►
From YouTube: 2021-03-02 - KEDA Standup
Description
B
Hey
how's
it
going
and
I
am
going
to
have
to
leave
a
bit
early
because
I've
got
some
kids
who
I
need
to
help
get
ready
for
school,
but
we
can
start
maybe
I'll
go
first
and
then
tom
I
might
hand
over
hosting
permissions
to
you
and
then
I'll,
hopefully
jump
back
on
in
a
few
minutes.
B
D
B
Could
hear
it
glad
to
hear
it
all
right?
Well,
we
can
jump
right
into
it.
Then,
as
I
mentioned,
I'm
going
to
have
to
pop
up
here
in
just
a
second,
yes,
great,
good,
first
agenda
item,
which
is
congrats
zibinek
on
the
baby.
I
love
it.
Yeah
everyone
was
thrilled
about
that.
I'm
gonna
go
through.
Let
me
paste
the
meeting
notes
into
the
chat
as
well
for
folks
right
here.
B
Thank
you,
tom
he's
going
ahead
and
going-
and
I
realized
I
probably
need
to
come
in
here
too,
and
remove
a
bunch
of
our
old
notes
into
the
archive
so
that
we
don't
see
this
starting
to
lag
up
again
because
last
time,
when
we
had
like
a
year's
worth
of
agenda
notes,
it
went
really
far
so
I'll
start
really
quickly
with
an
update
and
then
we'll.
If
we
can
just
go
around
the
room
and
do
a
quick
intro
and
then
share
any
items
that
folks
want
to
cover.
B
I've
got
to
go,
help
get
some
kids
off
to
school
and
then
I'll,
hopefully
be
back
here
in
a
few
minutes
if
the
meeting's
still
running
so
I'm
jeff.
Thank
you
for
adding
my
name
already.
I
work
at
microsoft
in
terms
of
updates,
there's
a
few
things
all
updates,
oh
wow,
so
this
is
one
I'll
just
weigh
in
right
now,
yeah,
I
think,
moving
to
the
github
container
registry,
I'm
all
for
it.
I
actually
been
using
the
github
container
registry.
B
More
recently
for
some
of
my
new
stuff,
I
think
that's
a
great
idea,
the
updates.
From
my
side,
I
spent
a
little
bit
of
time,
there's
two
more
customers
who
are
in
azure
who
are
kind
of
enterprises
who
are
looking
about
using
cada
that
I've
been
helping
with
one
of
the
results
of
that
is
that
there's
a
few
folks
who
want
to
use
cata
with
cosmos
db.
B
B
I
shared
I'll
go
for
it,
sorry
to
interrupt.
Will
they
work
on
it
or
that's
the
the
intent
is
that
they
would
work
on
it.
They
might
work
on
it
and
then
hand
it
off
to
to
like
the
functions
team
to
to
kind
of
support
and
maintain
the
same
way.
B
We're
supporting
like
some
of
the
other
microsoft
sponsored
ones,
but
the
the
plan
right
now
is
that
they
would
actually
help
build
the
thing
and
we're
trying
to
figure
out
right
now,
if
they're
going
to
be
building
it
and
go
with
their
go
sdk
or
if
they'd
be
building
it
in.net
with
their.net
sdk,
and
it
would
be
an
external
scaler,
we
just
sent
them
some
info
yesterday
and
ahmed
helped
out
there
too.
B
The
other
I
shared.
I
don't
know
if
jonathan's
on
the
call,
I
don't
think
so.
I
shared
the
recording
from
last
week's
recording
with
a
few
folks
as
well
to
look
over.
It
looks
like
there's
a
pretty
common
pattern
between
wanting
like
http
scale
stuff.
I
know
we
have
the
http
k2scaler
that
I
I
think
the
last
email
I
saw
was
that
we
were
just
about
to
publish
that
blog
post,
but
anyway,
that's
it
for
me,
so
so
a
little
bit
of
working
with
cosmos
db,
a
few
working
with
customers.
B
But
with
that,
I'm
gonna
stop
sharing
I'm
going
to
jump
off
and
then
folks
can
keep
going
around
the
room
so
maybe
to
keep
the
ball
rolling
tom
I'll
hand
the
ball
off
to
you,
and
you
can
do
a
quick
intro
and
either
give
your
updates
or
we
can
just
go
around
the
room
and
then
just
go
straight
to
the
agenda
yeah.
So
my
name
is
tom.
E
Yeah,
so
can
you
oh,
can
you
hear
me
yeah?
