►
From YouTube: Argo Contributor Experience Office Hours 10 Sep 2020
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
Okay,
I
I
think
we
can
start
like
16
contributors
already
and
I
will
be
surprised
if
we
get
several
more
okay.
So
my
name
is
alexander
and
I'm
a
software
engineer
at
intuit
and
I'm
a
contributor
and
maintainer
of
argo
cd
project.
I
used
to
work
on
argo,
workflows
and
rollouts,
but
right
now
I'm
almost
100
focused
on
argo,
cd,
okay
and
before
I
speak
too
much
can
I
get
thumbs
ups.
A
A
So
I
even
prepared
a
small
presentation
so
we'll
work
through
well,
I
will
try
to
give
you
overview
of
argo,
cd
architecture
and
components,
and
then
you
know
a
few
advisors
about
how
to
get
started,
how
to
set
up
your
dev
environment,
but
even
before
going
into
that,
I
would
like
to
propose
to
everyone
to
introduce
yourself
so
and
just
maybe
give
your
you
know,
give
your
name
and
maybe
a
couple
words
about
why
you're
interested
in
contributing
to
our
city
and
how
much
time
you
want
to
you
know,
spend
like
every
week,
let's
say
to
contribute:
okay,
and
we
have
a
lot
of
people,
so
I
will
try
to
kind
of
organize
the
process.
B
Oh
yeah,
sorry,
I
guess
I'm
the
only
william
yeah
yeah.
Yes,
I
work
for
a
wet
head,
so
I
think
we
have
a
few
other
folks,
also
from
whitehead.
We
we
are
in
the
git
ops
group
and
we
are
interested
in
contributing
to
other
cities
as
because
we're
gonna
be
using
it.
Quite
you
know
it's.
It
is
an
important
piece
of
of
our
power
strategy
and
yeah,
so
I'll,
be
probably
devoting
close
to
full
time
on
on
contributing
to
this
cd
project.
B
Oh
me,
so
I
worked
on
so
I
work
on
okay
youtube
a
little
bit,
that's
the
service
less!
So
I
I
joined,
I
said
so.
I
started
this
kubernetes
work
roughly
about
two
years
ago,
so
I
first
worked
a
little
on
the
techcon
and
then
some
sims
and
then
about
about
a
year
on
on
on
on
serverless
and
yeah
and
then
and
then
I
I
still
you
know,
start
looking
at
this.
You
know
roughly
about
half
a
year
ago,
like
the
githubs
in
general
yeah.
B
So
so
it
has
been,
has
been
all
open
source
experience
for
me
ever
since
I
joined
red
hat
and
I've.
B
Like
before
I
was,
I
was
involved
in
like
apache
stuff,
a
long
time
back
like
more
than
10
years
ago.
Yeah
right,
cool
thanks,
so
yeah
be
excited
to
join
the
community.
C
E
Yes,
hello,
everyone,
I'm
darshan,
chaudhary,
I'm
a
software
developer
based
out
of
bangalore,
india.
I
work
for
this
company
called
gojek,
it's
a
ride-hailing
plus
plus
company,
so
we
basically
do
radially
full
delivery.
E-Money
wallets
and
all
those
things
like
we
have
17
services
in
a
single
app.
It's
a
super
app
of
sorts.
E
E
E
C
Involved
with
the
argo
project
from
the
beginning,
obviously
it
started
when
actually
alex
and
a
bunch
of
other
people.
Sorry,
let
me
close
the
window.
C
I
Sure,
hey
everyone,
my
name
is
chetan.
I
work
at
red
hat
in
india
from
bangalore
office,
so
I
joined
red
hat
last
year
and
so
I'm
excited
to
contribute.
A
I
J
Sorry
is
it
me
alex
yes,
yes,
okay,
hi
everyone,
I'm
mei
meijong,
I'm
a
software
engineer
at
intuit
and
full
time
on
argo
city.
It.
A
Thank
you
next
is
jadeep.
H
No,
that's
fine!
Thank
you!
Hi!
I'm
jaydeep,
I'm
also
a
tradar,
william
and
the
others
as
a
part
of
the
get-offs
team,
so
I
started
as
a
currently
an
intern,
so
I
started
working
around
the
summer
and
I
haven't
had
a
lot
of
experience,
working
with
open
source
and
contributing
upstream
so
far.
H
So
this
is
like
a
interesting
opportunity
for
me
to
get
a
taste
of
that,
along
with,
like
all
the
kubernetes
stuff,
that
I've
been
working
on
at
red
hat
as
a
part
of
this
internship,
so
just
generally
very
looking
forward
to
exploring
as
much
of
the
space
as
I
can,
and
you
know
getting
better
at
like
open
source
contributions
in
general.
So
I'm
looking
forward
to
that.
Thank
you
for
having
me.
A
Welcome:
okay:
next
is
alek.
K
Hey
everyone
yeah,
I
I'm
oleg
from
codeflash.
M
N
You
can
see
I'm
full
time
on
days
so
before
that
I
joined
multiple
startups
and
I
worked
with
alex
jesse
and
mukulika
in
the
athletics.
So
we
started
our
group
project,
so
my
main
contribution
is
actually
in
argo,
workflows
and
currently,
I'm
the
manager
so
feel
free
to
connect
with
me.
If
you
have
any
questions-
and
I
can
help
you
thank
you.
B
N
O
Oh
yeah,
so
hey
everybody
nice
to
meet
you
all,
I'm
from
red
hat.
I
was.
I
started
interning
at
jan
and
it's
almost
been
eight
months
now,
which
I've
had
so
I'm
on
williams
team
as
well.
I'm
working
on
github
sometimes
we're
creating
our
own
github
store,
so
we
showcase
some
interest
towards
argo
cd
because
we
kind
of
use
it
for
our
deployments.
O
So
I
think
I
raised
my
first
pr
today
to
master
so
I'm
looking
forward
to
creating
it's
an
interesting
workspace.
