►
From YouTube: Layer5 Community Meeting (Feb 4th, 2022)
Description
Layer5 Community Meeting - February 4th, 2022
Agenda:
- Layer5 collaboration with Openforce
- Meshery Intro and deep-dive
- Meshery UI
- State Management In Meshery UI
- Implementation tracker
- Contribution guide
- Meshery CLI
- Meshery Documentation
- Service Mesh Catalog
- Meshmap beta sign-up
- Docker Desktop integration
Join the community at https://layer5.io/community
Find Layer5 on:
GitHub: https://github.com/layer5io
Twitter: https://twitter.com/layer5
LinkedIn: https://www.linkedin.com/company/layer5
Docker Hub: https://hub.docker.com/u/layer5/
A
A
Basically,
this
name
is
openco
and
I'm
most
on
the
machine
learning
side,
I'm
like
I'm
previously,
I'm
a
gsoc
student
under
tensorflow,
so
I
am,
I
contributed
in
tensorflow.js,
community
and
yeah
here
I
am
like
I'm
open
source
guys
just
want
to
explore
more
on
more
on
machinery
and
life
community.
So
thank
you.
C
Yeah
hi,
hello,
everyone,
I'm
aksharati,
so
I'm
an
engineering
student
from
hbo
kanpur
india.
So
this
is
my
first
time
like
exploring
open
source,
but
I
have
prior
experiences
with
internships
majorly
in
front
end,
where
I
work
with
react.
Redux
and
material
ui,
so
I'll
be
looking
forward
to
contributing.
B
B
B
E
Oh
okay,
so
yeah,
I'm
ruthless,
I'm
also
from
india
and
I
joined
the
community
last
week.
I
just
found
I
found
out
about
mushri
through
just
github,
like
someone's
github
profile,
and
I
I
learned
that
it's
a
part
of
cncfs
and
I
was
interested
in
exploring
cloud
native
projects
so
yeah
this.
That
brought
me
here.
G
Yeah,
thank
you
hey
there,
so
my
name
is
juliano
frantosi,
I'm
from
brazil.
I
was
just
recently
introduced
to
the
project,
so
you
know
I'm
I'm
still
a
very
early
stage
of
exploring
here,
but
I've
I've
really
enjoyed
the
community
support,
so
I'm
actually
quite
excited
to
get
to
know
about
everything.
I
heard
it.
G
F
F
G
H
So
hey
hi
hi,
guys,
I'm
yogi
and
I
I've
been
part
of
layer
five,
but
I've
been
very
on
and
off.
So
I
just
wanted
to
introduce
myself
to
my
name
is
yogi
porla
and
I
work
for
work
at
hpe,
and
this
has
been
a
I've
heard
great
things
about
this
community,
including
team
members
like
juliana,
who
it's
a
really
good,
has
great
deal
of
things
great
things
to
say
about
this
community
and
how
helpful
the
community
is.
So
thank
you.
F
F
I
K
L
H
L
N
I've
not
really
made
any
contributions
yet,
but
I'm
really
looking
forward
to
working
and
growing
in
the
community
most
most,
especially
in
terms
of
writing,
because
I've
seen
that
the
the
the
mercury
code
is
not
what
I'm
familiar
with.
So
I
think
I'll
just
stick
to
on
the
aspect
of
writing
and
much
much
more,
especially
the
like
documentation,
part
so,
and
I'm
happy
to
be
here.
Thank
you.
O
O
H
C
Hello
guys,
my
name
is
anjet
chandra,
I'm
from
india,
I'm
currently
a
third
year
student
in
information
technology
branch
and
I've
been
working
for
a
earlier
startup
name.
Amsos,
take
private,
limited
sd
intern,
like
I
always
wanted
to
contribute
in
open
source,
and
I
just
recently
came
to
know
about
the
layout.
I
like
this
community,
how
they
gave
people
opportunity
to
contribute
in
the
open
source
so
happy
to
have
you
all
here.
Thank
you.
Thank
you.
P
I'm
from
nigeria
I'm
a
front-end
developer
and
also
I'm
into
technical
writing
literature.
I
joined
the
newcomers
meeting
yesterday
and
I
actually
joined
this
community
called
today
and
I
really
hope
to
contribute
to
the
community
in
terms
of
documentation
and
technical
writing,
and
also
it
helped
me
improve
my
skills.
Thank
you.
F
Q
R
C
B
S
Cool
so
hello
good
evening
to
everyone,
so
I'm
proudly
announcing
that
we
have
collaborated
with
open
force
where
it
is
a
commenting
initiative
by
my
friend
haven't
got,
and
here
we
will
be
leading
for
different,
open
source
projects
and
we
will
be
guiding
students
how
to
contribute
to
their
health
and
yeah.
