►
From YouTube: Argo Contributors Office Hours Jul 27th 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).
B
A
And
it's
starting
off
with
triage
and
discussions.
So
last
week
we
add
myself
anishita
for
doing
triage,
so
I
I
took
a
look
yesterday
at
the
issues
that
we
have
nothing
stood
out.
Just
one
person
open
an
issue
asking
if
we
could
use
if
we
could,
if
we
could
have
a
better
configuration
in
our
go
mod
in
Argo,
City
repo,
because
he's
using
Argo
City
as
a
library-
and
he
is
having
hard
time
with
the
way
that
we
expose
our
our
go
module.
A
So
that
was
an
interesting
one.
I
upvoted
I
think
it
makes
sense
when
he's
asking
and
it
would
help
future
users
who
wants
to
use
argosity
as
a
library
as
well.
So
nothing
blocks
them
for
doing
so,
and
we
could
support.
Could
we
could
have
a
better
support
for
that
use
case
so
other
than
that
only
usual
stuff.
A
lot
of
application
set
related
issues
IC,
but
nothing
in
particular.
So
that's
it
for
me,
I'm,
not
sure
if
Ishita
is
in
the
car
same.
B
A
Oh
thanks,
thanks
Ishita,
all
right,
so,
okay
moving
forward,
we
need
to
select
the
next.
The
next
group
of
people
who's
going
to
be
leading
the
the
triage
for
next
week.
So
on
any
volunteers
in
the
call
someone
for
primary
secondary.
A
Otherwise,
I
go
by
the
list
Jesse,
you
know.
B
B
John
McCall
Abhishek.
Are
you
in
the
call
deep.
C
A
Hello,
all
right,
okay,
with
that
out
of
the
way,
let's
move
on
for
the
topics
we
just,
we
just
have
one
from
Michael
Michael
want
to
leave.
E
This
I
was
just
reading
an
issue
yesterday
about
ignoring
differences
in
application
updates
and
since
I
was
recently
looking
at
the
ignore
differences
code
that
is
used
for
resources
and
applications,
I
thought
hey.
This
could
actually
be
pretty
easily
cross-applied
to
to
applications
in
an
application
set.
So
I
just
put
up
a
proof
of
concept.
Pr.
E
This
section
that
you're
currently
scrolled
to
Leo
is
the
section
of
the
docs.
That
explains
the
feature
so.
B
B
E
That
section,
it
jumps
kind
of
weirdly.
If
you
scroll
up
a
bit
yeah
yeah
here
it
is
okay.
So
basically
the
idea
is:
it
supports
a
subset
of
the
ignore
differences,
rules
or
fields
that
are
available
for
for
application
resources.
So
first
thing
I
did
was
I
changed
the
name
to
ignore
application,
differences
just
to
avoid
confusion,
I
added
the
Json
pointers
field,
the
JQ
Expressions
field
and
the
name
field.
E
So
you
get
the
usual
Json
pointer,
JQ,
expression,
support
and
I
added
name
just
in
case
folks
want
to
ignore
differences
for
a
single
app.
That
name
isn't
a
glob
I
have
the
feeling
people
are
going
to
want
to
template
that
field
or
make
it
like
a
label
selector
later
or
something
but
I
think
it's
a
fine
start.
I
didn't
add
anything
for
managed
fields
or
there's,
maybe
obviously
group
and
kind
or
hard-coded
Sims.
The
only
thing
we're
going
to
be
ignoring
are
applications
yeah.
So
that's
basically
the
feature.
E
It
goes
a
little
bit
against
the
like
hardcore
get
Ops
mode
of
application
set
because
it's
like
you
know
you
can
make
changes
and
they'll
actually
stick
and
the
applications
that
control
controller
won't
override
them,
but
I
think
it
makes
sense
to
be
pragmatic,
especially
since,
like
applications
that
basically
destroys
the
rollback
feature.
Unless
you
configure
this
because
if
you
do
a
rollback,
the
application
set
controller
is
just
going
to
come
right
back
through
and
reapply
the
changes.
E
Anyhow,
that's
the
feature
I'm
curious.
If
people
have
comments
concerns,
my
main
concern
in
the
implementation
is
I'm
doing
a
lot
of
marshalling
back
and
forth
between
Json
and
struct,
because
I
need
to
convert
from
like
an
application
struct
to
an
unstructured
to
run
it
through
the
normalizer.
