►
From YouTube: Argo Workflow Community Meeting 17th Jan 2020
Description
http://bit.ly/argo-wf-cmty-mtng
Argo Workflows v2.5 is the biggest release since v2.0 and includes 3 big new features. Join us to learn more about these features and provide early feedback:
* 0:00 Introductions
* 5:00 Demo of new UI
* 11:00 Run scheduled workflows using Cron Workflows.
* 13:45 API server with enhanced security and more features
* 17:45 Keep workflows around for a long time with Workflow Archive.
* 19:20 Q&A
A
A
Close
this
I,
can
you
see
my
screen,
brilliant,
okay,
great
and
do
I?
Have
my
sound
on
that's
the
other
thing.
I
do
I!
Guess
if
people
speak,
then
we
can
find
out
myself
hello
good
morning
and
welcome
to
the
first
Argo
community
meeting
of
2020
and
also
the
first
community
meeting,
just
specifically
for
Argo
workflows,
we're
going
to
cover
some
exciting
new
stuff
today,
and
we
all
really
appreciate
you
coming
along
in
such
great
numbers.
It's
good
to
have
a
new
year's
resolution
may
be
every
reason.
A
Numerous
resolution
is
to
come
along
to
our
NGO
community
meetings.
That's
a
great
resolution!
If
that's
your
resolution
bit
weird
but
good,
you
know
fine
and
with
this
is
also
the
first
kind
of
arguable
workloads
only
community
meeting
and
in
the
past,
what
we've
done.
We've
combined
the
Argo
community
meetings
together
to
one,
but
what
we've
really
found
is
that
people
are
really
interested
in
one
particular
topics,
and
so
this
is
an
opportunity
just
to
kind
of
really
focus
from,
hopefully
around
30
minutes
on
on
what
we've
got
coming
up
in
our
NGO
workflows.
A
So
we're
going
to
do
the
introduction,
which
is
which
is
this,
then
we're
going
to
go
and
do
a
demo
of
the
new
user
interface
coming
in
version
2.5.
Then
we're
going
to
do
a
bit
of
a
deep
dive
into
a
couple
of
features
coming,
which
are
new
features,
called
cron
work,
clothes
and
the
API
server
and
thing
called
the
workflow
archiving.
A
Obviously,
for
those
who
are
new,
our
go
workflows
is
a
workflow
engine
runs
on
kubernetes
those
cloud
native
and
it's
particularly
popular.
We
found
recently
with
people
running
machine
learning,
workflows
and
we've
spoken
to
people
from
companies
like
Cray,
who
are
doing
a
lot
of
this
stuff
as
well.
So
that's
quite
interesting.
Who
are
we
well?
A
The
people
you'll
be
hearing
from
today
and
myself,
Simon
and
bar,
and
we
are
three
of
the
four
I
guess:
you'd
call
core
agro
workflows
developers
into
it
and
there's
a
team
around
eight
I
was
working
on
the
agro
projects
in
in
general,
and
we've
been
really
focused
in
the
last
kind
of
six
weeks
to
three
months
are
looking
at
Baran
on
these
new
capabilities
and
we're
pretty
excited
because
I'll
go
workflows,
2.5,
there's,
probably
gonna,
be
the
biggest
biggest
release.
Fargo
workflow
since
version
to
either
say
version
1,
perhaps
I
have
been
corrected
version.
A
B
A
A
A
Well,
although
the
end
of
this
is
your
QA,
and
maybe
we
can
talk
about
roadmap
and
people
can
tell
us
what
they
particularly
interested
in
or
kind
of
looking
forward
to
at
the
end,
and
if
you
want
to
come
introduce
yourself
just
say
you
know
hi
in
the
chat
room,
and
you
know
where
you're
from
as
in
you,
what
company,
you're
working
for
and
also
maybe
about
a
little
bit
about
what
you're
using
I'll
go.
Workflows
we'd
really
like
to
hear
about
that.
A
A
Okay,
that's
the
introduction,
so
I
promised
a
UI
demo
today,
and
this
is
an
early
version.
