►
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
B
B
So
events
had
63
contributors,
they
contributed
42
new
features,
71
bug
fixes
totaling
189
commits
we
had
over
4
000
monthly
visitors
to
our
documentation
website
and
we
grew
the
number
of
stars
by
30
year-on-year
to
1200
workflows,
had
200
contributors,
contributing
167
new
features
over
500
bug,
fixes,
which
is
nearly
a
thousand
commits,
and
that's
actually
one
commit
to
master
about
every
eight
hours.
So
every
eight
hours,
a
new
bug,
fix
or
feature
comes
into
arco
workflows
and
we're
not
actually
not
even
at
the
end
of
the
year.
B
We
gave
away
nearly
500
plushies,
though
I
have
found
out
today
that
some
of
those
plushies
are
actually
in
a
garage
in
san
diego.
So
there
are
some
more
available
and
we
have
nearly
10
000
stars
for
argo
workflows.
I'm
really
excited
I'm
really
hoping
argo
workflows
will
grow
to
be
10,
000
stars
and
more
in
2021
before
the
end
of
the
year,
but
even
better
it'd
be
awesome
if
it
grows
to
10
000
stars
during
argo
con
so
get
starring.
B
B
B
Argo
workflows
had
three
big
releases,
including
3.0,
which
may
even
be
the
biggest
relief
we've
ever
had
this
all
built
on
a
foundation
of
improved
performance.
We
find
that
users
these
day
are
running
not
just
100
node
workflows,
but
a
thousand
ten
thousand,
and
even
one
hundred
thousand
work
node
workflows
now
bear
in
mind.
Kubernetes
can
only
run
three
thousand
three
hundred
thousand
pods
in
a
cluster.
This
is
pushing
the
boundaries
of
what's
actually
possible.
B
With
kubernetes,
we
introduced
tighter
security
with
tls
and
client
authentication,
both
enabled
by
default
out
of
the
box,
and
that
version
3.0.
I
mentioned
previously
deeper
integration
with
argo
events,
new
api
endpoints,
making
easy
to
automate
argo
events
and
new
user
interface
features
the
ability
to
view
list
and
edit
both
event,
sources,
sensors
and
the
new
event
flow
view
which
allows
you
to
look
at
an
installation
of
argo
events
in
a
namespace.
B
B
The
chemistry
executor
is
as
powerful
as
fast
and
secure
or
more
powerful,
faster
and
more
secure
than
any
of
the
existing
executors,
but
it
also
enables
new
features
such
as
container
set
and
debug
ports,
which
I
think
are
really
exciting
features
coming
up
soon.
In
version
3.3,
we
introduced
three
new
template
types,
data
template,
http,
template
and
container
set
template
data.
Template
makes
it
easier
to
do
fan
out
workflows,
because
you
can
now
list
the
contents
of
a
bucket
and
then
run
a
single
task
for
each
of
those
items.
B
Http
makes
it
easier
to
integrate
with
third-party
systems
previously,
if
you
wanted
to
make
a
http
request
in
a
workflow,
you'd
have
to
spin
up
a
single
pod
per
http
request.
We
now
matrix
all
those
requests
into
a
single
pod,
making
it
much
much
faster.
This
was
a
feature
developed
in
collaboration
with
kubeflow
pipelines.
To
enable
some
of
the
use
cases
they
had
and
container
set
was
an
exciting
feature
introduced
in
version.
B
3.1
container
set
allows
you
to
run
a
set
of
containers
rather
than
a
single
container,
because
the
containers
will
run
in
a
single
pod.
They
share
process
namespace
and
volumes
allowing
those
containers
to
talk
to
one
another.
Much
more
easily
container
sets
are
also
cheaper,
because
pods
are
actually
quite
expensive
to
start,
and
you
don't
have
to
use
things
like
persistent
volume
claims
to
share
data
2021
was
also
the
year
of
the
python
sdks.
B
B
B
B
Let's
look
at
three
of
them
now
in
ci,
cd
it'll
enable
you
to
build
and
test
package
an
application
in
one
cluster
gate
that
builds
with
a
manual
approval
step
then
deploy
into
another
cluster
and
this
all
be
encapsulated
within
one
single
workflow.
Previously
you
needed
to
write
scripts
or
some
other
application
or
integration
to
achieve
this.