Maybe
there's
a
better
way
to
do
that
that
I'm,
unaware
of,
but
you
know
it,
it
works
right
now.
So
that's
a
start.
E
A
Okay,
I
can
I
can
work
with
that
with
you
in
that
front,
if
you
want
but
I'm
just
curious
to
understand
what
is
your
strategy?
So
if,
if
you
configure
the
application
set
with
ignore
difference,
if
I
remember
well
the
up
the
applications
that
creates
idocity
applications
right,
can't
you
translate
the
same,
ignore
difference
that
you
have
at
the
application
set
at
the
application
resource,
and
then
you
get
that
feature
automatically
applied.
E
So
within
the
application
spec,
the
ignore
differences
field
refers
only
to
the
resources
which
are
managed
by
that
application,
not
the
application
itself,
so
I
could
add
a
rule
to
my
application,
spec
to
say:
okay
for
group
kinds,
Argo
project,
V1,
Alpha,
One
application
ignore
changes
to
the
syncpolicy.automated
field,
fine
that
that
configuration
is
not
going
to
have
any
effect
and
the
application
set
controller
is
not
going
to
respect
that
rule,
because
that
rule
is
only
with
respect
to
resources
managed
by
the
app
not
the
app
itself.
E
This.
This
new.
E
A
E
E
D
I
had
to
ask
a
question
so
I,
first
of
all,
I
I
think
this
is
a
really
great
feature.
I
think
it's!
You
know
it's!
It's
very
useful
and
I
I've
heard
from
from
people
that
are
using
external
tools
like,
for
example,
diagnostic
image,
updater
right
You.
You
basically
couldn't
have
unique,
dedicated
image.
Updater
annotations
on
on
the
applications,
manageable
applications
that,
because
the
applications
that
controller
would
overwrite
them
from
the
template
again
so-
and
this
potentially
is-
is
also
valid
for
for
other
third-party
external
tools.
D
That
would
somehow
modify
an
application
resource.
One
question
I
had
is
so
the
ignore
application
differences?
Will
it.
D
E
So
the
way
it
works
under
the
hood
is
we
we
have
this
utility
function
called
diff.normalize.
It
accepts
a
list
of
live
unstructures
and
a
list
of
what
what
it
calls
configured
unstructured
like
the
desired
State,
and
then
you
pass
in
your
diff
config
and
what
it
spits
out
is
the
the
desired
State,
excluding
any
changes
that
were
made
by
the
yeah.
E
One
thing
that
this
this
might
hide
this
feature
might
hide
an
issue
that
I
think
we
still
need
to
fix,
which
is
that
templating
is
still
pretty
Limited
in
application
sets.
So,
whereas
people
could
theoretically
have
a
field
and
say
get
files
generator
config
file
to
turn
automated
sync
on
or
off,
and
they
could
like
really
truly
get
Ops.
That
setting
since
templating
is
pretty
limited.
People
may
choose
just
to
ignore
differences
to
that
field
and
make
the
changes
solely
in
the
cluster
instead
of
in
get
so.
D
Well,
I
think
this
discussion,
like
you,
know
the
true
get
UPS
way
of
things.
I
guess
you
know,
there's
there's
a
portion
in
the
in
the
docks
in
the
office
of
the
docks
that
says:
leave
room
for
imperativeness
right
like
if
you
look
at
how
the
horizontal
Part
auto
scale
works,
for
example,
I
I!
Guess
it's
fine
to
have
some
some.
You
know
yeah,
leaving
that
room
for
imperativeness
somewhere,
so
yeah
yeah.
B
E
One
small
node
Leo,
the
ignore
differences
feature
in
general
has
bugs
when
it
comes
to
the
respect,
ignored
differences
equals
true
sync
option
and
I
think
it
has
to
do
a
lot
with
the
facts
that
we
have
like
home,
built
diffing
logic
when
we
actually
construct
the
patch
to
apply
the
normalized
diff
Blake
Pederson
is
trying
his
level
best
to
improve
our
diff
function,
but
I
think
that
he
might
benefit
from
your
help.
E
Since
you
know
a
little
bit
more
about
how
diffing
is
done
in
a
way
that
respects
like
what
do
you
call
them
like
field
keys
in
a
list
so
like
a
list
of
name
value
pairs,
our
logic
doesn't
currently
respect
stuff,
like
that.