Let
me
just
it
doesn't
work,
particularly
what's
better
for
some
reason:
two-thirds
screen
of
the
new
in
user
interface
for
Argo
that
we're
going
to
be
introduced,
so
the
the
existing
user
interface
is
written
in
nodejs
and
reacts
and
we're
going
to
replace
it
with
a
thing
called
the
Argo
server
and
we'll
talk
a
bit
more
about
that
in
detail.
A
But
the
Argo
server
is
going
to
be
written
in
go
line
which
allows
us
to
leverage
a
lot
of
the
NGO
related
libraries
that
we
use
for
it
and
add
additional
features
more
successfully
and
also
use
kind
of
more
sophisticated
security
models
than
we
have
used
in
the
past
and
also
the
opportunity
to
kind
of
refresh
the
user
interface.
And
so
some
of
you,
people
will
probably
be
familiar
with
the
user
interface.
A
Okay,
so
the
first
area
here,
of
course,
is
that
is
the
existing
workflow
Timeline
view,
and
a
couple
of
changes
to
this
view
is:
there's
no
namespace
fills
from
the
top
right
hand
side,
so
you
can
filter
by
the
name,
space
you're
using
and
also
a
another
search
here.
So
you
can
search
for
your
workflows
by
name
and
go
into
them.
Have
a
look?
You'll
also
notice,
there's
a
new
button
here,
submitting
new
workflows
that
allows
you
to
create
a
new
workflow
from
the
user
interface.
A
You
just
click
on
that
you'll
see
you
get
this
yellow
here
and
you
can
just
enter.
You
know
details.
This
is
an
early
version,
so
the
names
basically
defaults
to
default,
but
actually
that
this
configuration
always
has
to
submit
only
has
permission
to
read
and
write
from
Iago
namespace
click,
Submit
and
I'll.
Just
take
you
in
and
your
new
workflow
will
just
start
right
there
and
then,
from
from
the
user
interface.
You'll
also
know
some
additional
changes
to
this.
A
The
existing
one
you'll
see
there's
a
number
of
new
buttons
on
the
toolbar
here
as
well.
The
ability
to
kind
of
retry
resubmit
and
delete
workflows,
so
that's
a
new
feature
and
there's
been
a
slightly
a
little
bit
of
a
refresh
to
this
part
of
the
user
interface
here
around
displaying
logs
now
so
there's
a
new
log
view
which
and
tells
you
which
container
you're,
you're
looking
at
but
otherwise
kind
of
functionally
the
same
to
it.
The
second
new
area
of
the
user
interface
is
around
workflow
templates,
I'm.
A
Just
thinking
how
to
describe
a
workflow
template
template
is
or
has
been
submitted
by
one
of
our
developers.
It's
books
turned
about
daisuke
and
this
basically
weighted
kind
of
template,
a
workflow
here
you
can
go
in
here
again.
You
can
create
a
new
workflow,
a
template
to
type
in
the
namespace
to
fix
that,
and
then
you've
got
a
new
template
here
that
you
can
edit
and
you
can
then
submit.
A
This
is
a
normal
workflow,
so
that
allows
you
create
a
library
of
workflow
templates
and
you
submit
ins,
new
new
workflows,
tab,
3,
temporaries,
cron,
workflows,
I'm,
not
going
to
talk
about
this
in
depth,
because
Simon's
going
to
come
round
to
talked
a
bit
about
chrome,
workflows,
but
but
in
summary,
it's
a
it's
a
cron
job
for
work,
clothes
and
you
can
go
in
here
again.
You
can
create
a
new
workflow.
You
can
and
you
can
edit
the
workflow,
and
it
shows
you
information
about
this
workflow
and
this.
A
A
A
couple
new
features:
the
ability
to
run
the
user
interfaces
as
a
local
server
on
your
on
your
laptop
using
our
go
server
and
it'll
actually
use
your
kubb
config
to
use
its
permissions
rather
than
what
the
other
UI
does
now
is
use
the
service
account
that
it's
been
deployed
with.
So
that's
a
new
way.