My
name
is
stewarship.
I'm
from
microsoft.
The
mainly
contributing
country
is
for
the
scaled
job.
E
F
I'm
erin
I
work
for
microsoft
and
I'm
working
primarily
on
the
http,
auto
scaling
project.
C
A
Jeff
is
not
here
now,
so
if
you
want
to
discuss
something
about
chat,
I'm
just
kidding.
Okay,
I
added
a
couple
of
things.
So,
let's
just
dive
right
into
it,
then
so.
Last
week
I
think
we
opened
our
annual
review
for
cncf,
so
it
is
on
the
pull
request,
link
there.
A
If
you
are
interested
basically
every
year,
we
need
to
mention
what
we
did,
how
it,
how
it
compares
to
the
roadmap
that
we
committed
to
and
where
we
are
going
so
yeah,
that's
all
open
there
and
then,
as
a
follow-up.
We
finally
started
looking
at
our
incubation
proposal,
so
first
we
need
a
sponsor.
So
I
am
checking
now
with
liz
rice,
who
was
one
of
our
original
sponsors
for
going
to
sandbox.
A
If
she
wants
to
sponsor
us
going
to
incubation,
and
then
we
will
start
writing
that
document
as
well.
C
Any
questions
on
this
is
there
anything
else
needed
for
the
for
the
review
or
for
the
proposal.
A
For
the
annual
review,
they
just
need
to
go
through
it,
but
I
don't
see
much
issues
there
and
then
for
the
incubation
proposal.
Once
we
have
a
sponsor,
there's
the
whole
due
diligence
process,
but
that's
basically
it
so.
Hopefully
we
can
also
reuse
a
lot
from
the
annual
review
in
terms
of
the
content
to
open
the
incubation
proposal.
Okay,
but
most
probably
we
would
like
we
would
need
to
present
to
the
doc.
A
So
this
is
the
process.
Let
me
see,
did
this
change
again?
No
yeah
I'll
just
add
a
link,
but
basically
you
just
need
to
have
a
healthy
community
and
some
other
things
that
we
normally
already
have,
and
then
they
will
review
it.
A
It
okay,
then
I'm
I
didn't
join
the
last
stand
up,
so
I'm
not
sure
what
the
outcome
was,
but
I
was
wondering
if
we
should
start
documenting
the
integrations
that
we
have
with
other
projects.
For
example,
I
know
apache
airflow
uses
scada
to
auto
scale.
A
A
Any
other
ideas
people
would
think
it's
another
good
idea.
That
sounds
good
to
me
too.
A
F
A
Then
the
proposal
to
move
our
docker
images
to
github
container
registry,
so
we
already
have
jeff
and
jebenek.
Agreeing
I
see
ahmed
is
here
as
well.
Do
you
have
any
concerns
on
moving
them?
So
the
I
don't
know
if
you
saw
the
issue,
but
the
point
is
that
github
container
registry
gives
you
a
lot
more
metrics
than
docker
hub
and
given
the
rate
limiting,
they
announced.
J
A
Fine
yeah,
so
we
started
it
as
a
test
for
the
http
scaling
as
well,
and
this
is
screenshot
that
I
added.
A
I
think
you
should
see
this
tab
as
well,
but
basically
it's
just
as
simple
as
you
can
do
it
publicly
as
well.
It
gives
you
a
docker
pull
and
then
it
also
shows
you
the
pools
per
tag,
which
means
that
we
would
have
a
better
understanding
of
of
how
many
people
are
still
using
older
versions
and
newer
versions
and
github
will
also
give
you
more
thorough,
metrics,
like
what's
the
the
comparison
between
arm
and
linux
and
windows
and
all
of
those
architectures.
C
A
A
I
retagged
them
and
published
them
to
the
new
registry
so
that
all
the
versions
are
in
the
new
registry
and
then
people
can
already
migrate
or
not,
and
then
just
drop
a
note
that,
as
of
the
next
version,
we
will
only
publish
to
github
container
registry,
but
this
release
is
the
last
one
on
the
upper
hub.
So
you
don't
have
to
change
anything
yet,
there's
also
the
benefit
of
helm
charts.
We
can
change
it.
People
will
implicitly
use
the
new
registry,
so
if
you're
doing
yaml
deployments,
those
will
be
the
only
ones.
A
C
A
F
I
haven't
had
any
problems
for
what
it's
worth
using
it
for
the
http
project.
It's
been
pretty
seamless.
