►
From YouTube: Kuma Community Call - November 25, 2020
Description
Kuma hosts official monthly community calls where users and contributors can discuss about any topic and demonstrate use-cases. Interested? You can register for the next Community Call: https://bit.ly/3A46EdD
A
Okay,
so
let's,
let's
start
quickly
so
welcome
to
the
bi-weekly
or
twice
a
month,
the
monthly
progress,
the
better
wording
here,
community
meeting
of
kuma-
I
see
some
new
faces
here.
I
hope
this
is
the
result
of
our
participation
in
kubecon
and
specifically
the
boot
that
we
had
there.
A
So
I
will
start
with
a
quick,
quick
wrap-up
of
what
happened
last
week
for
us
and
also
what's
what's
coming
next
in
terms
of
events,
at
least
to
the
best
of
what
I
know.
If
there's
something
else,
please
comment
here.
B
A
Good
so
last
week
was
kubecon,
which
was
a
big
event
for
everyone:
that's
interested
in
okay.
How
do
I
open
this
thing
for
everyone?
That's
that's
interested
in
in
the
cloud
native
technologies.
A
So
not
only
service
mesh,
but
everything
around
it.
So
it
started
with
a
day
of
pre-conference
events.
Collocated,
let's
say,
although
you
know
the
virtual
sense,
that's
a
little
bit
strange
like
how
do
you
collocate
a
virtual
event,
but
okay,
it
was
the
same
platform
probably
so
our
participation
was
specifically
on
the
service
mesh
con
mark.
Who
did
the
opening
remarks
and
then
there
was
also
a
talk.
A
I
was
not
able
to
attend.
I
don't
know
if
anyone
was.
I
guess
that
some
of
these
may
be
published.
I
don't
know
what
I
know
that
the
main
conference
is
going
to
be
published,
but
in
any
case
the
slides
are
here,
so
you
can
go
check
at
least
what
what
marco
is
saying
here.
I
guess
that's
at
least
to
the
people
that
are
for
longer
with
us.
These
are
probably
not,
let's
say
breaking
news,
yet
I
I
guess
it
was.
It
was
an
interesting
thing
to
see.
A
A
I
hope
that,
for
the
next
cube
con
we'll
probably
have
something
in.
I
was
actually
a
track
chair
of
this
of
the
service
mesh
track,
so
they
were
quite
good
talks.
Also,
we
had
a
couple
and
then
the
selection
was
very,
very
very
hard.
So
what
what
was
left
here?
These
four
were
just
a
very,
very
small
fraction.
I
cannot
share
my
number,
but
it's
a
small
fraction
of
what
was
actually
selected.
A
I
was
I
was
on
some
of
these.
It's
it's.
It
was
really
interesting
when,
when
the
community,
when
the
recordings
appear,
I
definitely
suggest
that
you
go
and
and
check
this
interesting,
also
to
compare
our
experience,
what
we
do
with
what
everyone
else
is
doing,
especially
this
one
I
I
was
not
able
to
attend
it,
but
I'm
when
I
was
reading
the
proposal
I
was,
I
definitely
have
a
high.
A
I
think
that
this
was
one
of
the
highest
ranked
talks
so
yeah,
I
don't
know,
if
I'm
I
should
say
this,
but
yeah
in
general,
it's
clearly
seeing
100
k,
line
and
void
configuration
in
production.
That's
that's
definitely
something
that's
that
is
instead,
it's
interesting
for
anyone
in
the
dome
domain.
A
Apart
from
that,
we
had
our
boots,
which
we
saw.
We
saw
some
interesting
discussions,
people
coming
asking
things,
and
I
also
hope
that
this
this
is
one
of
the
one
of
the
results
of
this.
Is
you
know
this
spark,
let's
say
the
spike
in
the
interest
of
tokuma.
So
last
week,
as
we'll
speak
later,
we
announced
our
100
release,
so
we
received
a
lot
of
feedback.
A
A
Okay,
I
saw
that
was,
let's
say,
more
or
less
our
participation
in
last
week
conference.
So
there
were
a
lot
of
interesting
things
going
on.
I
think
what
actually
concerns
the
the
community
here
is
one
of
the
remarks.
I
think
it
was
from
constance
on
one
of
the
keynotes.
A
So
as
a
recap
of
2020,
they
announced
that
they
they
have
accepted
a
lot
of
a
lot
of
new
projects.