So
the
timeline
will
be
from
third
master
28th
march
and
yeah.
I
can
share
the
doc
link
with
you
just
so
we
don't
I'm
not
able
to
go
with
it.
F
S
So
this
is
a
link
for
the
doc
that
we
have
prepared
for
layer,
five
collaboration
with
open
force,
so
we
need
the
organizing
team
and
we
would
be
dividing
the
areas
of
responsibility
to
every
like
volunteer.
Who
would
be
able
to
volunteer
for
this
course
and
like
we
will
be
looking
for
people
who
would
be
making
certain
issues,
as
we
have
already
certain
issues,
we
can
work
on
it.
S
Otherwise
we
can
make
certain
new
issues
for
the
beginners
to
contribute
to
layer
5,
and
we
would
be
also
needing
people
for
promotions
on
linkedin,
instagram
and
twitter.
Obviously,
twitter
we
have
and
yeah.
So
if
anyone
is
entrusted
for
the
same,
they
could
write
their
names
in
the
like
below
in
document
today,
and
we
would
be
happy
to
have
you
on
board
for
this.
F
S
Also,
it
is
a
great
cause
since,
since
we
are
promoting
the
beginners
to
contribute
to
such
open
source
organizations
and
many
organizations
such
as
apache
eclipse,
adaptive
is
coming
for
the
collaboration
as
well.
So
it
will
be
a
great
start
for
layer.
Five,
too,
I
had
to
be
in
a
like,
in
collaboration
with
a
student
community
event,.
B
F
Yeah
sure
you
know
so
by
the
way,
there's
oh
there's
36
of
us
on
the
call
today,
I'd
like
to
say
that
that's
a
record.
It's
not
so
tell
your
friends
because
we're
close
not
quite
there.
Yet,
though
speaking
of
telling
your
friends
and
your
colleagues
mr
porla
who's
on
the
call,
is
well
instrumental
in
getting
max
and
giuliano
and
and
the
rest
of
the
crew
here
and
and
just
with
the
influx
of
a
number
of
newcomers.
F
So
if
you
don't
interrupt
with
the
question
I'll,
just
assume
you're
not
paying
attention
or
you
don't
care,
I
guess
I'll
just
be
offended.
So
anyway,
point
is
just
oh
38.
Okay,
we're
up
we're
mesh
mate,
aditya
you'll
have
to
keep
score.
If
that's
like,
I
think
that's
you
have
to
keep
scoring
on
the
official
record.
F
Don't
recall
all
right
so
so
there's
there's
a
ton
to
talk
about
like
measuring,
is
about
almost
a
two-year-old
project
at
this
point
and
actually
no
actually,
it
is
so
mesherie's
been
boy,
a
lot
to
say,
measures
been
open
from
the
very
beginning.
It
looked
totally
different
than
it
does
today.
It
wasn't
a
cncf
project
until
a
few
months
ago
it
started
out
to
be
well.
F
It
started
out
running
on
top
of
play
with
kates.com
playwithdocker.com
we're
teaching
people
about
service
meshes
people
kept
asking
questions
like
which
one
should
I
use
and
what's
the
difference
between
them?
What's
the
overhead
involved
there's
a
lot
of
value
here?
How
do
you
characterize
the
performance
of
these
things?
How
do
you
know
if
you're
running
a
service
mesh
well
gosh?
I'm
scared
to
make
it
change.
F
How
do
I
know
if
the
traffic
you
know
the
configuration
that
I'm
applying
is
doing
what
it's
supposed
to
do
and
all
those
types
of
questions
right,
understanding
that
there's
a
lot
of
power
inside
the
network
and
that
we've
got
this
newfangled
serve?
You
know
software-defined
networking
layer
called
service
meshes
and
that
there
are
many
of
them
and
that
there
are
they're
here
to
stay,
and
many
of
them
are
here
to
say
some
of
them
will
die,
that's
unfortunate
and
some
of
them
and
new
ones
will
be
announced.
F
I
suspect
one
of
you,
one
or
more
of
you-
that's
been
paying
attention
to
an
upcoming
announcement
about
a
new
service
mesh,
we'll
probably
list
off
that
service
mesh
in
the
zoom
chat.
So
we'll
send
you,
you
know
you'll
get
the
today's
prize,
which
is:
is
it
the
pot
up
to
half
a
million
dollars?
That's
what
we're
gonna
send
whoever
says
it
first.
I
think
that's
what
it
is.