A
A
C
A
Okay,
so
I'll
take
up
those
issues
then
probably
also
migrate.
The
existing
images
you
yeah
I'll
just
write
a
quick
script
that
does
that
okay,
so
kubecon
you
we
may
or
may
not
have
a
session,
maybe
I'm
not
allowed
to
say,
but
maybe
we'll
have
one
which
is
good.
Finally,
a
kubecon
with
gator.
A
F
Sure
yeah
we
have
released
a
first
alpha,
which
is
the
core
functionality.
F
F
So
we
are
working
now
on
the
blog
post
that
you
see
tom,
put
a
link
in
there
for
that's
gonna,
be
an
announcement
for
the
alpha
one
we
just
mentioned.
Actually
the
images
for
alpha
one
are
up
on
github
container
registry
as
well,
and
I
think
those
are
all
the
the
basic
pertinent
details
of
this
thomas
king,
big
time.
F
I
don't
think
you're
slacking
on
this.
Oh,
yes,
so
yeah,
so
alpha,
one
was
technically
released
on
the
github
page,
the
github
releases
page
a
couple
weeks
ago.
So
since
then
we
are
moving
forward
with
some
of
our
alpha
2
features.
You
can
go
in
and
see
the
milestone
for
that
on
the
hdp
add-on
repo.
F
There's
some
discussion
over
the
scope
of
this.
I
think
we
might
talk
about
that
now.
The
scope
of
how
far
the
alpha,
how
far
the
hdb
addon
should
sort
of
reach
into
your
infrastructure
and
how?
How
many
things
it
should
create
for
you
versus
things
that
you
should
have
control
over
yourself.
F
But
we
probably
will
be
ready
with
an
alpha
2
release,
either
next
week
or
early
the
week
after
that,
and
the
goal
for
that
is
to
improve
stability
and
improve
the
general
user
experience
for
deploying
a
new
scaled
workload,
because
right
now
there's
a
matter
of
just
how
fast
it
happens
after
you
submit
your
http
scaled
object
and
also
how
many
things
you
have
to
manually
set
up
in
your
cluster.
F
A
I
think
it's
very
nice
piece
of
work.
Indeed,
the
only
thing
that
sometimes
comes
up
is:
do
we
also
create
the
services
and
the
ingresses,
and
all
of
that
or
do
we
only
support,
bringing
your
own?
I
can
see
value
in
both,
but
today,
noah
had
a
good
point
on
slack.
Is
we're
building
an
auto
scaling
solution?
A
So
is
it
our
responsibility
to
do
this,
or
is
that
part
more
of
a
separate
project,
maybe
that
we
carve
out
and
runs
fully
independent
or
not
because
clearly,
this
is
it
is
a
nice
thing
to
have
certainly
if
you're
new,
but
I'm
not
sure
me
personally,
if
it,
if
it's
part
of
cada
or
not,
and
that's
not
not
an
offender
or
anything,
but
merely
just
do
one
thing
and
do
one
thing
well
and
I'm
I'm
interested
to
hear
what
the
others
think
about
it.
F
I'll
just
throw
out
there
my
plan
is
to
build
a
layer
on
top
of
this,
it's,
like
you,
said
tom
different,
different
repository,
probably
outside
of
the
keto
org,
so
whatever.
If
we
take
stuff
out
of
this,
if
there
are
things
in
this
that
are
out
of
scope,
it
won't
be
thrown
away
in
case
that
is
off-putting
to
anybody.
So
if
anyone
thinks
that
there
are
things
that
are
out
of
scope
of
this
project
that
are
in
the
project
now,
rest
assured
it
won't
be
just
thrown
in
the
trash
cans.
J
K
J
Would
work
because
somebody
like
the
service
needs
to
point
to
the
http
activator
rather
than
the
actual
deployment
that
lands
on,
because
otherwise
we
won't
be
able
to
intercept
requests.
A
It
twofold
you'll
have
a
service
for
cada
for
for
receiving
okay.
I
get
two
points.
F
So
right
now,
right
now
the
operator
as
it
stands
in
the
alpha
one.
F
The
operator
has
to
create
the
the
activator
workload,
the
deployment
for
the
activator
and
the
service
for
the
activator
and
an
external
scaler
and
a
service
for
the
external
scaler,
and
that
is
because
the
external
scalar
may
be
maybe
obvious
why
it
has
to
create
the
service
for
that,
but
for
the
activator.
We
would
call
it
the
interceptor
in
this
case,
but
either
way
that
that
workload.