They
gave
a
number-
I
don't
remember
it,
but
the
fact
everyone
can
see
it
in
the
in
the
on
the
cncf
slide.
But
one
of
the
interesting
things
is
that
they
also
plan.
I
don't
think
that
they
have
done
it
so
so
much
so
far.
A
They
plan
to
have
some
of
the
sandbox
projects
like
pass
through,
but
probably
all
of
the
sandbox
projects
pass
through
a
very
serious
evaluation,
and
maybe
some
of
them
will
will
be
dropped.
I
don't
think
that
kuba
is
endangered
in
this,
but
our
plans
are-
and
we
work
very
strongly
towards
incubation
for
this.
A
We
need
use
cases
so
if
any
one
of
you
is
using
kumar
improv,
especially
but
okay,
maybe
some
in
other
other-
I
don't
know
testing
pipelines,
developer
pipelines
using
using
in
any
form
definitely
reach
out
to
us.
We
are
interested
to
gather
and
get
some
case
studies
that
we
can
we
can
put
forward,
and
this
will
definitely
help
us
move
the
project
into
the
next
stage.
A
Okay,
I
think
that
I'm
done.
Finally,
with
this
10
minutes
already,
okay,
quickly,
going
of
going
forward
cucumber
virtual
con
europe
was
was
announced
the
the
call
for
papers
or
presentations
or
proposals
these
proposals.
Okay,
is
it's
open
now,
it's
the
deadline
is
13th
of
december,
which
is
slightly
more
than
two
weeks.
So
we'll
have
one
more
call
before
that.
I
think
on
night
of
december,
is
our
next
call
and
yeah
you
can
see.
A
Kuma
is
on
the
list
here,
so
anything
that
concerns
any
of
these
projects
and
specifically
kuma.
What's
what
we're
interested
here
is
welcome.
There
are
the
topics
here,
service
meshes
here
and
so
on
and
so
forth.
So
I
would
say
being
on
the
program
committee
on
on
the
last
couple
of
cube
cons.
A
I
definitely
see
that
the
most
interesting
talks
are
cross
company
cross
teams
talks
like
especially
if,
if
you
are
able
to
combine
when
you
submit
a
talk,
combine
a
vendor
like
someone
that
is
developing
a
solution,
plus
a
user
and
present
the
use
case
and
how
how
how
this
particular
solution
actually
solved
any
real
use
cases.
That's
that's
that's
very
much
useful
and
helps
into
the
talks
being
accepted
so
again.
A
If
anyone
is
interesting
to
do
interested
to
do
like
a
common
talk
together,
we
we're
definitely
here
and
so
what
would
support
this?
Okay
with
this,
I
will
stop
being
boring
and
release
to
the
more
exciting
part.
Okay.
So,
as
you
said
last
week,
we
released
one
zero
zero,
which
was
a
long
weighted
version
for
us.
A
We
we
put
a
lot
of
effort
into
bringing
this
to
together.
The
changelog
is
pretty
big.
I
think
that
it
was
in
the
announcements,
but
I
think
about
70
pr's
were
in
a
lot
a
lot
of
of
new
features.
There
are
some
some
also
some
breaking
changes.
A
We
still
haven't
updated
our
upgrade
md,
but
I
was
I
was
preparing
this
and
going
through
the
braking
changes
and
it's
yeah.
We
definitely
would
suggest
that
this.
My
my
opinion
here
is
that
if
you're
running
kuma
in
production,
you'd
better
consider
reinstalling
from
scratch,
I
know
that
jacob
was
working
with
one
of
our
users
or
maybe
a
couple
of
our
users.
I
don't
know,
but
he
came
with
some
some
interesting
feedback
from
doing
the
the
parts
of
upgrades
jacob.
Do
you
have
anything
to
share
about
this
here.
C
Yeah
I
mean
yeah,
we
introduced
many
changes,
so
yeah
we
need
just
upgrade
md
with
the
instructions,
but
overall,
like
the
biggest
breaking
change
on
kubernetes.
Is
that
the
resources
like
the
most
of
our
resources
are
now
cluster
scoped
right.
So
you
need
to
kind
of
dump
your
resources
and
delete
the
crds
upgrade
kuma
and
then
apply
them
again.
That's
that
that's
the
biggest
one,
but
we
try
to
automate
as
much
as
we
could
on
kubernetes
right.
