►
From YouTube: Argo Contributors Office Hours May 25th 2023
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
Hi
everyone
welcome
to
this
week's
contributors.
Meeting
I'm
gonna
be
your
host
today,
I'm
Leo
and
I'm
sharing
my
screen
and
starting
off
with
triage
in
discussion.
A
So
I
was
the
one
signed
for
triage
last
week
and
jaydeep
as
well.
Jdp
is
in
the
call
thanks
JDP,
so
I
what
I
did
I
added
the
the
bullets
that
we
used
to
have
for
all
the
stats
and
Michael
suggested
to
add
a
few
more
with
the
top
reviewers
and
top
mergers
for
the
week,
which
is
pretty
cool
so
yeah.
A
So
previous
week
we
had
101
open
issues,
32
close,
so
we're
we're
a
little
bit
out
of
balance
there
in
terms
of
PR,
we
had
24
open
15
closed,
which
is
which
is
better
and
still
issues
which
means
issues
not
updated
for
the
last
year.
We
are
on
the
number
of
101
issues
in
this
condition,
which
is
which
requires
some
some
action
here.
We
we're
aware
of
that.
A
So
we
just
need
to
to
yeah
plan
how
we're
gonna
address
that
so
for
top
reviewers
we
have
Michael
Alex,
Blake
C
Santana
in
Buddha
Mubarak
and
top
mergers,
Michael
Alex,
Keith
and
Jesse.
For
the
last
week,
I
provided
the
links
to
the
Dev
stats.
So
you
can,
you
can
see
by
yourself
if
you
want
and
that's
pretty
much
it
for
three
years
of
the
week
for
myself
JDP,
you
have
anything
to
add.
B
Thanks
Leo
no
another
one,
unfortunately
I
wasn't
able
to
find
as
much
time
to
triage
this
week.
Okay
I
saw
some
issues
this
morning,
but
I
wasn't
able
to
spend
as
much
time
to
decide
if
they
needed
to
be
brought
up
here,
but
yeah.
Sorry
about
that.
No.
A
Problem
I
reply
to
a
few
a
few
issues:
I
wasn't
able
to
close
them,
but
but
nothing
that
worth
mentioning
is
yeah,
I,
think
it's
standard,
triage
and
stuff,
and
now
we
need
to
decide
who's
going
to
be
taking
the
the
role
for
next
week.
So
any
volunteers
for
primary
and
secondary.
A
I
can
take
primary
thanks,
then
all
right,
any
volunteers
for
secondary.
A
I
can
be
secondary.
It's
Keith
here,
thanks
Keith,
all
right,
okay,
so
thanks
everyone
moving
forward.
We
have
two
topics
for
the
day,
both
from
Michael
so
preview,
environments
for
Argos,
cdprs
Michael,
you
wanna,
say
a
few
words
about
it.
D
Yeah,
if
you
would
pop
open
that
issue-
or
actually
it
was
just
a
few
hours
ago,
a
company
called
bunny
show
basically
offered
to
help
us
set
up
preview
environments
for
our
PRS
I,
don't
know
anything
about
bunny
shell.
Their
website
has
the
cncf
logo
on
it,
which
seems
good,
I.
Think
that
having
a
PR
preview
environments
would
really
help
us
move
things,
especially
like
the
issue
author
said
UI
issues,
because
you
can
just
quickly
test
it
rather
than
you
know,
changing
branches
locally,
pulling
down
someone
else's
remote
building
it
running
it.
D
So
I
think
this
could
be
a
huge
win.
I
know
that
they're
probably
like
issues
with
making
sure
everything's
cool
cncf,
making
sure
the
maintenance
burden
on
us
will
be
minimal,
but
at
a
glance
it
looks
awesome.
D
D
C
D
A
B
D
E
Yeah
I
think
it
could
also
be
beneficial
like
if
you
know,
depending
on
how
customizable
those
those
environments
are
right,
if
I
mean,
if,
if
they,
for
example,
come
with
something
like
a
metrics
collection
and
visualization,
so
you
could
also
validate
whatever
change.
There
will
be
to
the
core
right,
I
mean
for
reconciliation
and
stuff
like
that.