I
mean
I've
always
thought
about
contributing
to
like
an
upstream
project
in
the
open
source.
So
this
is
my
first
opportunity
at
it
and
I
think
I'll
be
working
full-time
towards.
So
thank
you
all
for
the
opportunity
and
look
forward
to
working
with
you
welcome.
P
P
Yes,
so
I
work
for
our
trip
advisor,
so
I'm
a
person
principal
software
engineer
for
our
ml
infrastructure
team.
So
we
use
our
argo
cd
and
argo
event
hardcore
flow
for
our
ci
cd
heaven.
So
as
far
as
now,
it
works
perfectly.
So
I'm
glad
to
join
the
community.
P
P
So
I
think
I
can
like
few
hours
a
week,
so
many
is
problem
driven.
So
I
think
so.
I
already
maybe
like
create
some
issues
to
the
project
and
also
I
contribute
one
commit
back
to
the
article
workflow.
F
F
N
A
Yes-
and
I
kind
of
I
think,
is
too
she
understates
your
contribution.
Q
Hey
my
name
is
jonathan.
I
work
with
william
at
red
hat.
Like
tim,
I
actually
joined
red
hat
this
month
and
also.
D
Q
R
Hello,
I'm
regina
I've
been
with
red
hat
since
june
this
year.
I
would
be
out
of
the
westford
office,
but
I'm
working
at
a
greater
boston,
I'm
generally
still
really
new
to
the
tech
industry,
and
my
background
is
actually
in
public
health,
so
I'm
great
to
test
out
new
user
experience,
things
with
so
still
learning
a
lot
about
argo,
but
really
looking
forward
to
contributing
to
growing
the
community
and
documentation
and
any
other
any
other
pull
requests
testing
things
out.
I
would
love
to
do
that
so
looking
forward
to
contributing.
A
Thank
you,
cakes
he's
next.
S
Hi
I'm
keith
chong.
Similarly,
with
tim
and
jonathan
joined
red
hat
last
week,
you
know
work
with
william
a
little
bit
of
background.
I've
contributed
to
several
open
source
projects,
including
the
eclipse
web
tools,
platform
or
wtp
and
coldwin
and
double
with
the
open
api
generator,
including
providing
the
ide
support
for
vs
code
and
clips,
so
yeah
I'll
be
working
full
time
on
this.
A
Okay,
next
contributor
is
right
now.
D
Hey
guys,
this
is
weibo.
I'm
software
engineer,
blackrock,
I'm
contributor
maintenance,
fargo
events,
so
if
you
guys
had
any
doubt
or
you,
you
guys
want
to
contribute
on
our
weapons
feel
free
to
connect
with
me.
Thanks.
T
Sure
turn
my
video
on
for
a
sec,
so
hey
everyone,
I'm
dan.
I
work
for
factset
research
systems,
although
I
contribute
as
an
individual
contributor,
I'm
pretty
much
going
to
be
doing
part-time
stuff
on
the
weekend.
T
U
Alex
hi
everyone
great
to
see
everyone
here
today,
at
the
first
meeting
my
name
is
alex.
I
am
principal
engineer
on
argo,
workflows
and
events.
That's
my
focus
a
moment.
I
have
previously
worked
on
argo
cd,
including
implementing
the
helm,
support
and
the
sync
waves
support.
A
Thank
you,
okay.
Thank
you.
It's
like.
I
noticed
that
again
we
didn't
get
too
personally
the
same
name
except
you
know
me,
alex
and
and
alex
collins
and.
D
A
We're
going
to
have
the
same
confusion,
we
usually
go
by
alex
m
alex
c,
and
you
know
even
in
meetings.
We
sometimes
use
this
acronym.
So,
okay,
so
welcome
everyone,
and
I
didn't
really
know
how
much
time
introduction
will
take.
It
took
way
longer
than
I
expected,
because
of
so
many
people
joined.
A
But
still
I
prepared,
like
a
small
presentation
for
this
first
meeting,
just
to
get
started,
and
I
will
try
to
you
know,
be
sure
to
leave
some
time
for
questions
and
feel
free
to
just
interrupt
me.
You
know
in
the
middle
of
presentation.
I
can
answer
questions
right
away.
Let
me
share
my
screen
first
and
yeah.
A
You
know,
architecture
and
then
touch
development,
environment,
setup
kind
of
the
minimum
set
of
things
you
have
to
do
to
start
contributing
and
then
walk
through
the
you
know
how
to
how
we
organize
tickets
and
how
you
can
use
this.
You
know
label
system
labeling
system
to
find
ticket
which
you
know,
which
you
want
to
contribute.
A
First,
we
need
to
get
manifests.
Then
we
need
to
take
a
look
at
those
manifest
and
compare
it
with
kubernetes
cluster
and
then
we
need
to
apply
changes.
So
it's
extremely
simple
and
that's
why
a
lot
of
users
even
question
the
need
for
githubs
too,
and
I
think
they're
right.
So
if
it's
so
simple,
then
why
automate
it
at
all
and
each
gitobs2
kind
of
has
you
know
different
answer,
and
I
personally
believe
that
argosy
answered
to
that
question
like
why
tool
is
needed.
A
So
it's
a
githubs
tool
for
enterprise,
and
I
wanted
to
mention
it
because
it
explains
that
why
argo
cd
is,
you
know
what
it
is
so
because
it's
for
enterprise.
That
means
you
automate
these
three
simple
actions
for
thousands
of
developers,
and
that
means
you
know,
argo
cd.
It
adds
some
complexity,
you
need
to
run
it
but
at
the
same
time
you're
helping
to
thousands
of
people.
So,
even
if
you
help
a
little
bit,
this
impact
is
still
big,
and
so
that
explains
why
argo
cd
is
needed.