So
I
can
point
you
to
the
pr
that
he's
working
on
and
kind
of
talk
you
through
that
I.
Don't
think
it's
going
to
impact
this
feature,
but
working
on
this
just
got
me
thinking
about
the
the
other
feature
as
well.
A
Okay,
so
you
mentioned
about:
is
this
the
one
that
we
have
here
in
the
dock.
A
A
Okay,
so
I
can
take
a
look
at
this
and
so
you're
working
on
this
blink
is
that
right?
Yes,
exactly.
F
As
I
put
this
in
I
guess
it
ties
into
the
next
next
topic.
Slash
question,
because
Mike
was
mentioning
that
the
gist
of
it
is
that
when
it
comes
to
array
entries,
they
need
to
be
keyed
by
like
emergency
like
by
their
name
usually,
and
the
gist
of
it
is
I
mean
trying
to
dig
through.
This
is
that
if
we
do
not
properly
group
array
entries
by
like
a
certain
merch
kit
in
some
cases,
basically
things
would
get
worse.
F
So
basically,
what
I'm
trying
to
do
is
what
I'm
working
on
now
is
trying
to
properly
group
the
entries
by
by
by
name.
But
then
the
question
leads
into
like.
Is
that
going
to
be
sufficient?
F
Are
they
I
mean
I'm
sure,
there's
going
to
be
more
edge
cases
where
this
may
or
may
not
work,
but
what
I've
been
doing
has
been
working
on
the
rain
trees
that
have
names
and
then
basically
grouped
by
by
names
in
in
order,
but
I
I
still
need
to
work
and
I
mean
I,
have
a
failing
test
case
in
that
PR,
which
I
will
need
to
get
working.
F
So
that's
that's
one
I
mean
I've
been
working
through,
but
there
was
this
guy
that
was
helping
out
with
some
some
test
cases,
which
was
very
helpful,
but
there's
definitely
going
to
be
more
dragons
on
the
way
so
I
will.
We
can
discuss
that
at
a
later
time.
E
Blake
just
curious
you're.
Are
you
hard
coding
name
as
like
a
emerge
key
or
you
do?
You
have
some
way
to
like
infer
the
merge
key
for
a
field.
F
E
A
A
Okay,
so
apparently
we're
gonna
be
working
together
in
a
few
hours,
late,
nice,
good
stuff,
all
right,
so
any
other
topic.
Anyone
any
last
minute
topic.
E
Test
the
RCs
I'm
going
to
push
a
bunch
of
stuff
today,
oh
Blake.
We
need
to
cherry
pick
the
the
other
problem
that
you
fixed,
you're,
fixing
a
lot
of
problems
lately
and
then
I
think
we
can
cut
another
2.8
RC,
which
will
contain
a
lot
of
fixes.
E
Least,
one
I
think
will
have
actually
knocked
out
all
of
the
blockers
by
this
next
RC,
which
means
we'll
have
like
a
full
week
of
hopefully,
the
final
RC
yeah.
F
That's
one
caffeine
particular
stairs
one
which
I
was
working
on,
which
has
been
really
frustrating
because
locally
running
the.
Let
me
see
if
I
can
grab
it.
This
is
the
one
where,
where
the
hell
in
the
repo
server
with
the
with
that
check
and
for
helm,
basically
to
to
verify
the
that
the
hell
project
for
the
helm,
repo
has
actual
permissions,
so
there's
PR
for
that,
but
the
issue
there.
Let
me
just
get
that
one
out
issuaders
that
I
have
yes
losing
Source,
not
permitted
hammers.
F
So
the
background
here
is
that
we're
on
a
Surface.
We
want
to
be
able
to
surface
home
repos
where,
where
the,
where
there's
not
actually
permissions
in
the
app
projects
to
access
this,
because
the
what
happens
is
before
it
ends
is
the
repo
server
the
repos
which
do
not
have.
The
permissions
are
filtered
out
problems
that
works
a
bit
too
well
and
I
have
a
fixed
this
one.
The
issue
is
that
I
have
in
this
one
is
that
locally
I
can
run
the
end-to-end
tests.
F
It
works
fine,
but
then
here
in
CI,
it's
not
which
has
been
incredibly
frustrating.