If
and
if,
if
we
get
access
to
the
cluster
from
the
outside
as
well
like,
you
know,
Cube
CDL
access,
basically
I,
but
I,
don't
know.
E
E
This
is
basically
something
I
wished
for
a
long
time.
You
know
someone
submits
a
PR
and
yeah,
not
not
only
for
the
UI
I
mean
the
UI.
They
can
supply
some
screenshots
too
right,
but
for
for
everything
else
that
needs
validation
like.
A
For
the
reviewer
right,
so
the
reviewers
can
test
things
without
having
to
check
out
the
the
branch
and
deploy
and
build
everything
in
their
laptops.
You
save
some
time.
F
F
Oh,
do
they
I
think
for
open
source
and
projects.
They
do
provide
that
facility,
at
least
that's
what
they
started
with
I.
Don't
know
if
they
change
the
terms.
F
C
C
D
F
C
A
Them
to
provide
a
POC,
and
so
so
we
can
take
a
look
and
see
how
it
looks
like
yeah.
A
It
does
you're
right
cool
all
right,
so
moving
forward
next,
one
plug-in
generator.
Is
it
your
GD
project.
D
That
is
so.
This
is
the
pr
that
ashitosh
and
I
validated
in
our
GE.
So
for
folks
who
aren't
super
familiar
with
into
its
internal
stuff,
we
have
a
thing
called
Global
Engineering
days
or
we
get
a
week
to
just
kind
of
hack
on
something
so
ashitosh
and
I
hacked
on
creating
a
plug-in
generator.
F
A
C
We
did
their
their
project
was.
Basically,
it
was
a
pretty
neat
thing
that
we
did
is
we
called
an
internal
Intuit
API
that
got
a
list
of
Intuit
clusters
and
basically
used
instead
of
using
apps
that
it
was
a
dynamic.
D
D
Says
that
gave
a
good
summary.
The
the
main
reason
I
wanted
to
put
it
on
the
contributor
meeting
agenda
is
I,
feel
really
good
about
the
architecture
and
the
way
it
works
like
we
validated
that
it
actually
runs.
I
would
love
if
Jan
or
Zach-
or
you
know,
someone
who
is
kind
of
knows
or
go
see.
These
overall
architecture
would
take
a
look
and
just
do
a
coherence
check.
Make
sure
that
you
know
we're
not
going
to
lock
ourselves
into
a
system
that
is
problematic
in
the
future.
D
So
that's
the
main
thing
I
want
it's
just
a
second
second
solid
review,
not
of
the
code
but
of
the
architecture.
D
A
nitpicky
thing,
I
wanted
people's
feedback
on
is
Leo.
Can
you
scroll
down
to
Center
the
code
in
simple
example?
So
there
are
several
things
you
use
to
configure
a
plug-in
generator.
One
is
a
config
map,
ref
that
config
map
just
holds
like
the
URL
that
you
call
to
get
the
parameters.
D
Another
thing
is
we:
have
we
used
the
word
parameters,
a
lot
in
application
sets,
and
it's
actually
something
that
is
Thoroughly
frustrated
me
about
Argo
workflows
is
there
are
five
different
ways
we
use
the
word
parameters,
so
I
want
to
see
what
people's
feelings
are
about.
Using
this
phrase,
input
parameters
to
describe
okay
here
are
the
things
that
we
send
to
the
plug-in,
the
the
rest,
API
and
output
parameters,
to
describe
the
things
that,
like
the
the
plugin,
sends
back
to
you
that
you
use
to
actually
write
into
your
template.
D
D
D
C
D
D
Yep,
that's
fair
and
I
think
this
is
consistent
within
CD.
The
only
other
place
we
use.
The
word
parameters
like
in
yaml
like
this
is
applications
spec
Dot
source.helm.parameters,
and
that's
very
much
like
a
Helm
concept.
It
has
The
Hyphen,
P
flag,
so
I
think
it's
fine
to
drift
away
from
Hell
A.
C
C
F
E
If
so
did
we
did
we
decide
on
arguments
because
I
I
thought
you
you
have
like
the
concern
about
the
camel
case
for
input,
parameters
and
output
parameters
right
and
how
about
I
mean
we.