So.
A
Okay,
so,
as
you
see,
there's
just
I
think,
even
less
than
a
week
between
the
releases
of
one
zero
zero
zero
one
zero
one,
there
was
some
quickly
found
problems
that
we
wanted
to
fix
for
the
users
to
to
ensure
their
smooth
part,
and
we
have
even
more
coming.
So
that's
why
we
are
planning
already
for
next
week
to
have.
A
Okay,
yeah,
we
have
one
zero
two.
I
guess
that
by
okay
before
the
end
of
next
week,
for
sure
we
will
have
one,
I
don't
want
to
put
a
fixed
date.
A
We
will
do
our
best
to
do
wednesday,
thursday,
some
sometime
this
way,
but
but
we
definitely
fix
this.
This
is
this
is
mostly
to
say
that
we
hear
your
own
slack.
We
try
to
help
as
much
as
possible
and
so
the
most
let's
say
critical
issues
we
try
to.
We,
we
do
prioritize
and
put
put
them
on
our
immediate,
fixed
list
and
we
release
versions
fast.
I
don't
know
if
marco
is
with
us.
A
Not
he's
not
with
us,
but
essentially
our
plan
is
that
for
the
remaining
weeks
within
2020,
which
I'm
sure
that
most
of
the
people
would
just
like
to
to
be
already
yet
but
yeah
for
the
for
the
remaining
time
in
2020,
we
will
not
will
not
do
anything
major
in
terms
of
a
release,
so
we
will
be
focusing
on
fixing
issues,
so
yeah
keep
them
coming
to
us
on
slack.
A
We.
We
gather
these
and
and
try
to
address
okay
with
this,
I
don't
have
much
of
my
agenda
and
I
was
not
really
very
polite
in
the
beginning,
because
again
we
had
some
new
faces
here.
So
first
of
all,
we
have
a
new,
a
new
member
on
the
team
on
the
kong
side.
It
is
parts
of
bart,
would
you
please
say
a
couple
of
works.
A
D
Everyone,
I'm
sorry
yeah
yesterday
was
my
first
day
at
kong
and
I
will
be
working
with
nikolai
and
jakub
on
kumar,
so
I
will
be
here
more
and
before
I
was
working
at
vmware
as
you
can
see
in
my
shirt,
which
I
didn't
change
yet
and
I
was
working
at
a
kubernetes
project.
Mostly
my
field
of
expertise
was
the
infra
and
I'm
also
the
co-chair
of
kate's
intro
working
group
still.
A
So
anyone
that
feels
that
they
want
to
introduce
themselves.
I
don't
want
to
to
point
here.
I
mean
if
someone
wants
to
say
a
couple
of
words.
This
is
the
time
and
then
probably
we
have
an
open
q,
a.
E
Hey
guys,
I
would
like
to
introduce
myself.
My
name
is
hakeem
and
I
come
from
an
app
development
background
and
I'm
trying
to
set
up
a
service
mesh
for
a
use
case
after
a
lot
of
googling,
I
came
across
kuma
and
since
past,
from
past
couple
of
weeks,
I've
been
trying
to
learn
service
mesh
technology
because
I'm
new
to
all
of
this,
but
at
the
same
time
I've
made
some
progress
and
I'm
having
a
lot
of
fun.
You
know
working
with
us.
F
Yeah,
maybe
I
could
introduce
myself
as
well,
so
I've
been
working
with
some
of
you
already,
so
I'm
greg
I'm
working
for
for
a
company
called
arctic.
F
So
yes,
we,
we
are
independent,
a
consulting
company
and
I'm
working
with
a
big
telco
here
in
canada,
who
has
managed
us
to
discover
what
are
the
possibilities
with
kumar,
so
I'm
exploring
the
hybrid
setup
so
on
the
last,
the
last
thing
I've
been
working
on
is
trying
to
extend
the
kuma
demo
to
demonstrate
external
services
to
offset
the
database
to
a
ubuntu
machine
and
trying
to
automate
also
the
multi-zone
installation.
F
So
on
that
note,
thank
you
a
lot
for
simplifying
the
the
setup
process,
because
now,
with
the
auto
ingress,
I've
been
able
to
have
a
github
action
pipeline
that
spins
up
a
cluster
and
with
the
auto
ingress
by
using
the
m
chart,
I'm
able
to
do
a
real
end-to-end
test
that
spins
up
two
clusters
sets
one
as
a
global
control
plane.