We
call
that
a
local
mode
configuration
and
actually,
obviously
you
can
still
run
this
user
interface
in
a
hosted
mode
and
all
take
the
cube
config
of
the
service
account
or
we're.
A
A
A
A
C
D
C
Some
feedback
and
I
really
was
a
very
well
requested
feature
by
the
community,
so
we
are
very,
very
excited
that
we're
able
to
finally
ship
this
to
you
guys
this
week,
so
Chrome
or
close
a
quick
overview
is
that
it's
an
independent
CRD
that
is
managed
by
an
independent
controller.
This
controller
is
packaged
with
the
workflow
controller
binary.
So
there
is
no
additional
installation
required
for
developers.
C
Conversion
is
very
simple,
so
a
cron
workflow
is
essentially
a
workflow
with
less
options
so
to
convert
a
workflow
that
you've
been
using
already
to
Akron
workflow.
All
you
have
to
do
is
I
change.
The
kind
to
a
cron
workflow
add
some
options,
such
as
a
schedule
and
in
some
other
options
which
I
will
talk
about
shortly,
and
you
should
be
good
to
go
these
options
pretty
much
mimic
exactly
what
a
cron
job
cron
job
options
are
from
standard
core
Nettie's.
C
Obviously
these
include
the
schedule,
but
also
include
things
such
as
concurrency
policy,
successful
jobs,
history
limit,
failed
jobs
to
limit,
and
so
on.
These
haven't
tested
to
work
pretty
much
identically
to
to
how
they
work
on
cron
jobs.
We
also
have
time
zone
support.
It
appears
to
be
murder
soon.
C
So
if
you
have
tried
this,
if
you
have
tried
this
sort
of
on
your
own
over
the
last
couple
days,
it's
not
implemented
yet,
but
it
should
be
within
this
week
and
then
we
also
have
a
cron
workflow
CLI
that
you
can
use
to
manage
all
of
this.
It
would
be
out
our
go
cron
and
then
the
standard
verbs
such
as
get
create
list
lint
delete,
will
be
available.
So
just
so,
you
guys
don't
try
to
make
cron
workflows
using
our
go
submit,
which
is
only
for
normal
work.
C
E
D
So
the
total
on
2.5
we
are
introducing
another
binary
is
like
not
a
binary.
We
are
adding.
Another
capability
is
a
PA
survey
which
which
will
support
all
the
APA
functionalities
forward
flow
or
flow
template,
chrono
flow
and
portal
archive.
That
episode
will
support
the
gr,
PC
support
and
rest
calls.
D
So
the
client
can
make
a
GPC
our
rest
and
it's
a
standalone
executable,
so
you
can
start
it
as
a
local
mode
or
hosted
mode,
and
it
will
support
the
three
types
of
indications,
basically,
no
server
authentication,
which
will
take
it
from
where
you
are
running
like
a
kernel.
Local
mode
means
the
current
current
cube
context
to
country
the
host
or
mode
means
the
service
at
one.
Where
we
are
running,
then
crank
mode
will
be
like
it
know,
always
looking
for
the
claim
token
from
the
request,
header
and
hybrid
mode
is
like
ignore.
D
B
D
So,
in
what
throw
template
we
have,
you
can
create
our
float
and
plate,
get
the
workflow
template
list
and
add
a
leaf
and
the
chrome
you
can
create
a
throw,
not
flow,
get
the
preneur
flow
list.
Delete,
suspend
and
resume,
and
archive
is
like.
Let
me
get
your
archived
or
throws
list
and
the
particular
arcade
or
flow
and
a
delete
this
other
things
you
can
programmatically
manage
workflows
instead
of
using
a
Argos,
ela
or
UI.
D
D
A
Ok,
I.
Take
you
back
over.
Thank
you
very
much
Bala,
so
barlas
been
working
on
this
for
some
time,
so
I
think
he's
probably
pretty
specially
and
proud
of
it.
I
think
it's
pretty
really
exciting.
