag
or
some
some
way
to
be
able
to
when
we're
doing
the
coop
city
applied
to
providing
this
dash
server
side
flag
to
tiptoe
into
that
feature.
To
make
sure
there
is
no
regression
problem
for
us,
so
we
won't
getting
the
full
benefit
of
the
feature,
but
I
think
we
should
go
with
the
kubernetes
direction.
That's
that's
my
proposal.
C
E
This
has
implications
right
this
this.
This
may
have
implications
on
on
the
result
that
we're
seeing
in
things
so.
E
H
That's
that's
their
idea
using
the
management
field
to
making
the
reconciliation
to
be
simpler.
Is
it
it
won't.
Does.
E
But
I
think
so
there
there
is
a
pretty
old
issue,
an
enhancement
issue
from
jesse
right
regarding
server-side
apply.
I
think
he
he
mentioned
it
like
when
it's
when
it
was
early
beta.
C
H
E
E
Yeah
it
does
it
it
changes.
It
changes
the
way
that
the
reconciliation
is
done.
So
all
these
fields
that
hong
showed
these
manager
fields
in
the
in
the
metadata,
so
multiple
controllers
can
manage
the
same
resource
but
different
fields
in
it
right.
So
you
you
could
have
like
a
controller
that,
like
the
the
deployment
controller,
which
creates
the
replica
sets,
owns
a
few
items
of
the
of
the
resulting
pot
and
another
controller,
could
manage
other
parts
of
of
the
ports.
H
The
young,
so
what
is
the
implication
to
us
means
because
we,
anyway
around
the
coop
city,
apply
anything
has
to
be
dramatically
changed
in
our
side.
Regarding
rocd
behavior,.
E
That
they
don't
patch
a
resource
but,
like
you
know,
replace
it
so
using
the
three-way
diff
like
it
is
now
so
yeah.
Basically
it's
patched,
but
yeah
and
server
side
apply,
would
just
manage
a
part
of
the
resource,
and
that's
that's
why
I
think
we
should
give
the
users
like
the
option,
use
server,
side,
apply
or
use.
Client-Side
apply.
H
C
C
H
Out
of
things
I
see
well,
is
that
causing
I
read,
the
documents
is
mentioned,
something
regarding
the
conflict,
so
I'm
just
wondering
whether
there
will
be
cases
we
need
resolving
those
conflict
also
from
our
side.
H
A
A
I
had
one
quick
presentation
I
can
make,
maybe
make
it
next
week
if
that
makes
sense,
so
anything
else.
Anyone.
A
No,
so
thank
you
and
see
you
next
week.