F
So
brief.
Super
brief
recap:
what's
a
service
mesh,
oh
architecturally,
high
level,
a
data
plane
and
a
control
plane,
lots
of
proxies
with
something
controlling
them
and
configuring
those
proxies
good.
If
you
don't
have
a
control
plane,
you
ain't
a
service
mesh.
I
was
speaking
with
the
founder
of
well
a
company
that'll,
let
go
unnamed,
but
he
was
pedantically
stating
that
if
you
don't
have
observability
you
ain't
a
service
mesh.
As
a
matter
of
fact,
he
was
saying
that,
in
reference
to
the
mesh
that's
about
to
be
announced
anyway,
networking
is
difficult.
F
F
You
might
want
to
manage
that.
Can
you
optimize
your
service
mesh
configuration
yeah,
I'm
sure
you
can
help
you
assess
or
a
management
plan
can
help.
You
assess
that
mastery
can
help.
You
assess
that.
Is
that
a
one-time
assessment?
No,
it's
a
continual
thing
anytime
that
any
aspect
of
your
infrastructure
changes.
You
deploy
a
new
application.
You
reconfigure
your
your
mesh.
You
reconfigure
your
kubernetes
cluster.
You
deploy
a
new
version
of
your
app.
F
F
I
think,
in
a
boring
way
kind
of
focused
on
horizontal
support.
Like
doing
things
horizontally
like
supporting
we're,
what
are
we
at
10
service
meshes
right
now,
we'll
see
if
maybe
next
week
we
can
get
some
of
the
contributors
to
demo
the
latest
beta
adapter,
so
maybe
we'll
be
at
11,
although
it's
time
to
deprecate
one
so
we'll
do
a
plus
one
minus
one,
we'll
stay,
probably
at
10.
F
anyway.
The
project's
been
really
focused
on
how
to
provide
this
functionality
uniformly
across
service
meshes.
As
such,
the
project
and
its
vision
are
well
aligned
with
a
specification
called
service
mesh
interface,
which
is
which
we'll
describe
another
time
but
mesherie
as
it's
being
built
out
as
a
management
plane.
This
is
something
of
a
logical
architectural
view,
view
of
its
componentry,
and
it
is
you,
don't
have
to
think
of
meshi
as
being
this
complex.
If
you
don't
want
to.
F
F
That
those
are
pretty
intentional
and
they're
sort
of
littered
throughout
the
project,
so
someone
didn't
clean
up
their
code,
actually,
quite
the
opposite,
there's
very
intentional
consideration
given
for
making
measuring
pluggable.
So
a
prominent
example
is
something
like
an
adapter.
We
were
just
saying:
there's
10,
you
know
10
different
adapters
for
one
for
each
different
type
of
service
mesh.
If
you
want
to
add
another
one,
you
can
plug
into
that
extension
point.
There's
a
grpc
interface
you
plug
in,
and
the
same
goes
for
a
number
of
the
other
areas.
F
We're
going
to
see
some
extension
points
in
action
today.
Some
of
them
are
just
going
to
you
know
actually
all
of
them.
The
point
of
the
plug-in
model
is
that
you
can't
really
tell
that
they're
extensions,
so
measuring
from
a
functional
perspective,
kind
of
the
big
feature
categories,
if
you
will
is
like
life
cycle
management
of
the
individual
service,
meshes
there's
much
room
for
more
sophistication
there
and
there's
a
lot
of
sophistication
coming.
F
Well,
that's
coming
in
the
release,
that's
about
to
be
made
for
dot
six,
but
even
more
4.7,
and
how
it
is
that
you
can
in
a
sophisticated
way
manage
the
ongoing
configuration
of
your
service
meshes
a
lot
of
times.
You'll
hear
people
refer
to
that
type
of
configuration
as
a
pattern-based
approach,
so
we'll
be
clarifying
what
service
mesh
patterns
are,
so
you
should
be
able
to
do
workload
management.
Why
are
you
running
the
infrastructure
while
you're
running
the
service
mesh
in
the
first
place?
F
F
So
smi
is
a
collection
of
four
specifications
that
define
how
you
can
interact
with
any
given
mesh
that
implements
these
specs
smi
is
another,
is
a
sibling
project.
Meshri
is
the
conformance
tool.
So
if
you
want
to
know
if
the
service
meshes,
those
six
or
seven
service
meshes
that
have
implemented
smi
are
running
according
to
spec.
Meshery
is
the
tool
to
use
to
to
do
that?
There's
a
public
facing
dashboard
of
what
that
is
of
their
current
status.