A
A
So
first
one
is
argo,
cd,
repo
server
and
this
component
is
responsible,
for
you
know
talking
to
git
and
extracting
manifest
from
git.
So
in
real
life,
it's
not
as
simple
as
just
call
on
the
repository
and
read
files.
Typically,
you
want
to
cache
who
won't
you
positive,
so
repo
server
kind
of
takes
care
of
it.
It
clones
repository
once
and
put
it
into
a
local
file
system,
and
then
we
use
it
for
subsequent
calls
and
so
a
lot
of
users
don't
want
to
just
use
plain
gamma
files.
A
Instead,
you
just
want
to
use
something
like
helm,
customize
jsonnet
case
on
it,
which
is
dead
already,
but
still
a
lot
of
users
use
case
on
it,
and
there
are
more
tools
and
repo
server
allows
to
connect
additional
tools
using
just
cli
integration,
so
you
can
configure
it
to
invoke
a
script
which
does
whatever
you
want
it
to
do
and
it
will
produce
manifest,
and
you
can
see
a
lot
of
boxes
here,
and
these
boxes
means
like.
Basically,
you
can
scale
that
component
and
in
real
life.
A
You
actually
want
to
scale
it
because,
especially
if
you
use
config
management
tools,
the
process
of
manifest
generation
is
cpu
consuming
memory
consuming-
and
this
is
right
now
kind
of
I
would
say,
focus
of
our
next
release.
We
want
to
make
that
component
even
more
kind
of
performant.
We
wanted
to
scale
better
and
we
want
to
empower
each
replica
of
equal
server
to
be.
A
A
Okay
feel
free
to
interrupt
me
if
you
want
to.
You,
know,
stop
and
ask
questions
about
components.
So
if
there
are
no
questions,
let's
move
to
the
next
one
rcd
application
controller.
So
this
is
maybe
one
of
the
most
challenging
pieces
of
argo
cd
because
it's
doing
the
most
difficult
work.
A
So
the
responsibility
of
controller
is
to
compare
changes.
Compare
manifests
through
changes
manifest
from
weapon
server
with
the
live,
kubernetes
resources
detect
the
difference,
and
then
let
users
take
a
look
at
the
difference
and
diffuser
proof.
So-Called
thinking
push
the
changes
into
a
cluster
so
and
it's
kind
of
difficult
to
apply.
A
But
the
challenge
here
is
that
controller
manages
a
lot
of
hundreds
of
clusters
so
and
hundreds
of
clusters
means
you
have
to
compare
a
lot
of
resources
and
the
problem
is
how
to
do
it
very
fast
just
to
give
you
better
understanding
into
it.
As
a
user
of
argo
cd
and
we
use
ergo
cd
to
manage
200
plus
clusters,
and
we
we
have
some
metrics
about
how
many
objects,
how
many
sources
we
need
to
compare
and
you
know,
push
changes
to
so
last
time.
A
A
A
Okay
and
this
little
box
here
is
a
radius
and
I
feel
like
radius
becoming
more
and
more
important
part
of
fargo
cd,
it's
mostly
again
because
it's
an
enterprise
tool,
so
it
manages
a
lot
of
data
and
we
want
it
to
be
very
reliable
since
it
have
a
lot
of
users
we
want
to.
You
know
every
every
component
must
be.
A
You
know
it
should
be
possible
to
upgrade
to
the
downtime
and
we
rely
more
and
more
on
the
on
radius.
For
example,
in
in
most
recent
release,
we
kind
of
we
start
moving
some
data
out
of
kubernetes
crg,
we'll
talk
later
about
clds
into
radius,
to
improve
performance
and
so
controller
in
particular.
So
every
every
result
of
work
of
controller,
which
user
potentially
interested
in
is
saved
into
radius
and
controller
itself.
A
It
doesn't
have
any
api,
so
it's
impossible
to
talk
to
controller
directly,
but
controller
kind
of
present
data
to
user
through
radius,
and
it
takes
input
from
some
kubernetes
clds
and
next
we
can
move
to
clds
and
to
the
last
component,
which
is
argo,
cd,
api
server
and
ips
server.
As
you
can
understand
from
the
name.
It's
a
basically
gateway
into
rbo
cd,
so
ips
server.
A
It
contains
as
well
ui
and
set
of
ipis,
which
powers
that
ui
as
well
as
cli,
and
so
basically
it
allows
you
to
see
a
list
of
argo
cd
applications
and
the
state
of
every
application
and
application
is
stored
as
a
customer
source
in
control
in
control
plane
cluster.
So
this
is
the
crd
which
I
mentioned,
and
basically
the
flow
looks
like
following
so
argo:
cd
server
makes
a
change
or
creates
a
new
application.
A
Cid
controller
notices
that
change
and
process
the
settings
of
the
crd,
and
eventually
it
will
produce
kind
of
additional
data
about
application.
The
data
will
include
what
resources
are
part
of
that
application.
That's
the
state
of
every
resource
and
then
argo
cd
server
will
get
that
additional
data
from
radius
and
present
it
to
end
user
and
yeah.
So
so
the
challenges
of
ipi
server.
It
has
to
scale
really
well
because
it's
a
user
facing
component.
So
if
it
down,
then
basically
users
won't
be
happy
and
your
deployment
pipeline
might
fail.
A
A
We
used
to
it
used
to
be
totally
stateless
component,
so
it
was
just
a
proxy
to
you
know,
kubernetes
api,
which
is
here
and
to
radius,
and
it's
not
anymore.
The
reason
is
kubernetes.
Ipa
is
not
suited.
Well
for
so
it's
not
user
facing
api.
Basically,
so
it's
supposed
to
be
used
by
controllers.
A
Okay,
I
think
yeah,
that's
it
about
architecture,
good
time
to
ask
questions
before
we
move
before
we.
I
A
So
we
have
kind
of
two
types.
You
know
two
ways
we
talked
to
reaper
to
git.
One
is
just
the
git
cli.