F
The
external
scaler
needs
to
talk
to
that
as
well.
So
it
actually
creates
two
services,
one
for
the
external
scalers
to
talk
to
it
and
the
other
than
for
for
ingress
from
the
public
internet.
So
to
me
that
seems
like
oh
and
one
of
I
think
one
of
the
things
that
prompted
this
discussion
is
there's
a
pr
in
there
now
to
also
create
an
ingress
to
route
traffic
to
the
public
proxy
server
in
that
that
activator
component.
F
F
But
then,
beyond
that
we
might
be
able
to
say
I
actually
am
starting
to
think.
Even
though
I
submitted
the
pr
to
do
the
ingress
part,
I
actually
am
starting
to
think
we
can
say.
Okay,
the
operator
has
created
the
bare
minimum
functionality
for
you.
You
are
able
now
to
route
traffic
to
this
service.
F
I
I
am
starting
to
think
that
we
can
end
the
scope
there
and
then
add
a
documentation
that
shows
people
fairly
easily
how
to
get
their
traffic
into
the
system.
At
that
point,.
A
Yeah
me,
as
a
user,
I
would
prefer
that
I
can
still
manage
my
own
service
and
ingress
and
that
I
can
just
plug
in
keira.
Let's
say
personally.
H
Hey
I'll,
introduce
myself,
I'm
ryan
new
to
the
project,
just
for
my
howdy
from
a
user
perspective,
I
like
both
tracks
one
having
a
an
easy
like
get.
This
working,
get
everything
up
and
running
for
me,
so
that
I
can
just
understand
how
it
works.
H
Creating
you
know
having
an
option
whether
it's
an
example
or
or
built-in
option
to
have
it
build
everything
for
me
end-to-end,
but
personally,
most
of
our
services
are
going
to
be
exposed
through
a
multi-cluster,
ingress
or
or
something
that's.
Not
the
standard
single
cluster
ingress
so
being
able
to
have
options
to
bring
my
own
would
would
be
something
that
would
be
be
useful
for
for
our
work.
Our
use
case.
F
Another
question
ryan
that
you
just
reminded
me
of
is
right.
Now
we
have
the
operator
that
you
can
just
give
it
an
image
name
and
it
will
automatically
even
create
a
deployment
for
you
and
there's
a
pr
in
now
that
you
can
turn
that
behavior
off
and
just
have
the
operator
route
to
an
existing
deployment.
F
So
we
might
want
to
take
completely
remove
the
option
to
have
the
the
operator
create
a
deployment
as
well,
because
that
that's
kind
of
another
thing
that
feels
like
it
may
be
out
of
scope.
Is
it
that
it's
not
dealing
with
auto
scaling
or
http
routing
anymore.
A
Yeah,
I
think
that
the
deployment
side
is
definitely
one
that
I
would
leave
out,
maybe
because
typically
deployments
you'll
need
to
add
environment
variables.
You
need
to
do
mounting
and
all
of
that
stuff.
So
if
we
support
it,
we'll
need
to
support
a
lot
of
different
scenarios
and,
frankly,
for
one
reason,
so
I
agree
there.
F
So
I
can
propose
this:
I
can
propose
that
we
define
a
scope
and
the
scope
is
is
expressed
in
relation
to
where
traffic
starts
and
ends
in
the
system
and
where
http
traffic
starts
and
ends
in
the
system.
J
C
Nothing
makes
sense
to
me,
and
I
guess
that,
like
this
automatic
creation
of,
for
example,
the
deployment
from
the
from
the
image
or
or
the
the
other
things
is
maybe
more
of
in
charge,
some
kind
of
cli,
or
something
like
that.
Basically,
that
maybe
that's
yeah
built
on
top
of
on
on
top
of
the
core
core
solution.
C
So
I
I
kind
of
like,
like
the
the
shape
that
aaron
just
described,
that
the
the
distinction
between
the
traffic
inside
the
cluster
yeah.
So
I
agree.
J
A
A
J
I
mean
I
feel
for
the
service
like
we,
as
I
mentioned,
I
think
either
the
operator
has
to
be
the
one
creating
it
so
that
it
knows
where
to
route
the
traffic,
so
it
does
the
interception
or
activation,
or
if
you
allow
the
user
to
create
it
themselves,
then
there's
gonna
have
to
be
some
convention
about,
like
what
selector
does
that
service
uses,
so
that
the
traffic
goes
through
http
components
and
data
to
know
to
activate
it
right.