F
Measuring
also
recognizes
that
web
assembly
is
a
thing
and
there's
a
lot
of
power
that
can
be
well
exposed
and
controlled
through
filters.
So
there
are
some
just
as
in
the
current
releases
that
are
coming
out
now.
There's
can
out
of
the
box
filters
that
that
ship
with
meshri
meshri,
then
that's
kind
of
an
example
of
an
extension
point.
F
It's
just
like
so
long
as
you
can
avoid
a
lot
of
overhead
and
a
lot
of
performance
impact.
There's
just
a
lot
to
be
explored
with
what
people
can
do
with
those
types
of
filters,
and
hence
that's
why
we
focus
on
performance.
We'll
talk
about
performance
in
a
minute,
because
the
more
that
you
do
with
your
infrastructure,
the
more
you
would
expect
it.
As
has
overhead
that's
incurred,
it's
interesting,
because
it's
not
just
about
measuring
overhead
and
speeds
and
feeds
it's
also
about
measuring
value
and
your
benefit.
F
F
So
there's
that's
a
whole
topic
onto
its
own,
but
the
reason
I
mention
it
is
because
this
way
this
unit
of
measure,
a
mesh
mark,
would
indicate
to
you
how
well
your
how
efficiently
your
infrastructure
is
running
your
service
mesh,
your
your
cluster,
but
also
incorporate
it
into
that
unit
of
measure,
would
be
a
consideration
for
the
benefit,
the
value
that
you're
you're
gaining
as
well.
So
so
in
a
similar
vein,
by
the
way
the
do
interrupt
as
we
go,
I
can,
and
so
in
a
similar
vein.
F
F
Mastery
acknowledge
we
try
to
acknowledge
in
the
like,
in
the
community
and
in
the
project
that
there's
any
number
of
functionalities.
That
service
meshes
offer
circuit,
breaking
as
an
example
or
retries,
or
what
have
you
and
people
aren't
quite
sure
if
they're
doing
those
things
correctly?
Is
your
circuit
breakers
too
sensitive?
Is
it
not
sensitive
enough?
Does
that
sensitivity
need
to
change
over
time
based
on
runtime?
You
know
telemetry
runtime
things
that
are
happening.
F
Everybody
like,
like
almost
everyone,
has
that
type
of
a
question.
So
what
are
the
common
patterns?
What
are
the
templates
by
that?
We
can
use
to
make
sure
that
we're
operating
in
accordance
with
how
others
are
doing
it
or,
as
we
take
a
given
pattern
and
apply
that
to
our
infrastructure,
about
how
to
run
a
given
how
to
configure
your
environment,
you
might
you
might
want
to
change
that
and
modify
that
over
time,
and
so
you
can
persist
those
configurations
so
we'll
get
into
that.
F
Actually,
the
so
so
there's
a
a
collection
of
patterns
that
are
coming
forth,
some
that
are
being
written
into
an
a
book,
that's
an
early
release.
So
if
you're,
an
o'reilly
subscriber,
you
can
go
check
it
out
if
you're,
not
an
o'reilly
subscriber
like
come,
contribute
to
the
book.
There's
a
lot
of
patterns
to
be
published,
there's
in
concept
a
couple
of
books
on
service
mesh
patterns,
there's
a
few
that
have
been
published
already
or,
like
tentatively
published.
F
F
F
Allow
you
to
compare
different
performance
differences
between
different
configurations,
that
type
of
functionality
is
works,
whether
you're
testing
services
off
kubernetes
on
kubernetes
on
a
service
mesh
or
office
service
mesh.
That's
in
part,
one
of
our
other
sibling
projects
within
the
cncf
is
a
project
called
service,
mesh
performance,
and
so
smp.
Spec.Io
is
where
you
can
learn
more
about
that
particular
project.
F
Nice,
so
the
question
is,
you
know,
so.
Patterns
are
basically
ways
of
how
we
combine
different
service
meshes
okay,
so
our
pattern
is
basically
ways
of
how
we
combine
different
service
meshes,
or
is
this
more
about
the
workflow
and
specific
to
one
service
mesh
yeah?
That
makes
that
makes
sense?
Well,
here's
the
unfortunate
confusing
answer
is
it's
both,
which
is
to
say.
T
Yeah
can
I
can,
I
add
something
sure
thanks.
So
maybe
lee
can
correct
me
something's
wrong
with
what
I'm
about
to
say,
but
I
like
to
see
patterns
like
a
way
to
declaratively,
define
and
a
well
basically
a
way
of
architecting,
a
specific
workload
and
the
reason,
because
sometimes
it
could
be
a
mesh
agnostic,
and
sometimes
it's
not
it's
basically,
because
it
really
depends
on
what
you're
trying
to
achieve.