The
you
know,
comment
give
git
tool
so
basically
paper
server,
literally
fork,
exec
git
and
execute
git
clone
plus.
We
also
use
go
git
and
we
use
that
library
to
detect.
A
If
upstream
has
any
changes-
and
I
feel
like
it
translates
into
gita
last
remote
command,
but
it
kind
of
it's
just
much
quicker
because
it's
executed
in
memory
so-
and
one
thing
I
didn't
mention
is
there-
is
a
line
between
repository
and
and
git.
So
I
think
I
I
can
better
explain
it.
If
I
give
an
example,
let's
say
you
have
an
application
and
that
application
store
resources
in
repo
abc
under
the
path
like
my
application,
so
controller
would
send
requests
to
ripple
server
and
repo
server.
First,
it
will
check
if
repository
already
cloned.
A
A
So
next
it
will
check
if
manifest
for
that
commit
already
in
in
radius
and
if
it's
a
radius,
then
just
use
it.
If
not,
then
it
will.
You
know,
check
the
configuration
of
application.
It
will
fork
exact
config
management
tool
like
home
or
customize,
get
the
manifests,
save
it
in
regis
and
serve
back
to
application
controller,
so
yeah
and
the
biggest
disadvantage
is
because,
because
it
clone
repository
once
it
kind
of
pretty
much
it
can
handle
requests
for
it
can
have
it
can
handle
up
to
one
request
for
a
positive
at
the
same
time.
A
Your
first
one
will
be
executed
and
then
the
second
one-
and
this
is
that's-
why
people
server
have
to
be
scaled,
because
I
know
that
some
companies
have
like
50
replicas
of
ripple
server,
and
this
is
like
so
such
a
waste
of
resources
and
the
main
reason
is
they
have
hundreds
of
applications
in
the
same
repository,
and
the
only
suggestion
we
have
right
now
for
them
is
to
either
scale
ripa
server
a
lot
and
have
a
lot
of
replicas
or
just
split
your
applications
across
many
positives,
and
this
will
kind
of,
and
then
you
can
have
like
two
replicas
and
that
will
be
totally
fine,
and
this
is
not
the
best
answer.
A
A
B
All
right,
so,
if
you
will,
you
know
if
you
got
another
quick
question,
maybe
yeah,
I
wonder
if
you
have
another
diagram
or
or
architecture
diagram
that
can
show
like
a
multiple
cluster.
This
is
a
single
cluster
right
single
cluster.
Yes,.
A
A
A
So
yeah,
I
I
kind
of
yeah
I
didn't
notice.
I
didn't
mention
one
thing,
so
argo
cd
yeah,
so
it
manages
multiple
clusters,
so
you
have
you
usually
install
it
like
once
and
then
provide
credentials
to
multiple
clusters
and
manage
resources
in
those
clusters,
and
this
is
it's
kind
of
it's
a
trade-off
between
security
and
convenience.
A
So,
if
you
provide
you
know,
basically,
if
someone
breaking
into
the
target
cd,
then
your
whole
production
is
compromised.
At
the
same
time,
your
only
other
option
is
to
install
argo
cd
into
each
and
every
cluster.
It
works
as
well.
So
you
can
do
it
too,
but
in
this
case
you
have
like
one
control
plane
per
cluster
and
it's
up
to
you,
whatever
you
prefer
to
do
so
at
intuit
we
kind
of
went
into
like
we
decided
to
go
for
like
a
compromise.
A
We
don't
have
single
argo
cd,
which
manages
whole
aggressive
whole
intuit
production.
We
kind
of
just
decided
to
limit
blast
radius
and
we
split
it
into
segments
and
it
works
pretty
well.
So
we
have
lot
of
fargo
cities
and
each
one
kind
of
responsible
for
piece
of
intuit
and
and
the
good
thing
as
well,
that
usually
users
of
those
argo
cds,
they
kind
of
they
don't
care
about.
So
you
know
you
the
developers
of
one
segment.
They
usually
don't
need
to
go
into
another
segment.
A
Okay,
thanks:
okay,
we
have
20
minutes
and
I
will
try
not
to
spend
too
much
time,
so
I
will
go
through
how
to
deploy.
You
know
aggressively
on
your
laptop
and
first
I
want
to
talk
about
kind
of
what
must-have
components
you
need
to
have
on
your
laptop
and
list
is
not
so
long,
so
we
have
goldlink,
you
know,
so
you
can
build
lcd.
We
have
docker
and
docker
is
for
for
radius
and
for
kubernetes.
A
Next
is
kubernetes
itself
and
good
thing
that
you
kind
of
you
need.
Almost
any
kind
of
kubernetes
would
work
because
during
development
locally
we
just
use
it
as
a
database
and
I
really
feel
like
even
testing
api
provided
by
kubernetes
might
work
and
it's
a
good
project.
You
know
in
future,
but
for
now
we
recommend
to
use
either
mini
cube
or
kts.
A
Plus
you
need
capacity.
L,
argo
cd,
won't
use
cube
ctl
itself,
but
for
debugging
it's
like
really
useful,
so
it's
kind
of
optional
component,
but
it's
impossible
to
develop
without
it
and
two
more
optional
components
is
customize
and
help
a
reason
they
are
optional
is
that
if
you
don't
use
customize
or
home
based
applications,
if
you
don't
kind
of
create
these
applications
in
your
dev
instance
or
fargo
city,
you
don't
need
them.
A
But
if
you
want
to
develop,
you
know
some
feature
for
customized
integration.
Then
you
need
customize
and
and
and
plus
it's
maybe
not
perfect,
but
if
you
want
to
run
unit
tests
locally,
not
in
docker,
but
in
your
on
your
laptop,
then
you
need
to
have
customize
on
home,
because
some
unit
tests
fork
exact
customize
go
home.
B
Another
good
question
maybe
we'll
cover
later,
can
I
run
end-to-end
tests
on
my
machine
on
my
laptop.