Another
one
has
a
remote
and
then
setting
up
the
cluster
so
setting
of
the
the
service
mesh.
F
So
that's
what
I'm
working
on
and
lately
so
I'm
working
with
ilya
as
well
on
a
small
bug
regarding
vips.
So
by
the
way,
maybe
you
want
to
update
the
thread,
because
maybe
marco
will
let
you
have
some
feedback
on
that.
B
Oh
yeah,
I
will
remove
the
feedback.
I
mean
I
today
fixed
the
parts
of
the
park
that
we
found
yesterday
with
collision.
If
you
remember
so,
I
hope
it
will
be
soon
merged
and,
second
part
is
still
in
progress.
I'm
trying
to
figure
out
this
external
services,
okay,
on.
B
C
Thank
you
thanks,
thanks
for
the
for
the
feedback,
I'm
very
happy
to
hear
that
the
improved
multi-zone
workflow
worked
well,
because
I
I
actually
implemented
this
so
so.
Thank.
E
Yeah,
I
would
also
like
to
extend
my
introduction
and
say
the
same
thing.
I
really
like
the
automatic.
You
know
with
this
new
version
that
that
was
super
helpful.
E
Having
said
that,
I
am
struggling
with
a
few
smaller
things
which
part
of
that
is
might
be
because
I'm
new
to
this
service
mesh
world
but
at
the
same
time
I'm
trying
to
set
a
multi-zone
similar
case.
For
the
most
part,
I've
been
able
to
do
it
successfully,
but
I
we
have
a
use
case
where
the
data
plane
runs
on
a
very
small
embedded
device
and
I
need
to
lift
out
the
control
plane
up
in
cloud.
B
A
So
lately
we
have
been
hearing
okay,
so
apparently
the
multi-zone
is
the
is
the
how
to
say
one
of
the
the
big
feature
that
people
are
seeing
in
in
kumas
like
let's
say
most
the
most
important,
and
we
see
a
lot
of
interest
into
this,
and
also
a
lot
of
the
problems
that
we
see
are
related
to
some
kind.
Some
kind
of
let's
say,
communication
problems,
no
matter
if
it's
a
firewall,
if
it's
just
you
know,
misconfiguration
of
ports
of
host
names
or
whatever.
A
So
we
we
definitely
started
some
internal
discussion
of
how
we
can.
A
We
can
improve
the
tooling
or
messaging,
or
I
don't
know,
maybe
with
there's
obviously,
obviously
a
problem
there
that
we
need
to
improve
so
any
any
kind
of
suggestions,
even
on
slack
or
like
things
like.
It
would
be
great
if
I,
if,
if,
if
I
had
seen
this
at
this
step,
but
I
find
it
like
to
stand
two
steps
down
there
for
I
have
to
get
back
to
my
firewall
and
et
cetera,
et
cetera.
A
So
I
don't
know
anything
like
this
from
from
user
point
of
view,
could
could
help
a
lot
on
on
taking
the
decisions
of
what
we
want
to
improve
there,
but
definitely
we
do
recognize
that
there
is
a
a
big.
Let's
say,
like
a
big
amount
of
the
of
the
problems
are
related
to
this,
at
least
the
initial
ones.
Let's
say
when
deploying
okay.
A
G
I
I
have
a
question
so
I
wanted
to
okay
yeah.
I
wanted
to
learn
a
little
bit
more
on
multiple
proxy
on
the
same
virtual
machine
for
the
for
the
universal
plane.
Is
there
any
limitations
there
or
any
kind
of
work
which
is
to
happen,
or
can
I
try
that
out.
B
F
F
No,
it
was,
it
was
on
a
vm.
So
on
a
vm,
I
set
up
the
the
ingress
data
plane
and
I
was
trying
to
have
one
data
plane
per
service
and
when
I
was
trying
to
spin
up
the
second
one,
the
promises
port
was
colliding
with
the
first
one.
So
it
would
not
come
up,
but.
C
Yeah,
so
that's
true
yeah,
you
can
go
both
ways.
You
can
go
with
many
services
with
one
data
plane
right,
but
the
trade-off
is
that
then
you
have
one
certificate
for
both
services
right.
So
you
don't
have
this
granule
control
over
the
traffic
and
I
believe
yes,
you
can
do
like
two
proxies
on
one
machine.