So
that's
so
either
there
needs
to
be
some
way
of
surfacing
this
either
surfing
this
surfacing
this
locally
or
actually
fixing
this
or
rolling
back
the
previous
PR.
So
that's
probably
something
else
that
needs
to
be
addressed.
E
Or
forgot
that
this
one's
a
blocker
I
will
help
you
with
the
ede
test
exactly
when
that
might
depend
we
should.
We
should
talk
and
make
sure
that
we
have
enough
time
before
August
7th.
Yes,.
B
F
A
All
right
thanks
thanks
Blake
any
other
topic.
Someone
wants
to
bring.
C
Yeah
there
is
a
two
PRS
that
I
have
mentioned
in
the
in
the
chat
right
now.
One
of
them
is
a
14th
through
sorry
fourteen
three,
eight
one
they
are
both
related.
So
this
is
about
so
the
the
subject
is
auto
respecting
airbag
for
this
copyright
and
sync.
C
So,
basically,
when
there
is
a
airbag
presenting
watch
of
some
type
of
resources,
the
idea
is
to
make
the
dynamic
plane
to
ignore
these
resources
to
be
watched
in
order
to
save
resources
on
the
API
server
side
and
also
on
the
application
application
controller
side.
C
Yeah,
so
I
have
been
investigating
some
performance
issues
that
are
creating
a
few
cases
that
you
have
received
from
customers
and
I
was
looking
at
this
one
as
a
potential
Improvement
that
we
can
improve,
that
we
can
introduce
sorry
I
think
that
Tiana
has
a
started.
The
review
of
this
PR
and
yeah
I
wanted
to
raise
your
attention
about
this
one
as
a
potential
proposal
for
performance.
C
Improvement
I'm,
not
the
author
of
the
the
author
of
The
VR
of
the
two
PRS,
but
this
is
something
that
can
be
interesting.
I
have
done
a
few
tests
on
my
side
with
the
with
the
build
of
this
one,
but
until
now,
I'm
I'm
I'm
not
able
to
to
have
a
stable
cluster
to
test
out
properly
every
time
that
I
try
to
increase
the
load
of
my
cluster
I'm
breaking
the
cluster,
so
I,
don't
I,
don't
think
it
is
related
to
this
specific
PR,
but
I
would
be
interested
about.
C
I
would
be
interested
in
a
review
of
of
your
faults.
What
for
this.
A
Yeah,
looking
at
the
changes
in
GitHub
singing,
it
introduces
a
client
in
a
closer
cache.
Yeah
is
I
think
this
is
to
be
reviewed
carefully,
so
it
does
not
introduce
performance
issues
or
anything
like
that.
C
Okay-
and
you
know
what.
C
The
impact
human,
so
I
I
think
that
it
would
improve.
Yeah
Ishita
is
mentioning
an
another
way
to
improve
performance,
but
the
impact
would
be
that
there
would
be
less
of
dynamic
clients
watching
some
specific
resources
that
would
be
created
and
maintained
alive
on
the
application
controller
side.
So
we
should
have
less
throttling
happening
on
the
client
side
and
we
should
also
consume
less
resources
on
the
API
server
side.
C
The
thing
that
I
was
wondering
also
and
I-
didn't
see
that
in
the
code,
exactly
but
someone,
maybe
with
a
better
knowledge,
can
can
see
that
if
a
resource
is
ignored
from
being
watched
by
this
current
PR
and
it's
excluded,
I
I
think
that
the
exclusive
the
exclusion
is
final.
So
even
if
we
add
the
permissions
later
on,
I,
don't
see
it
as
being
reintroduced
as
a
potential
watch.
A
C
No,
no
I'm
I'm,
just
like
a
collateral,
tester
I,
would
say,
because
it's
I'm
interested
in
interested
in
having
some
improvements
in
performance.
So
I
was
looking
into
this
other
potential
helpers.
A
B
A
B
C
A
Sure
all
right,
thanks
for
bringing
this
up
yeah,
it
looks
a
good
one.
E
If,
if
I
review
this
it's
going
to
be
kind
of
late
in
the
2.9
cycle,
I'm
not
gonna
have
time
for
a
while.
So
we
may
have
time
around
and
see
if
we
can
find
another
final
reviewer.
C
A
Okay,
any
other
topic:
someone
wants
to
bring.