It
could
be
something
that
the
different
layers
of
of
the
the
stack
right.
F
The
yaml
here
is
an
example
like
so
so
we're
doing
ourselves
a
disservice
and
we
have
to
stop
this
after
the
dot
six
release,
and
that
is
that
that
we've
conflated
two
things
within
the
project
kind
of
the
difference
between
a
pattern
like
a
template
for
like
hey
here's,
a
here's,
a
here's,
how
you
do
retries
well
and
here's
how
you
might
configure
it,
either
on
a
specific
mesh
or
across
different
different
surface
meshes,
and
that
you
could
go
to
kind
of
a
catalog
of
patterns
and
say
yeah.
The
circuit
breaking
pattern:
yeah!
F
F
That's
the
gist
of
how
we
will
eventually
be
using
the
term
patterns,
there's
a
different
term
that
we
should
bring
forth.
I
think-
and
you
know
like
I'm,
I'm
expressing
this
on
behalf
of
some
of
the
other
maintainers,
but
that
like
in
this
ui,
we
would
probably
this
would
end
up
being
called
like
designs
or
it
would
be
called
just
life
cycle
management
of
the
the
mesh.
F
So
this
would
be
your
service
mesh
configurator,
the
the
same
so
as
you
go
through
and
kind
of
design
how
you
want
a
specific
service
mesh
deployment
to
run
that
gets
it
gets
expressed
in
this
in
the
yaml
format.
That
is
it's.
That
is
the
same
format
that
we
express
patterns
within.
So
if
there's
a
best
practice
kind
of
approach
to
how
you
do
circuit
breaking
that,
and
that's
a
pattern
that
pattern
is
expressed
in
yaml,
that
that
same
yaml
format
is
this
same
format,
so
you
can
so
you
could.
F
Think
of
so.
I
think
the
thing
that
we
want
to
try
to
begin
to
like
convey
better
and
grasp
yeah-
and
I'm
saying
this
like
toward
myself
as
well-
is
that
like
patterns
are
that
that
you
master,
you
can
use
meshrey
to
go.
Configure
your
service
mesh
if
you've
identified
and
you
can
take
certain
designs-
have
been
done
if
those
are
kind
of
tried
and
true
that
those
can
be
qualified
or
sanctified
as
a
pattern
sort
of
in
in
your
in
the
official.
F
The
official
way
that,
like
hey
these
are
generally
reusable.
So,
if
you
think
about
docker
hub
and
the
fact
it
has,
you
know
hundreds
of
thousands
of
container
images,
some
of
those
are
official.
Some
of
those
are
verified
publisher,
and
so
some
of
those
are
well
supported
and
highly.
You
know
highly
used.
Others
of
those
are
published
by
you
know
joe
schmoe,
and,
and
so
I
think
we
we
consider
that
that
service
mesh
catalog
is
essentially
becomes
the
same
thing
that
there
are
some
of
those
designs
that
are
imbued
with
best
practices.
F
We
would
refer
to
those
as
patterns
and
that
as
a
community
or
as
a
project,
we
would
stand
behind
and
say
you.
You
know
this
is
the
way
that
generally
agreed
upon
way.
You
should
probably
go
and
run
this
type
of
behavior
within
your
mesh
and
there
may
be
any
number
of
other
individuals
that
come
along
and
create
a
design.
Maybe
it's
a
fantastic
pattern.
F
They
too
can
take.
You
know
it's
the
same
descriptor,
it's
the
same
format,
so
they
too
could
take
and
publish
that,
and
so
we've
as
a
community
we've
yet
to
like
arrive,
have
the
discussions
that
kind
of
arrive
at
like
well.
What
would
be
an
official
pattern
and
and
well
supported
and
incorporated
into
the
integration
tests
and
that
kind
of
a
thing
versus
what
are?
How
do
we
delineate
between
any
other
design
that
someone
else
might
create
so.
F
F
The
load
generators
that
we've
spoken
of
they're
kind
of
bundled
in
there
as
well.
There's
individual
containers,
one
one
for
each
mesh
readapter
to
the
extent
that
you
connect
measuring
to
kubernetes
you'll,
find
that
there's
an
operator
that's
deployed.
The
operator
has
a
couple
of
custom
controllers,
the
most
notable
of
which
is
a
controller
called
mesh
sync
and
so
running
inside
of
kubernetes,
whether
meshery
is
running
in
kubernetes
or
whether
you've
deployed
mesherie
on
a
docker
host
either
of
those
two
deployment
models
are
supported.
