►
From YouTube: Kuma Community Call - December 23, 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
Yeah
I
mean,
as
as
marco
mentioned,
it's
it's
like
christmas,
so
let's,
let's
try
and
quickly
keep
this
short.
So
in
terms
of
events,
I
have
cleaned
up
everything
that
we
had
so
far.
A
A
So
for
the
future
the
call
for
papers
go
for.
Presentations
for
kubecon
eu
has
closed
will
be
the
notifications
will
be
sent
on
march.
I
guess
that
probably
a
week
later,
like
mid-march
will
have
the
schedule.
If
someone
is
interested
in
this,
we
we
released
like
24
hours
ago,
maybe
20
hours
ago,
we
released
the
last.
We
did
the
last
release
for
this
year.
A
The
change
lock
is
updated,
so
with
all
the
nice
stuff
that
we
added
a
quite
good
representation
of
how
to
say
external
contributors.
So
that's
that's
a
good
good
thing.
Other
than
that.
I
have
outlined
two
two
new
interesting
features.
A
How
we
can
how
we
can
use
it
on
universal,
it's
like
still
slightly
more
complicated
than
one
would
would
call,
but
the
world
out
there
is
complex.
So
this
is
what
we
have
that's,
why
people
use
kubernetes,
because
you
know
make
life
easier
and
then
we
have
obviously
our
very
first
big
contribution
from
our
brand
new
member
bart.
A
It's
a
new
policy.
Configuring
crit
rise.
It's
fully
documented
here
with
examples,
and
so
I
just
heard
that
bart
and
jacob
tried
this
cross
zones
cross
cluster.
So.
A
This
is
this
is
something
nice
new
that
we
we
added
as
a
last
touch
last
couple
of
touches
before
we
we
closed
the
year
overall,
we
were
just
chatting
like
half
an
hour
ago
with
marco
and
others
that
that
has
been
a
great
year
for
for
kuma,
like
starting
in
january,
when,
like
january
2020,
I
haven't
been
on
the
team
yet,
but
I
mean
it's
been
just
a
couple
of
months
old
project.
A
At
the
end
of
the
year,
it's
a
cncf
sandbox
project,
a
little
like
external
contributors,
lots
of
people
on
the
team,
lots
of
features
you
have
one
dot,
release
and
yeah.
We
have
many
great
plans
for
the
next
year.
A
I
don't
know
what
we
do
is
there's
anything
specific
that
that
we
cannot
hear
for
sure
in
this
104
release.
For
example,
we
have
added
the
the
great
work
of
jacob
for
preparing
for
migrating
to
xds
v3,
which
essentially
will
open
the
door
for
us
for
the
next
envoy
upgrades,
because
invoice
is
going
to
drop
the
v2
support
any
day
day
now
as
they
are
promising.
A
So
that's
probably
not
something
very
exciting,
but
this
kind
of
paves
the
way
they
pave
the
path
for
the
you
know,
smoother
upgrades
and
staying
on
the.
A
Like
with
the
with
the
novelties
that
the
the
the
data
plane
is
going
to
introduce
in
the
next
year,
I
don't
know
we
were,
we
were
discussing
some
things
about
innovating,
how
we
use
ip
tables
and
etc,
etc.
I
don't
know
what
what
of
this
can
be
can
be
discussed
here,
but
essentially.
B
I'd
like
to
I
would
like
to
chime
in
on
that.
I
I
think
I
think
that
there
is
lots
of
innovation.
That's
happening
with
bpf
ebpf
these
days
in
the
in
the
linux
world
and
would
be
would
be
amazing
to
explore
a
native
support
for
kuma
on
ebpf
in
such
a
way
that
perhaps
number
one
we
can
get
rid
of
the
ip
tables
rules.
Therefore,
we
can
make
it
faster
to
implement
transparent
proxy
across
the
board
and-
and
perhaps
you
know,
ebpf
is
quite
powerful.