It's
just
a
matter
of
parts
right.
You
need
to
be
careful
with
parts,
so
you
don't
collide
parts
with
two
comma
dps,
but
other
than
this.
I
believe
it
should
work.
C
No,
not
really.
We
have
this
logic
implemented
in
comma
dp
that
we
try
to
pick
the
free
port
starting
starting
from
1500.
F
B
B
C
I
think
I
think
30
000
right,
so
it
also
should
be
handled
automatically
by
comma
dp.
If
not,
you
can
always
fall
back
to
just
passing
another
import
to
to
kuma
dp
run.
G
Okay,
is
there
a
possibility
to
block
the
admin
port
at
all.
C
G
I
know
so
why
do
I
say
that
is
because
the
onwise
is
living
on
the
application
developers
vm
and
then
he
gets
an
access
to
the
admin
port
and
he
can
misconfigure
it
or.
G
C
C
Yeah,
I
understand
yeah,
I'm
thinking
what
can
he
actually
change?
C
G
Or
can
he
he
can
see
the
certificates.
C
G
G
Okay,
so
maybe
I'll
try
that
out
and
I'll
I'll
come
back
to
you
I
mean
looks
like
we
should
not
have
a
lot
of
issue,
except
that
quick,
quick,
quick
thing.
Other
thing
is
the
the
slack
channel
in
the
slack.
Basically,
I
was
having
a
conversation
around
around
the
stored
procedure,
and
I
see
that
you
are.
You
are
going
to
basically
have
more
implementation
around
that.
Can
you
shed
some
light
around
that,
as
in
what
more
functionalities
are
you
trying
to
do.
C
C
C
But
you
know
sometimes
it
just
it's
not
possible
right,
and
especially
with
the
database,
the
only
alternative
that
we
saw-
and
maybe
you
have
some
better
idea
how
to
do
this
is
to
just
pull
the
database
for
the
for
the
for
the
ch
for
the
changes
right
and
imagine
that
you
have
like
one
control
plane
with
5
000
of
data
planes
with
the
refresh
of
one
second
right,
so
we
are
doing
5.
000
requests
a
second
to
a
database
just
to
see
if
anything
changed.
C
So
our
idea
was
that
okay,
we
can
just
use
this
stored
procedure
and
wait
for
the
database
to
to
to
notify
us
for
this
right.
So
it's
not
like,
like
it's,
not
our
intention
and
plan
to
move
more
and
more
logic
into
star
procedure
or
something
like
this
right,
but
for
this
one
we
felt
that
there
was
no
other
same
choice
to
do
this.
G
Yeah
yeah,
I
mean
I
understand
that
perspective.
I
mean
you
had
to
do
it,
but
do
you
think
there
was
a
alternate
implementation
possible?
Where
is
where
an
where
an
event
can
be
sent
to
all
the
other
remote
remote
cps,
as
in
whenever
a
database,
not
databases,
and
whenever
a
data
plane
connects?
It
is
already
it's
an
event
and
it
does
not
happen
every
second.
So
if
that
can
be
channelized
properly,
maybe
we
can
we
can.
G
We
can
remove
the
notification
logic
that
we
have
written
with
the
database
so
that
way.
C
G
C
That's
not
true,
I
mean
the
goal
of
the
control.
Plane
is
to
have
a
dynamic
call
configuration
right,
so
you
apply
traffic
permission
or
traffic
route
or
whatever,
and
it's
just
applied
in
in
immediately
right.
That
was
the
goal
of
the
of
the
command
that
you
don't
have
to
like.
You
know,
apply
and
restart
unvoice,
or
do
something
like
this.
G
Sure
all
right,
maybe
I
need
to
look
into
it
to
understand
more,
but
I
mean
it
is
still
a
configuration
which
is
coming
in
whether
it
is
getting
triggered
by
coma
ctl
or
whether
it
is
getting
triggered
by
a
data
plane
and
ultimately
your
control
plane
will
definitely
know
that
there
is
a
configuration
change
done
and
if
you
can
have
a
separate
channel
to
redistribute
this
knowledge
to
all
the
other
interested
control
plane
nodes
we
can
potentially
get
out
of
this
is
our
multi-zone.
C
Put
no,
it's
not
even
called
connected
to
to
multi
zone
deployment
right,
it's
even
with
the
standalone
yeah.