F
When
you
do
connect-
or
if
you
do
connect
meshri
to
kubernetes,
it
will
deploy
an
operator
so
that
mesh
we
can
stay
in
sync
with
in
real
time.
So
both
it
can
stay
in
sync,
with
real-time
configuration
changes
that
are
happening
within
kubernetes
and
within
your
service
mesh,
also
so
that
it
can
perform
discovery
of
any
existing
service
meshes
and
any
existing
workloads
that
you
might
have
deployed.
F
F
If
you
choose
the
local
provider,
you
are
having
a
single
user
mode
experience
or
sort
of
a
no
user
mode.
Experience
perfect.
If
you
choose
to
log
into
a
provider,
different
providers
might
offer
different
providers
can
give
you
might
offer
up
plugins
different
providers
will
yeah
have
different
have
geez.
How
do
I
say
this
like
have
have
different
characteristics
like
some
of
them
might
help
persist
performance
tests
that
you've
run,
so
you
can
pull
up
the
history
of
those
tests.
Some
of
them
might
be
for
research
purposes.
F
F
R
Okay,
so
as
we,
everyone
is
not
like
a
ui
person,
some
people
wants
to
interact
with
terminal.
We
have
a
cli
of
our
own.
That's
message:
cpl
I'll,
just
present
my
screen
to
showcase
some
of
the
features.
Okay,
so
mystery
ctl
is
our
cli
and
we
have
a
lot
of
features
here
like
app.
We
can
manage
our
messaging
applications.
R
R
F
K
F
Yeah,
so
so
to
usher's
point.
Not
every
use
case
involves
the
mouse
and,
and
so
mescheri
the
mastery
cli
is
not
only
it's
designed
for
humans.
Primarily
push
can
tell
you
all
about
the
pain
involved
in
designing
for
humans,
but
it's
also
understood
that
it's
pretty
common
for
people
to
kind
of
treat
a
cli
in
a
programmatic
way,
as
well
as
an
unofficial
api
anyway,
because
of
some
of
those
use
cases,
people
want
to
run
performance
tests.
That
measure
will
will
do
within
their
pipeline.
F
People
want
to
run
maybe
conformance
tests
from
within
their
pipeline.
People
might
want
to
have
meshri,
apply
a
pattern
or
like
deploy
a
service
mesh,
deploy
their
application
apply
a
pattern
run
a
performance
test
like
do
all
that
kind
of
a
thing
within
their
build
system
within
their
pipeline,
and
so
to
the
extent
that
projects
are
using
github
as
their
workflow
is
their
pipeline
as
their
their
build
engine.
F
This
community
is
published,
I
think,
in
total
about
six
github
actions,
but
like
three
of
which
are
probably
pretty
well
user-facing,
so
there's
one
for
performance
management
or
one
for
conformance
tests
for
smi
and
hirsch.
If
he's
on
the
call
calling
he's
working
on
one
around
patterns
so
to
apply
a
pattern.
M
A
A
So
if
you
see
what
it
and
another
thing
that
you
could
possibly
do,
is
you
know
we
have
these
actions
for
someone
to
you
know
they.
They
just
want
to.
Let's
say
that
they've
done
something
new
and
they
just
want
to
test
out
some
patterns
and
stuff,
and
they
it's
it's,
not
a
very
massive
pattern.
So
what
they
could
do
is
they
could
just
simply
spin
up
a
bit
of
action
to
see
whether
you
know
the
they're
applying
something
and
that
that's
working
or
not
and
specifically,
for
measuring
ctl.
A
So
if
you're
doing
some
changes
to
this
particular,
let's
say
if
you're
making
some
changes
with
regards
to
the
command
line
right
and
you
want
to
see
whether
this
is
working
or
this
is
not
working.
So
this
is
a
particularly
good
action
that
you
could
go
ahead
and
spin
up
and
see
whether
you've,
absolutely
whether
you've,
you
know
not
made
a
mess
there
and
see,
okay,
that
okay.
This
is
something
that
that
serves
as
a
sort
of
a
visual
test
as
well.
A
So
this
is
something
that
that
you
could
do
it's
pretty
this.
This
one
is
pretty
straightforward.
There
is
another
one
in
the
chat
which
has
been
shared
by
navendu.
I
guess
yeah,
that
is
the
service
mesh
performance
action.
So
if
you
just
want
to
go
ahead
and
check
the
performance
of
your,
you
know,
you
could
fork
from
that
repository.
A
F
K
F
F
So
yeah,
one
of
the
one
of
the
features
of
mastery
we're
talking
about
is
extensibility.