Some
of
the
benefits
you're
gonna
get
out
of
the
API
server,
so
we
talked
a
bit
about
in
2.4,
there's
a
feature
called
off
offload
node
status,
which
allows
you
to
run
very
large
workflows
and
that
does
it
by
offloading
into
an
SQL
database.
Obviously,
none
of
the
kubernetes
api
can
support,
reading
and
writing
from
SQL
database.
A
The
new
api
server
provides
you
with
all
those
features
and
they're,
obviously
all
tailored
to
running.
You
know
workflows,
so
we're
hoping
that
people
will
find
it's
really
useful
and
there
will
be
able
to
make
it
easy
to
build
tooling
and
SDKs
and
that
kind
of
stuff,
okay,
so
I
think
we
are
now
done
with
the
demos
and
looking
around
the
room
again
here:
cool
well.
C
A
Archives,
ok,
we
could
talk
a
little
bit
about
what
archives
I
haven't,
provided
any
any
slides
in
it
assignment.
So
so
in
2.5,
where
we're
introducing
a
new
feature
called
the
workflow
archive
now,
workflow
archive
is
effectively
just
a
table
in
an
SQL
database
where
we
will
save
workflows
the
current
way.
This
works,
and
you
know
we're
interested
in
feedback
on
whether
we
use
the
right
terminology
and
it's
working
in
the
way
that
people
would
like
is
that
when
a
were
foe
completes.
A
So
that's
when
the
status
and
status
phase
moves
to
a
computer's
data
such
as
fails,
error
or
successful,
then
we'll
actually
then
write
the
the
whole
of
the
manifest
into
a
database
table
in
the
SQL
database
and
there.
And
then
you
can
then
delete
that
workflow
using
our
go,
delete
command
or
deleting
it
from
the
user
interface
or
even
using
cube,
CTL
and
delete,
and
that
will
delete
that
and
workflow
from
the
working
step,
but
obviously
remain
in
the
archive.
A
And
then
when
you
want
to
resubmit
that
workflow
from
the
archive,
you
can
actually
what
you
can
do
it
in
a
couple
of
different
ways.
But
the
basic
the
basic
way
you
do
is
you
do
an
Argo
archive
get
and
then
you'd
give
it
the
UID
of
the
workflow
and
then
they'll
give
you
the
JSON,
and
then
you
can
do
Argo
submit
on
that
JSON
document
so
that
that's
kind
of
how
it
works.
A
At
the
moment
we
interested
if
people
want
it
to
work
differently
and
we're
also
going
to
keen
for
people
to
try
it
out
and
see
if
it
kind
of
fulfills
the
UK
use
cases
they
would
like
to
do.
It's
very
easy
to
misspell
the
word
archive
as
I've
discovered.
When
writing
this
never
miss
spelt
it
with
ease
and
eyes
in
all
sorts
of
different
orders.
So
close.
So
if
you
to
structure
a
spelling
mistake
with
the
word
archive
to,
let
me
know
I'm
sure,
there's
still
some
lingering
around
in
Akko
beds.
A
B
From
the
one
question
is
how
many
of
you
really
want
to
use
the
persistence
feature
of
like
how
many
of
you
really
want
to
archive
workforce?
This
has
been
an
ass
for
a
long
time
and
it
took
some
time
to
build
it
because
of
other
priorities.
So
have
you
guys
have
some
of
you
already
built
persistence
of
workflows
internally,
or
do
you
want
to
use
this
anybody.
F
A
Yeah
do
that
that
kind
of
stuff
in
it
what
we
do
currently
is
we
there's
a
literature's,
a
text
column
on
the
database,
because
we
don't
know
how
people
want
to
pause
this
this
this
data
and
we've
added
a
few
additional
columns
based
on
what
we
think
people
are
likely
to
want
to
query
and
those
columns
are
named
namespace,
start
time,
finished
time
and
phase
so
they're.
All
in
the
database,
though
they're
not
necessarily
indexed.
F
A
A
A
Thank
You
Jesse,
Jesse
Jesse
didn't
introduce
himself
so
Jesse's
a
principal
engineer
on
the
Argo
team
and
he's
one
of
the
original
commits
remark.