B
B
A
A
B
I
would
say
that
there
is
you
know:
50
percent
of
the
project
really
is
about
innovating
and
you
know
pushing
what
a
service
mesh
can
do
for
the
end
user,
but
the
other
50
really
is
listening
even
more
than
50
listening
to
the
existing
users
and
understanding,
you
know
what
they
would
like
to
see
from
the
project,
and
I
think
that
most
of
it,
you
know
a
lot
of
it
if,
if
anything
will
involve
day
two
operations
of
kuma,
so
you
know
making
it
easier
to
understand
how
to
upgrade
the
control
planes,
how
to
upgrade
the
the
data
planes,
perhaps
instead
of
having
two
different
zones
standalone
and
multi-zone,
having
only
one
zone
that
can
then
gradually
organically
expand
as
multi-zone,
so
basically
simplifying
the
deployment
mode
as
well
as
perhaps
even
removing
the
database
dependency
on
the
remote
control
planes
and
treating
the
remote
control
planes
as
like
the
data
planes.
B
As
you
know,
you
know,
with
their
in-memory
storage
synchronized
with
a
global
control
plane
I
mean
there
is
a
lot,
a
lot
of
items
that
you
would
like
to
explore.
There
is
going
to
be
more
policies,
but
really
at
this
point
it's
also
about
setting
up
the
right
platform
in
the
community
to
understand
what
the
users
want,
and
I
think
we
will
have
to
do
some
work
around
that
when
it
comes
to
the
project,
we
are
not
very
organized.
I
think,
at
this
point
to
capture
this
feedback.
B
C
Yeah,
I
think
more
along
the
lines
of
the
second
point
that
you're
making
with
the
ease
of
upgrading
and
installation
and
just
understanding
the
deployment
flow.
I
think
also
the
actually,
I
know
one
of
the
people
who
put
in
a
request
for
more
policies,
the
the
circuit
breaker
extension
on
max.
So
I
think,
keeping
up
with
the
policies
on
between
kuma
and
like
the
other
service
meshes
is
also
something
that
I
would
like
to
help
out
on.
B
I
I
think
you
know
something
that
has
been
circulating
in
the
back
of
our
heads
was
to
make
the
policy
creation
a
little
bit
more
pluggable.
If
you
wish
right
now,
it's
a
little
bit
cumbersome,
whereas
we
would
like
to
make
it
easier
and
so
that
we
can
quickly
expand
to
new
policies,
new
modes
new,
whatever
new
filters,
you
know
at
any
time,
and
today
today
we
don't
really
have
this
kind
of
a
pluggable
dynamic.
B
You
know
abstraction
that
would
allow
us
to
do
that.
I
know
that
you
know
jacob.
This
is
something
that
me
and
you
spoke
about
a
long
time
ago,
but
maybe
you
know
we
can
create
a
doc
a
memo
and
see
if
somebody
from
the
community
can
also
help
with
that.
D
B
Yeah
yeah
I
mean
these
are
the
kind
of
projects
that
you
know
will
take
a
quarter.
Let's
say
to
you
know,
or
you
know,
they're
significant
right,
refactorings
of
the
project,
so
I
I
wouldn't
be
opposed
to
start.
You
know
prioritizing
some
of
this
work
which,
in
the
long
term,
will
accelerate
the
velocity
of
the
project.
B
Right,
yeah,
absolutely
so
austin.
Can
you
be
more
specific
on
day
two
operations?
What
would
you
like
to
see.
C
For
day
two
operations
for
upgrading-
and
things
like
that-
I
think
maybe
it's
as
simple
as
just
a
a
migration
guide
like
a
standardized
migration
guide
approach,
because
I
think
I
mean
it'll
get
easier
as
more
as
we
released
more
on
top
of
1.0.
But
I
think
the
1.0
release
was
difficult
for
us
and
having
to
rip
out
all
those.
A
C
Changes
but
yeah,
I
I
think
it's
just
as
the
project
grows.
I
don't
think
I
have
any
super
specific
things.
It's
working
well
enough!
So
far,
it's
just
about
the
the
adoption
and
kind
of
the
trying
to
keep
up
with
the
features.
B
I
think
that
something
that
we
can
do
better
when
it
comes
to
community
management
is
perhaps
even
you
know
sending
reminders
for
these
community
calls.