The
ability
for
remote
providers
to
create
plugins,
one
of
the
plugins
that's
being
worked
on
in
the
community,
is
referred
to
as
a
mesh
map,
and
so
meshmap
has
is
about
to
go
into
beta,
and
so
I'm
sure
someone
will
drop
a
link
to
where
to
go.
Sign
up
for
beta
to
introduce
it
briefly.
It
is
mesh
map
is,
is
really
has
two
modes.
F
And
understand
you're
currently
running
infrastructure.
So
if
in
your,
if
I
think
what
did
I
have
istio
and
linker
d
both
deployed
in
this
environment,
but
it
lets
you
analyze,
you're,
currently
running
infrastructure
go
through,
you
know,
get
details
you
want.
You
want
to
look
at
run,
invoke
performance
tests
against
your
services
that
are
running.
F
F
O
F
F
There's
kind
of
your
stereotypical
set
of
ways
that
you
want
to
view
your
infrastructure
and
understand.
What's
going
on
in
your
runtime
environment,
try
this
real
quick!
So
you
can
search
through
and
identify.
You
know
pick
out
the
ones
you
want
to
work
with
there's
another
mode
inside
of
mesherie
or
sorry.
I'm
sorry
inside
of
mesh
map
referred
to
as
a
designer
mode
and
so
in
designer
mode.
F
F
Yeah,
the
the
visualizer
mode
is
probably
keali
is,
is
most
similar,
yeah
yeah,
except
keali.
You
know
so
so
great
you
so
kelly's.
You
know,
I
don't
know
we
don't.
I
don't
spend
a
lot
of
time
with
it,
but
but
it
you
know,
is
istio
only
and
does
like
a
fantastic
job
of
hooking
up
with
prometheus
and
and
speaking
to
istio,
specifically
and
helping
you
understand,
what's
happening
in
your
environment
and
that's
kind
of
yeah,
it's
very
similar
to
the
visualizer
mode,
except
for
measuring
it's
really
any
service
mesh.
F
You
can
like
this
you'll
have
to
correct
me.
I
don't
recall
kind
of
where
this
is
at,
but
if,
like
for
different
style
configurations
that
you're
going
to
do,
there's
different,
you
know
there's
different,
meaning
involved.
So
if
you
want
to
apply
like
an
envoy
filter,
you
might
take
that
component
and
and
instead
of
being
able
to
place
it
onto
the
canvas
directly.
It's
not
gonna,
it
won't.
Let
you
it!
It
wants
you
to
apply
that
to
a
given
service.
F
That's
there
a
given
component,
that's
already
there
and
then
you
can
configure
that
you
know
the
binary
for
the
filter
that
you
want
to
run
and
any
you
know
text-based
configuration
you
want
to
apply.
You
know,
there's
the
concept
that
you
might
want
to
apply.
You
know
multiple
filters
in
a
chain
for
a
specific
service.
You
might
drop
over
a
different
one
and
configure
what
that
filter
is,
and
then
you'd
have
kind
of
your
inventory
of
envoy
filters
here.
So
we
really,
you
know.
V
U
F
F
It's
just
it's
layer,
five
io,
slash
mesh
map
and
so
so
jump
in
yeah
get
get
on
the
waiting
list
now,
but
fill
in
the
form
and
you'll
be
on
the
list
to
get
early
access
to
go,
go
break
something.
There's
there's
some
additional
functionality
that
we
really
didn't.
I
didn't
really
demo
in
there
there's,
there's
a
ton
more
work
to
be
done
within
there,
but
yeah
it's
it's,
but
that's
kind
of,
but
you're
sort
of
understanding
the
gist
of.
What's
going
on.
F
Like
I
think
in
the
future,
you
know
it
would
be
ideal
if
there's
a
bit
of
less,
maybe
segregation
between
the
two
modes.
You
know
we
kind
of
switched
from
the
visualizer,
where
you're
seeing
you're,
currently
running
infrastructure
and
then
the
designer
where
you
would
apply
patterns
and
create
your
own
designs.
But
we
want
to
make
that
a
little
more
integrated
such
that
like,
while
it's
not
necessarily
a
best
practice
for
people
to.
We
don't
necessarily
want
to
encourage
people
to.
F
Now,
what's
the
word
I'm
looking
for
to
to
not
to
make
ad-hoc
changes
to
their
production
infrastructure?
Sometimes
that's
necessary.