I
work,
those
he
knows
his
stuff
I'm,
just
gonna,
go
back
as
I
feel
I.
Just
noticed.
We've
skipped
over
a
question
from
from
a
couple
of
people
here,
so
I'll
try
and
do
that
I'll
try
and
do
them
in
order.
I'll
use
my
mouse
points.
It's
indicate
which
ones
we're
going
to
so
Evan
asks.
Do
the
container
logs
get
archived
with
the
workflow
and
being
archived
the
question.
A
The
answer
to
that
is:
yes,
no
it's
so
we
don't
copy
the
logs
in
to
it.
But
if
you're,
if
you're
already
using
in
the
configuration
under
artifacts,
you
can
set
to
archive
logs
and
if
you
do
that,
you
they'll
be
archived
wherever
you
archive
them,
which
for
us
is
ss3,
and
you
have
to
get
the
configuration
right
for
that
to
make
sure
that
new
workflows,
don't
overwrite
your
old
artifacts,
and
so
you
can
keep
your
archive
logs.
But
you
don't
actually
need
an
archiving
to
to
do
that.
A
C
So
if
you
buy
the
question
you
mean
of
having
persistent
workflow
like
you
can
run
at
any
time
in
such,
as
is
the
case
with
the
workflow
template,
then
the
answer
is
kind
of
yes,
because
Cornwall
close
being
their
own
CR.
These
are
persistent
on
the
cluster,
so
they
can
always
be
run.
If
you
mean
that
you
can
reference
a
cron
workflow
from
let's
say
a
normal
workflow
or
a
workflow
template,
you
can't
do
that
work
on
work.
C
Those
are
sort
of
the
sync
notes
in
this
dependency
graph
they
can,
but
they
can
fully
reference
other
workflow
templates.
So,
for
example,
if
you
have
two
workflow
templates
on
your
cluster
and
you
want
your
cron
workflow
to
run
one
or
all
of
them,
that
is
perfectly
possible
so
and
the
back
balance.
Okay,.
A
Oh
right,
what
is
the
time
line
2.5,
so
we're
hoping
to
have
a
release
candidate
and
available
immediately?
It
was
hopefully,
hopefully
today,
but
still
turned
over
to
tidy
up
on
that.
So
what
is
the
timeline
for
GA?
Hopefully,
in
the
next
four
weeks,
I
would
say
crossing
my
fingers.
Both
fingers
are
crossed,
so
we're
going
back
so
from
my
head
depend
how
authorization
and
authentication
will
work
work
for
the
newer
UI.
Are
there
gonna,
be
different
roles?
Those
two
missions
by
Allah
to
answer
this
yeah.
D
E
E
Basically,
I'm
operated
ice
run,
our
go
UI
and
everyone
can
access
it
and
the
permissions
that
it
has
is
just
what
the
service
account
that
the
argos
server
is
running
is
so,
if
you
give,
if
you
want
to
give
everyone
the
ability
who
can
reach
the
UI,
the
ability
create
workflows,
you
can
do
that
by
running
it
in
the
in
the
hosted
mode
and
giving
it
like
the
all
the
privileges
it
needs.
On
the
flip
side,
you
can
also
run
it
in
basically
with
almost
zero
Roth,
like
give
it
a
service
count.
E
That
obviously
has
nothing
except
just
reading
the
workflow
controller
config
map,
and
if
you
do
it
that
way,
then
basically
the
Argos
server
will
bring
you
configure
with
the
auth
mode
so
that
it
relies
on
a
client
provided
token
to
perform
operations
on
behalf
of
the
requester
using
the
supplied
token.
And
so
the
idea
is
that
it's
similar
to
the
way
communities,
dashboard
works
and
if
you
configure
it
correctly,
you
can
run
it
with
like
basically
no
auth,
and
then
the
user
can
paste
their
token
like
a
bearer
token
into
the
UI.
A
Okay,
so
we
have
I,
think
one
more
question
from
Orion
I.