It's
very
easy
for
folks
to
forget
that
there
is
a
bi-weekly
community
call
and
usually
when
we
do
remind
ourselves
to
send
a
reminder.
B
C
One
question
I
have
is
in
regards
to
development:
how
much
happens
internally
on
the
like
the
kong
slack
versus
the
kuma
development,
because
trying
to
keep
up
with
I
mean
I,
I
don't
have
too
much
time
right
now.
So
it's
definitely
on
me,
but
trying
to
keep
up
with
all
the
new
features
like
even
this
retry,
this
new
retry
policy,
trying
to
figure
out
the
development
design
around.
That
did
that
happen
like
if
I
was
just
trying
to
observe
that
where
would
I
go.
B
Yeah,
so
the
retry
policies
were
a
feature
that
was
requested
by
a
con
customer,
that's
using
kuma,
and
so
we
we,
you
know,
we
built
it
because
number
one,
it's
something
that
in
general
we
need
it's
a
nice
feature
to
have,
but
second,
it
was
something
that
would
have
made
that
cuma
casa,
a
human
user,
successful
into
their
production
environment,
and
so
we've
been
prioritizing
it.
You
know
externally,
in
order
to
then
push
it
into
the
project.
E
Yeah-
and
I
can
maybe
one
more
thing-
it's
probably
because
it
was
my
first,
the
issue
most
of
the
development
was
done
on
the
zoom
calls
between
me
and
the
other
folks,
because
I
tried
to
understand
everything.
So
it
wasn't
perfect.
We
didn't
like
spend
enough
time
to
talking
about
it
on
the
community
slack,
but
I
will.
I
think
it
will
be
better.
C
Yeah
not
as
a
criticism
I
I
just
I
come
in
and
out
kind
of
I
make
promises
sometimes
and
I
can't
devote
enough
time
to
keep
them.
So
I'm
not
trying
to
say
change
your
workflow,
I'm
just
trying
to
understand
better
how
the
development
works.
A
This
is
actually
the
first
time
where
someone
is
interested
into
the
development
process,
not
not
just
requesting
features,
because
most
of
the
people
are
coming
in.
Like
I
have
this
in
that
situation,
I
want
to
solve
it
with
kumar.
Would
you
deploy?
Would
you
develop
this
for
me
so
that
I
can
pick
it
and
forget
about
yeah?
A
A
I
mean
we
tried
to
have
like
once
the
once
the
code
is
developed
and
there
is
a
pr
or
source
to
be
shared
with
like
if
you,
if,
if
you
can
track
there,
I
mean
we
try
to
put
you
know.
A
And
then
you
know
discussions
in
the
reviews,
so
you
can
see
things,
but
let's
say
that
the
the
whole
inception
of
the
thing
is
usually
happening
somewhere
like
outside
of
big
channels.
Yeah
I
mean
if,
if,
if
you
ask
me,
I
was
not
even
I
mean
I
knew
that
this
thing
is
coming,
but
I
don't
know
what
happened
between
and
jacob
when
they
were
when
they
were
designing
it.
So.
D
A
D
Yeah
also
also
one
more
thing,
I
think
what
we've
missed
with
this
policy
is
that
we
did
not
create
a
proposal
for
this.
So
back
in
the
days,
even
in
the
repo,
you
can
see
the
directory
docs
proposals
right
and
we
were
kind
of
pushing
even
short
pro
proposals,
how
we
want
to
implement
some
stuff,
especially
user
facing
policies,
for
example,
traffic
trades
right.
There
was
a
pro
pro
proposal
for
traffic
trace.
D
I
think
it
would
be
very
valuable
to
have
a
proposal
for
new
policies
and
especially
to
track
some
decisions
that
we've
made,
because
we're
essentially
translating
some
of
the
unvoiced
config
into
our
configs
right,
and
we
try
to
make
this
better
for
users
right,
so
we
sometimes
squeeze
something
into
one
configuration
or
conditionally
do
something
right.
It
would
be
great
to
describe
this
like
our
thought
process.
Yeah
of
the
of
the
new
policies.