So
if
I
have
all
these
compo,
I
notice
the
the
end
test.
Look
for
a
k3s,
so
you
have
to
you,
have
to
have
kds
in
store
for
foreign.
A
It
depends
how
you
start
end-to-end
tests,
so
there
are.
We
can
actually
we'll
touch
it
a
little
bit
later,
but
there
is
a
way
you
can
just
you.
You
can
kind
of
manually
start
all
components
without
even
kps.
You
can
point
all
components
to
whatever
communities
you're
using
and
e2e
tests
themselves.
They
just
assume
they
just
like
to
connect
to
the
cluster
in
your
cube,
config
and
only
kind
of
hard
coded
thing
is
in
e2
tests.
It
assumes
that
all
the
tests
must
be
run
in
argo,
cd,
e2e
namespace.
A
B
But
but
you
can
do
stuff
you
can
you
can
run
your
e2e
test
on
your.
A
A
Okay,
thanks,
okay,
the
next
step-
and
it's
kind
of
it's
like
a
one-time
step.
You
need
to
do
after
you
get
your
mini
q
for
kts.
You
need
to
create
some
configuration
in
a
cluster,
and
this
is
the
list
of
things
you
need
to
create.
So
we
need
two
clgs,
two
customer
resources,
application
and
a
project
bunch
of
config
maps
and
one
secret,
and
this
is
how
you
can.
Basically,
if
you
just
run
that
one
command,
it
will
create
all
the
resources
and
there
are
different
ways
you
can
do
it.
A
I
feel
like
the
documentation.
Oh,
I
forgot
to
mention
yeah.
There
is
a
link
to
developer
guide
and
it
explains
how
to
configure
your
cluster,
how
to
install
these
components.
So
there
are
different
ways
you
can.
You
know
achieve
this.
Cluster
configuration
step,
developer
guide,
explains
one
way
and
that's
why
I
mentioned
another
way
here,
so
this
is
what
e2
tests
are
doing,
so
we
have
kind
of
a
set
of
manifests
specific
for
testing
environment.
A
So
if
you
just
run
this
one
command,
it
will
create
all
these
resources
for
you
and
it
has
additional
convenience.
It
will
configure
rgo3d
to
use
built-in
user
admin
and
it
will
reset
password
to
password
to
word
password.
So-
and
I
I
mentioned
it
here,
because
it's
been
like
useful
for
me.
Basically,
I
run
it
very
often
every
time
I
test
like
authentication
and
then
I
change
password
to
something
I
forgot
in
the
future.
You
can
just
rerun
the
same
test
and
you
will
be
back
to
kind
of
clean
environment,
yeah.
A
Okay,
so
once
you've
done
visits
next,
you
can
pretty
much.
You
know,
run
components
and
I've
seen
from
conversation
in
dev
channel.
I
think
everyone
already
kind
of
started
trying
option,
one
which
is
run
everything
in
docker,
and
I
think
this
option
was
developed
by
jan
I'm
sure.
A
So
I
wanted
to
discuss
option
two
and
the
reason
is
this
option.
Two
is
a
little
bit
more,
you
know
it
will
make
you
more
performant
plus
make
start
uses
basements
if
you
know
how
to
run
components
manually.
You
already
know
how
make
start
work
because
make
start
literally
execute
these
steps,
automate
the
steps
which
I'm
going
to
describe.
A
Okay,
so
option
two
is
pretty
much.
A
The
main
difference
between
you
know,
running
things
in
docker
is
that
you,
you
can
use
some
cli
commands
to
start
components
of
argo
cd
one
by
one
and
advantage
is
that
there
is
no
docker,
so
you
don't
need
to
use
volume
mounting
and
your
calculation
step
is
much
faster
and,
additionally,
you
can
choose
to
start
only
required
components
one
by
one
instead
of
starting
starting
everything
at
the
same
time,
yeah
and
the
only
kind
of
preparation
you
need
to
do
is
you
need
to
install
install
tool
called
gorman
and
I'm
not
sure
if
any
one
of
you
worked
with
ruby,
I
think
they
started
it.
A
First,
they
created
a
tool
called
foreman
and
then
node.js
added
nodemen
and
then
go
created
gorman.
So
it's
a
simple
tool
to
run
multiple
processes
and
tail
logs
from
all
of
them.
So
I
already
installed
the
tool
and
what
the
tool
does
it
simply
look
for
rods
file
in
your
root
directory,
look
for
lines
in
the
profile
and
basically
each
line
represents
a
process
which
is
you
know
it
know
and
which
it
supposed
to
start,
and
so
the
format
is
name
of
your
process
and
then
the
command
line
which
supposed
to
be
executed.
A
So
it's
extremely
simple
and
we
have
pros
file
in
argo
cd,
positive
with
a
line
for
every
argo
cd
component.
So
we
already
talked
about
controller
ipi
server.
There
is
dex,
I
didn't
mention
it,
and
hopefully
you
won't
need
it
very
soon.
It's
for
ssr.
A
Hopefully
it's
because
usually,
if
you
need
it,
it's
some
kind
of
tricky
bug
related
to
oidc
integration.
So,
and
I'm
mentioning
it
now,
because
it's
advantage
of
running
things
kind
of
independently
one
by
one
and
it
saves
resources
on
your
laptop.
So
if
you
don't
troubleshoot
or
idc
integration,
then
you
don't
just
don't
run
dex
next
is
radius.
A
A
The
mounter
is
a
tool
for
pretty
much
for
two
features:
gpg
keys
and
custom
certificate
testing
so-
and
I
kind
of
I
wanted
to
show
that
file
because
I
feel
like
it's
extremely
useful
for
developers.
A
Let
me
know
if
you
don't
see
my
screen.
Well,
I
think
it's,
hopefully
it's
big
enough
so
yeah.