Think
there's
two
questions
here:
actually
I
will
I
will
ask
one:
I
will
do
them
one
after
another,
when
teaching
I'll
go
to
others
in
my
companies,
a
lot
of
computer
on
the
term
template
and
with
the
additions
of
workflow
template,
it's
even
more
confusing.
Have
you
thought
about
this?
Naming
collision
I
I
do
completely
confusing
I'm
relatively
new
to
the
project.
A
I
feel,
like
probably
the
cats
out
of
the
bag
at
the
moment
in
terms
of
template,
but
perhaps
we
could
do
something
about
improving
the
documentation
or
providing
better
kind
of
signage
in
the
user
interface
about
that.
So
maybe
we'll
take
that
away
and
think
about
it.
There's
some
nodding
in
the
in
the
room
here
with
me.
C
So
one
of
the
main
reasons
we
decided
to
implement
cron
workflows
independently
from
Argo
events,
is
that
we
wanted
to
have
developers.
We
wanted
to
allow
developers
to
get
started
on
this
very
quickly,
so
we
want
developers
to
have
to
also
work
through
argue
events
and
get
that
installed
and
get
that
connected
and
then
add
extra
permissions
for
the
separate
controller
with
the
con
wear
clothes
here,
the
it's
all
native
to
our
go
work,
clothes.
C
So
as
a
result,
it
works
in
a
much
more
quick
and
efficient
manner
that
will
allow
developers
to
get
to
get
started,
get
started
using
it
quicker.
There's
also
always
argue
events,
so
if
developers
feel
like
they
need
more
control,
they
can
also
just
use
that
are
events,
of
course.
Okay,
thank
you.
So
much
I.
C
E
And
I
think
the
question
is
around
them
in
integrating
our
go
events
into,
for
example,
the
UI
and
I
think
the
answer
to
that
is
I,
yeah
I
think
we're
open
to
that
I
think.
If
there's
enough,
you
know
requests
or
usage
and
then
it's
definitely
something
we'll
consider.
The
other
thing
I'll
add
about
the
use
cases
for
events
versus
cron,
so
cron
is
really
to
to
address
the
more
simpler
use
cases
of
scheduling.
A
Great
stuff,
okay,
so
Evan
asks
hang
on
no
I'm
getting
right.
Srivasa
forget
the
right.
Thank
you.
I've
heard
that
Argos
decay
is
also
in
the
offing.
Is
it
coming
with
V
2.5
as
well?
If
so,
what
language
this
TK
4
supports,
so
kind
of
stepping
stepping
back
and
I'll
go
2.5
will
introduce
the
API
server,
which
will
introduce
gr
PCM
points
along
with
the
proto
files
and
swagger
definitions,
so
you
can
produce
API
clients
to
those
in
terms
of
more
kind
of
first-class
support
and
people,
obviously
particular
keen
on
Python.
A
A
E
I
think
we'd
like
to
have
a
fish
Alesi
case,
but
then
like
they're,
the
are
garage
so
that
there's
people
don't
repeat,
work
I,
think
so
and
then,
even
though
we
have
the
we
do
have
the
G
profiles
for
the
GOP
server
I
think
they're
they're
still
gonna
be
need
to
have
some
kind
of
glue
on
top
of
the
generated
files.
So
you
know
I'd
love
to
see.
People
contribute
something.
The
agro
labs
like
a
Python
SDK
Java
with
their
ESL
actually
already
a
Java
SDK
in
our
go
labs
but
yeah.
A
Fantastic,
wouldn't
it
okay
I'm
kind
of
new
questions,
so
maybe
we
can
kind
of
keep
our
answers.
Kind
of
really
shortened
to
the
point.
If
we
have
time
with
the
GRP
sees
it
possible
replaces
work
for
mo.
Wasn't
like
a
programming
language
like
python,
but
protobuf.
I
think
the
answer
is
yes.
Okay,
thomas.
Does
our
great
guy
eventually
replace
on
go
events
with
web
cookie
gee
can
I
submit
an
arc
or
workflows
API?
Yes,
yes,
yes,
can
I,
create
work
for
today's
class
too
wide
and
later
use
it
across
with
different
namespace.