We
could
also,
like
you
know,
have
a
Fiat
input
and
below
that
input
we
have
parameters
yeah.
So
that
would
just
move
the
parameters
down
one
level
and
then
we
we
could
have
input.
E
E
D
D
D
D
D
D
No,
so
that's
just
what
the
HTTP
server
responds
with
there's
a
it
expects,
so
the
applications
that
controller
calls
the
plug-in
API
and
the
controller
expects
to
receive
a
Json
objects
with
a
top
level
key
being
output
parameters.
Right
now,
that's
the
only
key.
We
just
left
waiting
for
it
to
return.
A
It
Michael
quick,
quick
question:
is
there
a
PR
that
we
can
continue
this
discussion
asynchronously
there.
A
A
Okay,
that
summarizes
the
the
topics
for
the
day
any
last
minute
topic.
Someone
wants
to
bring.
B
I
had
a
quick
question,
so
I
noticed
that
we
got
the
273
release,
for
which
picks
up
I
think
the
deadlock
fix
that
we
discussed
on
slack,
but
I
haven't
seen
if
we've
done
a
two
six
or
two
five
for
that.
Yet
is
there
something
we're
waiting.
D
I
kicked
those
off
I
kept
hitting
errors
and
the
actions
I
never
checked
whether
my
last
attempt
worked
though,
and
it
looks
like
it
didn't,
let
me
see
what
went
wrong.
There
failed
to
push
protected,
Branch
yeah
a
while
back
I
protected
the
release
branches
because
technically
for
salsa
compliance
we
need
to,
but
apparently
the
branch
protection
is
breaking
our
old
build
pipeline.
So
let
me
I
need
to
fix
that.
D
I,
don't
know
what
branch
protection
rule
it's
breaking
right
now:
yeah
I'll
I'll
try
to
kick
another
another
release
for
two
five
and
two
six.
In
a
few
minutes.
H
Yeah
myself
have
a
cooking
PR
this
one.
Regarding
the
cluster
sharding
algorithm,
there
were
some
changes
requested
and
I
have
done
that
the
pr
is
still
building.
It
seems
wrong.
H
Yeah
and
I.
Think
I
have
been
some
some
folks
and.
A
F
A
H
Okay,
I'll
just
send
the
link
and
yeah.
Two
weeks
ago
there
was
some
change
requested
and
I
have
done
that
last
week
and.
C
H
And
I
added
documentation
also
because
it
was
a
asked
like
specifying
that
the
this
new
feature
is
experimental
and
the
default
is
XYZ
and
so
on.
Cool
I'll
take.
D
A
G
Yeah
quickly,
yesterday,
I
found
a
new
PR
about
adding
go
fmt
to
the
make
lint
should
be
a
pretty
straightforward.
One
I
was
trying
to
use
golang
City
Island,
which
is
already
implemented
in
in
the
making
step
to
run,
go
fmt,
which
it
normally
supports.
Yeah,
that's
the
BR
and
I.
Couldn't
convince
the
thing
to
actually,
you
know
show
the
diffs
of
of
what
it
would
do
of
all
the
simplifications
and
and
formatting
changes.
G
G
So
two
things:
if
anyone
has
time
to
just
like
give
a
look
at
this
PR
should
be
a
very
straightforward
one,
and
the
second
thing
is:
if
anyone
knows
figures
out
how
to
use
Google
like
Ci
lent,
instead
of
using
a
direct
ego
fmt,
which
you
know
would
be
more
convenient
again.
I,
don't
think
it's
possible,
but
maybe
I've
met
something.
D
G
But
I
was
since
going
CI.
Lint
is
used
already
the
line
above
where
I
inserted
my
script
in
the
make
file
I
was
trying
and
since
it's
also
supported,
he
was
trying
to
use
good
Google
links,
chat,
name,
I'm,
struggling
to
say
this
word.
Every
time
goding
see,
Island
I
was
trying
to
use
reuse
it
for
that.
But
the
other
thing
it
will
do
is
rewrite
the
changes.
It
sees
the
the
fixes
it
sees
right.