I
understand
what
what
what
you
are
saying,
because
the
problem
is
okay,
we
could
kind
of
emit
the
event
when
you
apply
something
on
the
control
plane,
but
then
the
problem
would
be
that
if
you
have
like
five
instances
of
the
control
plane,
this
event
would
be
only
emitted
in
in
one
instance
right,
so
you
need
to
kind
of
synchronize
this
to
to
other
instances.
G
Right
so
in
my
previous
organization,
the
way
we
solved
it
by
by
having
a
configuration
hash.
So
whenever
the
configuration
needs
to
change,
we
used
to
change
the
configuration
hash
and
and
and
and
whenever
you
do,
a
heartbeat
to
the
master
controller
or
the
global
control
plane,
you
get
the
hash
and
if
there's
a
change
in
the
hash,
you
go
and
download
the
delta,
but
I
mean
I'll
leave
it
on
to
you
guys.
I
mean
whatever
feels
good
for
you.
G
I
mean
because
because
this
particular
peer
is
not
helping
me
and
and
and
it
has
not
allowing
me
to
make
any
more
updates
that
you
do
any
more
pr
changes
that
you
do,
I
can't
stay
up
to
the
latest
by
the
way
I
mean
really
great,
to
have
the
version
one
out
I
mean
a
lot
of
changes
that
you
guys
did
is
like
amazing.
I
have
been
able
to
bring
it
up.
B
Yeah
yeah.
G
I
mean
if
you
want
to
have
like
more
discussions
around
it.
I
mean
I,
I
don't
know
how.
How
can
I
help
you
guys,
but
anything
that
you
want.
I
mean
I'm
really
perfectly
okay
to
spend
more
time
on
this
with
you
guys
or
are
everyone.
C
Okay,
yeah,
I
think
we
discussed
with
something
similar
with
this
hashes
right
here.
Do
you
do
you
remember.
B
Or
yes,
you
mean
it's
about
global,
remote
or
just
like
in
general,
just.
C
B
The
logical
steps
that
leads
to,
but
we
can
go
through
them
again
like
in
okay
out.
A
Of
this
meeting,
so
just
something
here
so
deep:
okay,
can
you
can
you
prepare
a
couple
of
slides
so
that
we
can
illustrate
your
problem
and
maybe
some
some
some
ideas
that
you
have
and-
and
we
can
do
a
brainstorm
next
time
like
we
can
dedicate
a
little
bit
more
time
because
we
are
like
almost
10
minutes
over
there.
G
B
Yeah,
just
throwing
it
out
how
other
platforms
do
change
data
capture
on
a
database
probably
overkill
for
the
solution.
But
if
there's
continued
problems
with
the
procedure,
storing.
G
By
the
way
I
mean
after
the
change
I
mean,
did
you
see
any
good
performance?
I
mean
performance
improved
or
something
like
that.
C
Yeah
yeah,
so
now
I
mean
for
this
change
we
for
the
events
for
now.
We
just
use
this
in
the
counters
that
are
visible
on
the
on
the
kuma
gui
right
on
the
dashboards,
because
we
didn't
have
time
to
re-implement
the
config
generation
right.
But
this
is
what
we
eventually
want
to
do
right
to
just
react
on
the
events
to
generate
the
android
configuration.
G
I
mean
I'm
already
over
time,
but
in
case
one
database
connects
to
like
multiple
cp
will
the
database
generate.
So
many
events
to
all
the
other
cps.
G
So
let's
say
you
have
multiple
remote
control
planes
and
all
of
them
are
connecting
to
the
same
database
and
and
there
is
change
in
the
configuration,
so
the
notification
will
be
sent
out
to
all
the
remote
cps
or
with
it
go
to
only
one.
C
No,
I
mean
I
we
we
don't
support
this.
Okay,
do
you
mean
remote
cps
as
many
zones
or
remote
cps
as
many
instances
without
with
within
one
zone,
yeah.
C
In
this
yeah,
so
we
yeah,
so
if
you
have
one
zone
and
many
instances
right
with
current
implementation,
every
instance
will
receive
the
notifications,
the
notification
from
the
database
right
and
if
you
have
many
zones
and
every
zone
should
use
their
own
schema.
The
database
schema.
Even
if
you
connect
this
to
one
database,
it
should
be
a
separate
schema.
I.