This
is
how
you
can
start
ips
server,
independently
ipa
server
and
yeah,
so
you
can
just
run
gourmand
start
and
then
the
name
of
component
from
the
pros
file.
A
A
Let
me
do
it
again,
so
I
started
api
server
and
ui,
and
this
is
pretty
much
if
I'm
working
on
user
interface.
Only
I
don't
care
about
controller
repo
server.
I
don't
have
to
run
all
these
other
components,
so
we
can
just
use
gorman
to
start
these
two
and
yeah
it's
much
quicker
plus.
If
you're
making
changes
in
equal
server
and
you
want
to
restart
it
quickly,
you
don't
have
to
kill
like
everything
and
then
recompile
everything
just
instead,
you
can
just
go
and
you
know
restart
just
deeper
server.
A
So
that's
what
I
did.
I
just
killed
that
process
and
I
started
it
again.
So
if
I
make
code
changes
in
between
I
just
decompile
on
the
only
api,
server
and,
and
then
things
still
working,
so
this
is
kind
of
my
local
argo
cd.
It
happens
to
have
two
applications
and
I
can
make
changes
and
play
with
it.
A
So
let's
say
you
want
to
test
some
controller
changes,
usually
what
I
do.
I
just
open
one
more
tab
and
then
start
the
rest
of
components.
So
gorman
start
reaper
server,
radius
and
controller.
So
if
I
start
that,
then
immediately
I
can
do
kind
of
more
things.
In
my
local
environment,
for
example,
I
can
refresh
application.
So
if
I
click
refresh
now,
okay,
it's
been
refreshed
because
controller
already
started
and
it
picked
the
change
and
yeah.
So
it's
working
as
soon
as
I
don't
care
about
controller.
A
And
yeah,
so
I
highly
recommend
to
use
this
way
and
it's
kind
of
it.
It
is,
I
think,
mike
start
is
a
good
kind
of
way
to
get
started
really
quickly,
and
then
you
can
kind
of
you
know
gorman
as
a
next
step.
So
it's
like
a
little
bit
more
difficult
to
use
but
easier
to
you
know
it's
better
for
people,
performance,
okay
and
next
kind
of
great
thing
about
profile.
A
Is
that,
as
you
know,
like
we'll
use
different
ide,
I
use
golend,
and
then
you
know
someone
use
vs
code,
someone
else
might
use
eclipse
so
and
we
cannot
commit
config.
You
know
configuration
for
every
ide
in
git,
but
profile
works
for
everyone
and
this
basically
it
has
all
the
information
about
how
to
style
components
locally
in
you
know,
dev
friendly
way,
and
you
can
use
it,
for
example,
to
configure
your
ide.
So
what
I
do
periodically,
I
just
copy
paste
this
line
from
here
and
put
it
into
debug
configuration
in
golet.
A
So,
for
example,
let's
take
a
look
at
ips
server,
so
this
is
how
you
can
run
ipi
server.
It
has
bunch
of
environment
variables,
it
has
a
port.
Sorry,
it
has,
you
know
downstream
service,
hostnames
and
ports.
All
I
need
to
do.
I
just
need
to
copy
paste
it
and
put
it
into
configuration
of
intellij
or
the
goland.
So
this
is
pretty
much
almost
the
same
line
here
and
yeah.
A
A
We
keep
introducing
new
flags,
maybe
name
some
flux,
introduce
new
environment
variables
and
you
know
team
make
sure
that
pros
file
works,
and
so,
if
something
broke
in
your
environment,
you
always
can
go
here
and
basically
make
sure
that
your
configuration
in
your
ide
matches
this
file
and
if
it
doesn't
match,
then
you
have
answered
so
basically
less
communication
required.
So
this
is
our
way
to
maintain
you
know
becoming
genius
development
environment.
A
So
my
ips,
wherever
I
started,
I
can
kind
of
check
it's
still
working
and-
and
this
is
you
already
know
how
to
it
kind
of
helps
you
to
debug
things.
So
I
know
how,
if
I
want
to
debug
something
yeah
all
I
need
to
do.
I
just
need
to
do
this
configuration
of
my
configure
my
debug
configuration
and
then
create
a
breakpoint
in
in
any
methods.
A
A
This
is
to
teach
cargo
cd
to
connect
to
cluster
in
your
queue
config,
if
you're
trying
to
use
queue,
ctl
default
svt
and
the
reason
is
it's
like
it's
a
magic
reserved
url
inside
of
kubernetes,
it's
hard
to
configure
it
locally.
Instead,
with
this,
we
just
have
environment
variable
for
argo.
Cd
to
you
know,
do
this
switch
and
few
other
environment
variables.
So
we
have
several
features
related
to
gpg,
keys
and
custom
certificates,
and
the
problem
is,
we
have
to
use
volume
mounting
kubernetes
volume
mounting
and
it's
hard
to
get
it
working
locally.
A
A
A
Let
me
find
one
and
then
from
here
you
can
just
you
can
kind
of
click
run
from
here
or
you
can.
You
know
another
way
you
can
use
golang
cli
and
you
can
type
go
test
and
provide
the
cli
command.
So
this
is
enough.
So
basically
what
it
will
do
it
will
try
to
use
config
to
connect
to
your
cluster.
It
will
make
sure
yeah
one
prerequisite
is
it
will
check
if
argo
cd,
it
will
try
to
use
argo
cd.
A
A
Okay,
let's
we
almost
run
out
of
time.
I
was
hoping
we'll
have
some
time,
for
you
know
two
questions,
but
let
me
try
to
finish
it
so
we
do
have.
This
is
the
url
to
get
good
first
issues
filter,
so
we
have
a
label
good,
first
issue
and
every
time
so,
every
day
we
have
a
meeting
where
we're
trying
to
assign
that
label
to
you
know:
tickets,
which
are
good
for
starters.
So
if
you
open
it,
there
are
already
some
urls
some
tickets
and
what
we
usually
apply
that
label.