B
This
will
enable
a
hub
and
spoke
model
from
automation
where
the
workflows
that
do
that
automation
exist
in
one
cluster
and
they
can
reach
out
into
other
clusters
to
perform
the
specific
maintenance
tasks.
You
need
to
do
I'm
just
giving
these
as
three
examples,
but
I
think
there'll
be
a
whole
other
bunch
of
use
cases
we'll
be
able
to
do
with
multi-cluster.
B
The
next
thing
I'm
excited
about
in
2022
is
plug-ins.
We
don't
have
a
plug-in
system
for
argo
work
phones
today
and
I'm
actually
kind
of
more
excited
about
plug-ins
than
multi-cluster.
Let's
have
a
little
look
at
that
now.
One
of
the
biggest
challenges
about
extending
argo
workflows
today
is
it
takes
time.
Typically,
you'll
have
to
come
up
with
your
idea.
B
Learn
golang,
maybe
learn
react,
maybe
learn.
Typescript
learn
the
argo
workflows
code
base
submit
a
pull
request.
Have
that
pull
request
reviewed
iterate
on
it?
That
poll
request
then
gets
the
most
to
master
until
it's
ready
to
go
generally
available
now
that
that
first
bit
of
the
process
can
take
up
to
two
months.
B
So
the
whole
time
frame
from
introducing
a
new
feature
to
getting
into
production
could
easily
be
six
to
nine
months.
It's
a
really
long
period
of
time,
and
it
might
be
that
you
can't
even
introduce
that
feature
into
argo
workflows
because
your
company
is
unable
to
open
source
it
or
it's
based
on
propriety
technologies.
B
B
Let's
take
a
little
look
at
a
plug-in
in
detail,
so
this
is
kind
of
a
canonical
slack
plugin.
This
template
runs
and
sends
a
slack
message
and
you
can
see
how
the
plugin
is
formed
and
this
particular
type
of
plug-in
is
called
a
plug-in
template
and
there
are
other
types
of
plug-in.
But
this
is
just
a
very
useful
one
to
send
a
slack
message
and
you
can
see
how
this
leans
on
existing
argo
workflows
features
such
as
templates
and
also
templating
system,
to
insert
the
workflow
name
into
the
slack
message.
B
How
hard
is
it
to
implement
a
plugin?
I
hear
you
ask,
I
don't
you're
all
muted.
I
don't
hear
that
at
all,
so
this
is
39
lines
of
code
that
implement
the
entire
plugin.
It's
so
small.
You
can
fit
it
onto
a
single
slide.
B
B
B
It's
because,
as
we
developed
plug-ins,
we
spoke
to
people
about
them.
We
came
up
with
new
ideas
and
new
things
that
we
could
do
so.
For
example,
when
we
demonstrated
this
to
pratik
a
few
weeks
ago,
he
suggested
python
first
class
support
and
it
was
possible
to
implement
a
basic
version
of
python.
First
last
support
in
argo
workflows
in
about
90
minutes
from
from
his
idea
to
implementing
it,
and
if
we
want
to
do
that
today,
we're
not
talking
90
minutes,
we're
not
talking
90
days,
we're
talking
up
to
9
months.
B
So
it's
a
two
orders
of
magnitude
game-changing
feature!
That's
why
I'm
so
excited
about
it
now
I
said
four
things.
The
next
thing
I'm
quite
excited
about
is
mentoring.
It's
pretty
difficult
to
get
started
with
argo
workflows.
If
you
want
to
become
a
contributor,
so
we're
going
to
introduce
a
mentoring
program,
that'll,
allow
you
to
buddy
up
with
somebody
else
and
get
started
and
skill
up
really
really
quickly
and
ultimately,
I'm
hoping
that
we'll
be
able
to
go
to
google
summer
of
code
and
how
do
mentoring
there
as
well.
B
B
A
Yes,
you
can,
that
was
amazing
and
alex.
You
know
the
comments
in
the
chat
and
we've
done
a
survey
online
people
are
saying
they
want
to
know
more
about
road
map.
They
want
to
know
where
it's
going
so
please
feel
free
to
join
them
in
the
chat
and
continue
to
have
that
discussion.
If
y'all
haven't
filled
out
the
survey,
please
go
ahead
and
do
it
and
up
next
we
have
jesse
to
talk
about
more
maintainer
updates.
Thank
you.
So
much.