G
It
will
not
produce
an
eight,
a
a
diff
and
I
looked
at
the
code,
and
it
seems
they
don't
implement
the
dash
d
parameter
for
for
go
fmt
in
including
silent,
so
they
don't
have
an
option
for
that.
So
I
think
it's
I
mean
from
what
I
see
it's
not
doable,
but
but
I
would
just
love
to
reuse
it.
If,
if
you
know,
if
we
can
in
any
case
What
I've
Done
instead
is
just
write
a
little
script
and
edit
it
in
the
hack,
subdirectory
and.
H
G
Mean
directorate
and
what
it
does
is
simply
produces
a
diff,
and
if
that
diff
is
on
it,
then
it
shows
it
and
it
errors
up.
A
Does
yeah
and
the
problem
is
if
we
diverge
between
the
two
tools,
we
might
get
yeah
differences
right,
because
golang
CI
runs
in
CI
during
during
our
CI
job.
So
I'm
not
sure
if
having
both
tools
in
parallel
is
a
good
idea
or
if
we
should
dedicate
time
to
see
why
galencia
is
not
doing
what
you
wanted
to
do
so.
G
A
G
A
G
G
A
Or
maybe,
if
yeah
I,
don't
remember,
golang
side
generating
diff
at
all,
maybe
it
doesn't,
it
just
fails
and
provides
a
a
message.
Saying
hey.
This
is
not
following
the
the
the.
A
G
Yeah,
so
what
I?
What
I've
done
is
I've
before
I
did
wrote
this
script
to
use?
Go
fmt
I
modified
the
make
file,
so
so
we're
not
using
a
a
config
for
Google
anxiety,
Island,
we're
just
passing
parameters
on
command
line,
so
I
just
added.
You
know
for
convenience,
I
added
a
config
file
and
just
added
the
enable
the
go
fmt
linter
in
there
and
it
it.
It
did
just
that
it
just
like
it
rewrites
the
the
all
the
stuff
that
it
sees
that
that
should
be
improved.
A
D
I
think
the
CI
Behavior
should
be
consistent
with
the
make
file
behavior,
which
may
mean
dropping
the
golang
CI
lint
action
in
favor
of
just
a
plain
old
run,
step
that
runs
precisely
what's
in
the
make
file
because
I
agree,
diverging
CI
versus
local
Dev
experience
is
probably
a
bad
thing.
G
A
I
think
Michael
there
it's
a
little
bit
more
complicated
than
that,
because
golang
CI
goes
beyond
go
fmt.
It
does
a
lot
more
than
that,
and
this
is
what
we
run
in
CI,
so
I.
If
I
remember
well,
there
is
a
make
Target
that
allows
us
to
run
golang
CI
in
a
Dev
environment
in
our
laptops.
A
What
I'm
saying
is
I
agree
that
it
should
be
centralized
in
the
in
the
mid
Target,
whatever
we
do,
and
it
should
be
one
single
make
Target
that
does
all
linked
linked
actions.
We
want
in
the
project
and
if
we
start
spreading
things
in
different
places,
I
think
yeah.
It's
going
to
be
hard
to
maintain
in
the
future.
So
whatever
that
is,
let's,
let's
centralize
in
a
in
a
single
make
Target
and
yeah
and
then
go
from
there.
That's.
D
G
I
I
think
the
other
thing
I
wanted
to
add
to
that
is
that
as
good
NCI
island
is
used
today,
it
only
uses
the
default
printers,
which
are
not
very
many.
D
As
this
PR
goes,
I
think
that,
like
so
Leo
wants
everything
centralized
under
one
make
Target
I.
Think
that's
done
in
this
PR
I
think
that
we
also
need
CI
to
be
consistent
with
local
Behavior,
which
I
think
means.
We
need
a
new
CI
step
to
run
this
hack
script
or
something
else
basically
to
make
sure
that
there's
no
go
format
diff
before
PR
gets
merged
again.
D
D
G
A
Okay,
I
guess
that
wraps
up
for
today,
if
there
are
no
other
topics,
I
will
give
you
those
eight
minutes
back
and
see
you
all
next
week.
Thank
you
thanks.
Everyone
bye.