A
If
so
ticket
is
not
necessary
like
easy
to
do,
but
at
least
it
is
straightforward
and
no
design
discussions
are
needed
and
kind
of
all
the
information
is
inside
of
the
ticket,
so
it's
kind
of
self-explanatory
how
to
implement
it.
Plus
we
have
additional
labels,
which
you
can
use
for
filtering.
A
Basically,
every
label
which
starts
from
component
something
usually
we're
trying
to
assign
it
to
you
know
to
these
tickets,
and
this
is
the
label
this.
These
are
components
which
we
have.
A
I
think
there
is
around
10
different
component
labels
and
some
of
them
related
to
ipi,
some
of
them
related
to
audi
cli
for
red
hat
guys.
We
have
third
party
kubernetes
label
and
usually,
if
we
see
something
specific
to
openshift,
we
apply
that
label
or
enter
yeah.
A
L
A
A
It's
released,
it
really
depends
on
it's
changing
over
time,
so
it
used
to
be
kind
of
frequently
when
we
had
less
users-
and
I
think
now
we're
literally
transitioning
into
you
know.
Now
we
have
to
we're
forced
to
test
better
and
it
became
like
10
times
harder
because
of
number
of
users.
So
we
tried
to
increase
the
recycle
and
we
were
trying
to
you
know:
do
it
like
a
waterfall
way,
basically
break
master
and
then
have
a
kind
of
hard
stop
and
basically
don't
merge.
A
A
lot
of
you
know
at
least
big
prs
only
merge,
very
simple
ones,
and
you
know
focus
on
bug,
fixing
and
testing
and
it's
this
is
not
working
well
and
I
feel
like
we
are
transitioning
into
almost
like
okay,
at
least
internally
we're
going
to
we're
trying
to
release
things
continuously
and
yeah.
So
but,
okay
to
answer
your
question
right
now,
it's
released
around
every
two
three
months
and
I
feel
like
we're
looking
for
a
way
to
improve
quality
and
it
might
result
into
a
little
shorter
cycle.
A
Maybe
we'll
release
quicker
yeah,
okay,
and
it
really
depends
on
how
it's
going
to
work
forward
with
more
contributors,
because
when
argo
didn't
have
that
many
users,
it
was
kind
of
okay
to
break
things.
And
then
you
know
the
new
users
who
joins.
They
can
report
about
the
bug.
We
can
fix
them
and
everyone
is
happy
now
we're
really
kind
of
affecting
someone's
everyday.
You
know
work
and
every
bug,
even
if
it
doesn't
affect
you
as
a
user.
It
might
be.
You
know
it
might
make
someone's
production.
A
A
Enough
success,
okay,
yeah
fortunately
or
unfortunately,
argo
is
really
becoming
like
a
service
and
I'm
super
happy.
We
have
kubernetes
because
it
just
kind
of
abstract
away.
Although
all
this
complexity,
it's
still
possible
to
deploy
argo
cd,
just
using
cube,
ctl
applied,
but
internally
yeah.
We
have
like
components
which
talk
to
each
other
and
you
know
that
it
like
we
have
to
deal
with.
You
know,
resilience
and
you
know
components
have
to
try
to
heal
themselves
and
so
on.
T
I
I
came
across
another
github
organization,
known
as
approach
lab,
which
has
adversity
operator
and
couple
of
other
projects.
So
how
is
this
related
to
the
main
argo
organization
and.
A
The
goal
of
every
research
project
is
to
kind
of
graduate
into
arguable
and
we
decided
to
have
separate
github
work
to
kind
of
lower
the
barrier,
so
you
you
can
use
it
as
a
you
know,
kind
of
easy
way
to
advertise
your
project,
because
you
know
there
is
a
chance.
People
can
keep
looking
into
that
repository
plus.
It's
really
easy
to
you
know
iterate
quicker.
A
So
good
example,
I
guess
image
updater.
It
was
recently
created.
Image.
Updater
is
like
a
is
a
feature
which
asked
by
many
users
from
argo
cd,
but
feature
is
kind
of
questionable.
Some
people,
don't
like
the
feature
at
all.
The
feature
is
automatically
arrived
back
automatically
updates
git
manifest
when
you,
when
a
new
image
is
pushed
into
docker
and
this
feature
can
be
implemented
as
a
separate
project
and
you
just
run
it
as
a
cycle.
A
A
Maybe
you
know
it's
not
polished,
yet
things
can
change
as
soon
as
it's
stable.
I
think
we
want
to
move
it
into
our
approach
yeah.
So
this
is
so
kind
of.
If
you
have
your
idea
and
you
want
to
validate
it,
you
know
something
for
I
guess.
Fargo
cd
or
argo
workflows
put
it
into
our
google
apps
and
hopefully
you
will
get
users
sooner
then
you
can
play
with
it
at
it
and
then
eventually.
A
Yeah,
actually
it's
a
good
point.
I
feel
like
yeah,
then
it's
you
know.
In
order
for
it
to
move
into
algo
approach,
it
must
be
either
for
workflows
or
virgo
cd,
like
not
a
brand
new
project.
It
cannot
be
just
something
you
know
totally
unrelated,
but
I
I
feel
like
that.
Maybe
you
can
correct
me
if
I'm
wrong,
I
feel
like
if
we
decide
to
move
argo
cd
notifications
into
our
approach,
it
should
not
be
an
issue
right.
We
can
just
move
that
project
because
it's
kind
of
continuation
of
adversity.
C
C
Is
also
going
on
right
now
is
applications
that
makes
it
really
easy
for
to
manage
basically
the
same
application
across
multiple
clusters
and
that
probably
will
almost
certainly
become
a
part
of
the
argo
cd.
You
know
for
more
of
the
cd
project,
so
that's
something
to
keep
an
eye
on
as
well.