H
A
Create
issue,
then
we
can
realize
my
recommendation
is:
if
you
want
something
done,
make
sure
it's
an
issue
in
github.
Explain
what
yes
pretty
really.
What's
my
use
case
about
it
and
then
the
more
thumbs
up
it
has
them
the
more
likely
we
are
to
look
at
it
and,
of
course,
the
best
way
to
get
something
done
on
an
open-source
project
is
to
you
can
contribute
it
yourself.
I
was
watching
to
the
south.
A
Yeah
I
mean
we
can
I
mean
we
really
came
for
people
to
contribute
new
features.
So
isn't
particularly
one
and
you
actually
your
hands
dirty,
then
that
would
be
absolutely
wonderful
to
do.
Do
we
have
okay?
Is
it
going
to
be
a
workflow
status,
API
and
I
guess
I'd
be
returns,
a
lot
of
information
which
is
not
required.
G
G
E
A
You
know
have
a
look
at
the
code
yourself
and
you
know
if
you're
interested
in
doing
it
I'm
sure
we
can
do
something
soon.
As
this,
if
it's
been
done
before
them,
we
know
how
to
do
already.
Okay
request
for
an
SDK
for
Ruby
okey
dokey.
Okay,
thank
you,
Evan.
Thank
you
very
much.
Okay,
any
more
questions.
It's
gonna
give
you
can
of
10.
B
A
Big
thumbs
up
anything
you're,
particularly
interested
in
you
know,
so
we
need
to
know
what
is
important
to
you
and
you
need
to
tell
you.
You
know
that
if
you
tell
us
this
that,
obviously
the
best
way
to
do
it
if
you've
got
any
more
questions
that
come
to
your
mind,
you
hop
on
to
the
Argos
lack
charm
and
come
and
ask
us
there
we're
going
to
save
this
recording
and
put
it
on
to
youtube.
So
you
can
you,
can
you
can
review
it
again
and
or
review
the
features?
A
E
F
Building
a
custom
pipeline
that
requires
a
human
intervention
at
a
moment,
so,
basically
we're
going
to
send
out
a
message
via
slack
someone's
gonna
hit
it
button,
say
yay
or
nay.
When
we
resume
the
workflow,
we
want
to
know
whether
they
hit
yay
or
nay.
Okay,
we've
looked
at
different
ways
of
doing
that,
but
they
okay.
C
I
can
speak
to
this
a
bit
there's
actually
already
a
proposal
and
a
PR
that
that
sort
of
tries
to
tackle
this
problem.
You
should
be
able
to
I
think
it's
by
user
mark
nine
white.
You
should
be
able
to
look
at
it
on
the
PRS
and
issues
on
github,
so
this
would
be
a
great
time
to
add,
whatever
you
think
should
be
added
to
this,
to
make
your
use
case
work
with
this.
Thanks
for
bringing
this
up,
yeah.
E
E
It's
like
a
API
approval
likes
a
service,
and
basically
all
you
need
to
do
is
your
step
is
really
just
issuing
a
long
live
curl
to
this
command.
I
think
you
can
retry
it
forever
if
you
need
to,
and
then
it
hits
this
curl
of
this
a
service
and
that
service
connects
to
things
like
slack
and
email
and
stuff,
and
then
the
whole
point
of
the
service
is
just
to
to
do
approvals
in
some
way.
I'm,
not
sure
if
that's
something
you
looked
into,
but
I
can
find
that
link.
That
was
suggested.
F
E
A
Yeah
I
will
do
them
here
to
let
us
know
if
you're,
you
know
speaking
at
meetups,
and
you
want
us
to
promote
the
about-
are
going
your
your
or
you
know
things
you
want
to
promote.
We
can
we
can
tweet
them
for
you,
we
can
let
people
know
in
the
announcements
channel
as
well,
which
is
really
really
interesting.
Somebody's
doing
one
in
Germany
reasons
that
was
great
to
know
about
okay,
cool
great.
Thank
you
very
much.
Everybody
have
a
lovely
go.