F
The
the
service
that
it
creates
is
routing
to
a
deployment
for
internal
infrastructure
that
it
has
to
create.
So
I
agree
with
you
ahmed.
You
should
also
create
that.
J
A
So
it
should
be
as
straightforward
as
possible.
We'll
just
need
to
find
a
way
where
it
is
easy
to
do.
F
F
F
So
unless
we
add
the
activator
pattern
to
this
system,
we
could
give
someone.
This
is
the
naming
convention.
You
need
to
create
a
service
and
an
ingress
using
this
naming
convention,
and
now
your
traffic
will
be
routed
properly.
F
A
Maybe
it
would
be
useful
if
we
create
some
example
scenarios.
Imagine
you
have
this
http
application.
Hopefully
there
are
already
examples,
but
if
we
deploy
this
and
then
we
add
auto
scaling,
what
is
the
impact?
A
Another
reason
why
that
would
be
useful
is
also
to
explain
to
people
more
in
a
walk-through
fashion
of
a
bit
similar
like
the
samples
we
have
today
hey.
This
is
the
the
non-auto-scaled
application
in
order
to
add
auto
scaling.
You
need
to
do
this
and
this
and
this
and
that's
how
it
works.
Because
then,
maybe
we
have
more
concrete
examples
of
of
the
steps
that
need
to
be
taken.
A
J
A
A
A
An
ingress
no,
no
just
as
kind
of
documentation
or
walkthrough.
We
imagine
we
have
this.
A
This
nginx
deployment
that
we
have
today,
which
is
exposed
to
a
service
now
we'll
apply
auto
scaling,
and
then
it
basically
shows
what
you
need
to
do
and
what
what
is
being
created
or
what
the
steps
are,
because
the
examples
we
have
now,
I
don't
have
it
open
anymore.
It's
just
hey.
This
is
the
http
scaled
object.
You
need
to
configure
it
like
this
and
this,
but
maybe
if
we
have
the
more
those
high
level
overviews,
that
would
tell
the
story
better,
but
maybe
it's
just
me
yeah.
I
don't
know.
A
J
Oh,
I
wanted
to
ask
about
the
next
release.
When
is
that
I
have
a
few
items
that
I'm
working
on
that
I'd
like
to
get
in
before,
but
let's
kiss
when
it's
going
to
happen,.
A
Well,
actually,
I
was
recently
thinking
if
we
should,
I
think
we
once
discussed
it,
but
if
we
should
do
frequent
or
scheduled
releases,
maybe
not
really
have
them
as
hard
deadlines
but
more
to
stimulate
ourselves
to
release.
Because,
honestly,
I
don't
really
think
about
it.
Much
until
I
see
the
whole
changelog
and
thinking
maybe
it's
time
to
ship.
This.
D
C
K
C
C
So
so
don't
why
why
don't
we?
Why
don't
we
basically
discuss
the
release
next
next
standard
and
if
everything
is
going
well,
we
can.
We
can
release
that
big.
If
not,
we
can
postpone
sounds
good
yeah.
A
Should
we
start
doing
maybe
more
frequent
releases,
or
do
we
just
do
them
at
hog?
Now.
A
Perfect
anybody
else
who
wants
to
share
some,
maybe
one
more
thing,
chiosi
about
the
scale
jobs.
So
I
was
wondering
if
you
are
up
for
contributing
metrics
for
those
so
that
people
can
use
prometheus
to
monitor
them.
The
discussion
we
had
was
more
about
somebody
who's
looking
at
parsing
the
logs,
because
he
he
doesn't
have
the
metrics,
but
ideally
he
could
use
that.
E
A
I
don't
know
if
we
have
an
issue
on
that
one,
but
maybe
we
should
add
one
to
to
keep
track
of
that
and
okay.
Thank
you
for
looking
at
that
chucy
and
thank
you,
everybody
for
joining
and
see
you
in
two
weeks.
E
B
Weeks
it
I
got
back
here
for
the
last
few
minutes,
but
yeah,
that's
what
you're
saying,
but
actually
you've
been
just
here
snooping
around
yeah.
I
was
new,
my
my
daughter.
She
just
started
going
back
to
school
and
she,
when
she
did
this
when
she
was
in
kindergarten,
but
she
gets
really
anxious
when
she
goes
to
school,
and
so
when
she
was
in
kindergarten,
sometimes
she
gets
so
anxious
that
she
would
throw
up
which
usually
the
school
was
like
yeah.