A
Yeah
and
I
feel
like
as
a
community,
we
can,
you
know
there
are
other
projects
which
so
argo
applications
is
a
good
example.
I
feel
like
because
everyone
needs
it.
You
know.
Obviously
you
need
notifications,
image
updater,
it's
another
example
where
you
know
some
users
might
want
it
and
some
don't
so,
and
if
you
know
as
a
community
here
we
have
a
lot
of
interest
in
it.
A
Then
you
know
you
can
just
mention
it
and
then,
if
you
so
let's
say
you
have
new
project
which
is
which
you
think
is
useful
for
some
rvc
users
then
good
to
mention
it
here
and
then
it
will
get
fair
visibility
and
will
be
prioritized.
You
know
if
there
is
interest
in
the
community.
A
Yeah,
I
saw
a
question
about
flux,
integration,
yeah,
so
answer
is,
I
feel
like
we
basically
the
flux
for
those
who
don't
know.
We
worked
with
flux
team
and
we
were
trying
to
pretty
much
merge
two
projects
or
at
least
find
a
way
to
merge
projects,
and
we
started
from
trying
to
merge
code
bases
and
at
least
they
use
the
same
libraries
you
know
use
one
library
for
identical
features,
and
so
what
happened?
A
Is
we
learned
a
lot
and
we've
created
a
library
which
can
be
used
in
argo
and
flux,
and
right
now,
argo
uses
that
library
and
flux
is
pretty
much.
They
decided
to
build
version
2,
which
is
not
even
it
doesn't
look
like
flux
version
one
and
that's
the
focus
right
now
in
future.
They
might
want
to
use
it.
You
know,
I
think,
as
soon
as
they
get
exact
same
use,
cases
where
libraries,
which
we
built,
is
really
helpful.
A
A
No
plans
from
flux
team
to
use
the
library
which
was
developed
together
right
now
and
in
future
it
depends
on
the
you
know,
on
their
priorities,
and
the
good
news
about
this
library
is
that
there
are
some
other
projects
which
kind
of
at
least
working
on
pocs
right
now
and
they
contribute
so
the
work
which
we
did
it
kind
of.
It
was
not
a
waste
of
time
and
I
keep
referring
to
it
as
a
library.
So
the
library
called
github's
engine
and
it's
an
argo
approach
repository
so
in
our
go
approach.
A
A
Okay,
I
yeah,
I
was
hoping
we'll
get
more
time.
You
know
to
talk
about
next
meetings
and
for
you
to
give
feedback.
So
if
you
have
any
feedback
and
if
you
okay
to
stay
longer,
please
you
know
I'm
here
and
or
we
can
kind
of,
took
it
offline
and
try
to
figure
out
agenda
for
next.
A
Meeting
yeah
so
for
because
we
have
meetings
every
week,
I'm
hoping
to
kind
of
make
this
next
meeting
more.
You
know
oriented
for
you,
so
it
would
not
be
just
you
know
talking.
I
I
wish
to.
I
wish
you
a
kind
of
answer,
ask
some
questions
and
we'll
try
to
prepare
answers
for
your
questions
and
yeah.
So
please
use
next
week
to
play
with
argo
cd
devtools
and
then
give
your
feedback.
If
something
is
not
working-
and
I
know
that
you
know
we
do
have
some
issues
with
dockerized
environment-
it's
just.
A
A
I
keep
kind
of
pushing
to
use
gorman
and
let
us
know
you
know
if
you
want
flood
type
to
run
argo
cd
and
we
can
incorporate
it
in
different
environments,
and
I
personally
I
like
to
rely
on
ci
to
run
into
a
test.
If,
let's
say,
if
you
william,
wants
to
run
it
locally,
then
you
know
tell
us
please
if
it
doesn't
work
for
you
and
we
can
work
together
to
improve
it.
N
Yeah,
and
also
one
more
thing
I
want
to
just
bring
down,
is
since
we
have
a
bigger
community.
I
think
there
are
a
lot
of
the
questions
here
and
there.
So
since
we
are
the
contributors,
I
think
you
guys
actually
have
a
little
bit
more
knowledge
or
even
have
a
lot
of
knowledge
than
us,
sometimes
to
please
help
to
share
your
knowledge
with
others
to
help
the
questions
so
really
appreciate
that.
A
Yeah,
I
see
a
proposal
from
alex
yeah
and
it
contributed
a
third
way
to
iran.
Argo
cd
actually-
and
I
totally
didn't
mention
it-
could
you
brought
it
up?
So
there
is
a
way
to
deploy
argo
cd
into
the
cluster
and
debug
it
remotely
I'm
yeah.
I
never.
I
have
not
tried
here
just
because
I'm
fixing
bugs
frantically
for
1.7
release
as
soon
as
it's
done.
I
want
to
try
to
yeah.
K
A
N
N
I
think,
since
you
have
the
most
knowledge,
maybe
in
the
future
we
can
have
a
one
day
session,
just
like
see
how
we
can
get
the
inside
about
internally.
What
happened
like
the
memory
issues,
then
the
all
the
interaction
where
it
could
be
the
structure.
So
I
think
that
will
be
very
beneficial
to
the
community.
A
I
agree
and
I
think
beneficial
for
you
know
for
us,
meaning
you
know
you
have
to
for
for
end
users.
So
what
is
talking
about
after,
like
1.7
release,
introduced
a
lot
of
changes
and
we're
forced
to
make
these
changes
because
we
kind
of
we
we
have
to
scale
more
and
more
and
we
did
these
changes
and
now
we
kind
of
have
production
issues
as
soon
as
we're
done
with
fixing
them.
I
think
I
can
give
a
kind
of
post-mortem
and
explain
what
happened
and
then
you
will
have
knowledge
about.
A
G
Much
thank
you
thanks
for
joining.
Thank
you
thanks
a
lot
for
the
stationary
super
helpful.
Thank
you.
Thank
you.
Alex.