You
need
to
go
debug
things
and,
and
great,
hopefully,
mesh
map
helps
in
that
way,
but
also
we
want
to
like
you,
you're
you're,
able
to
connect
up
to
github,
github
and
read
in
patterns
and
not
yet,
but
ultimately
push
pattern
like
there
needs
to
be
a
get
ops
centric
flow
to
this
kind
of
an
approval
process
by
which
people
can
run
these,
I
don't
know
we're.
F
I
was
to
be
meeting
with
docker
cto
right
now.
We
just
had
a
conflict.
I
want
to
make
sure
we
were
here
for
the
for
the
new
folks
that
are
taking
a
look,
but
we
want
to
talk
to
them
about
an
integration
of
meshri
and
docker
desktop
docker
desktop
today,
fantastic
way
to
get
you
know
a
local
kubernetes
cluster,
so
that
if
you're
a
developer
or
an
engineer
that
needs
to
have
a
local
system,
it's
a
great
it's
a
checkbox
thing.
H
H
So
once
we
have
the
service
identity
in
there,
we
could
actually
using
this
as
editor
for
abac.
That
would
be
amazing,
because
that's
what
many
struggle
with
we
at
hp
we
are
struggling
with
that
on.
How
do
we
actually,
because
with
our
backs?
Yes,
you
do
have
the
user
involved,
but
with
for
the
a
backs,
we
do
need
this
service
identity.
Once
we
have
the
service
identity,
how
do
we
actually
route
traffic?
How
do
we
actually
issue
the
either
it
could
be
leveraging
oppa
or
leveraging
something
like
kyber?
Now?
H
F
It's
a
great
great
suggestion,
yeah
and
to
be
able
to
do
it
visually
and
like
verify
that,
like
even
to
be
able
to
test
something
like
to
put
an
assertion
like
if,
if
the
token,
if
the
user
inside
of
this
jot
inside
this
token
is
equals,
yogi
borla
is
the
traffic
allow.
Is
it
redirected
over
to
the
golden
cluster
or
is
it
you
know
what
like,
but
to
be
able
to
see
that
and
or
rather
it's
not
just
yogi
portal,
but
in
part
to
your
point,
yogi
is
like,
but
also
is
yogi
hat.
T
Sorry
to
interrupt,
I
just
wanted
to
say
that
the
github
approach
that
lee
mentioned
it's
also
like
super
valuable,
because
it's
not
besides
having
that
ability
to
design
visually
if
we're
able
to
back
it
up
and
have
these
get
ups
a
process.
You
know
it's
like
awesome
like
to
have
this
gate
integration
and
have
some
process
that
automatically
a
you
know.
It's
able
to
to
see
the
context
about
different
branches
and
make
sure
the
infrastructure
is
well
basically
in
accordance
with
that
desired
state.
That's
awesome.
F
F
So
obviously
that
needs
to
be
reviewed
critically
and
then
may
be
approved,
and
if
it
is
like
it
and
it's
built
and
then
deployed
in
the
runtime,
it
could
be
really
nice
to
see
when
that
goes
to
crap,
that
you
could
do
a
git
blame
and
see
that
it
was.
You
know
lee
that
made
that
change
and
then
make
sure
he's
on
the
call
to
go,
fix
his
problem
and
like
roll.
F
We
should
that's.
We
should
have
absolutely
like
double
click
on
on
that.
We
should
so
we're
kind
of
out
of
time
today,
but
we
should
get
a
a
design
spec
going.
We
should
get
some
thoughts
on
paper
and
talk
about
some
of
the
use
cases
and
and
get
get
messy
integrated
with
spire
with
spiffy.
Much
more
deeply
like,
I
think,
yogi
you
were
presenting
sort
of
just
the
tip
of
the
iceberg
on
all
of
the
things
to
do
around
identity,
abac,
yeah,.
H
So
jared
I
don't
know
if
you're
familiar
with
frederic
cards,
so
we
were
working
with
with
them.
So
frederick
is
he's
the
design
head
at
anthem
and
he's
also
part
of
the
cncf
community.
He
had,
I
believe
he
had
one
of
the
tag
groups
within
the
cncf
community,
so
one
of
the
ideas
was:
was
we?
Okay?
Previously
we
were
in
a
passive
mode,
mainly
because
there
was
no
federal
mandate.
H
There
was
no
timeline
for
that.
But
now
there
is
a
timeline
which
is
2024
is
the
hard
deadline
we
have.
We
have
to
be
actively
present
in
there,
so
I
think
identity
as
a
spec
within
smi
that
could
work
wonders
so
yeah.
Okay,
if
you,
if
you
don't
mind,
could
you
we
could
think
of
a
sync
up
offline
and
if
you
can
share
some
details
on
what
you
have
in
mind
and
what
I
could
also
do
the
same
thing.