►
From YouTube: Meshery Project Office Hours | KubeCon China 2021
Description
Meshery Project Office Hours | KubeCon China 2021
Join the community at https://layer5.io/community
Find Meshery on:
GitHub: https://github.com/meshery
Twitter: https://twitter.com/@mesheryio
LinkedIn: https://www.linkedin.com/showcase/meshery/
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
We
have
45
minutes
of
talking
about
my
favorite
project
mesherie
before
we
do
that,
let's
go
around
and
introduce
a
few
of
our
maintainers,
so
my
name
is
lee
calcote,
I'm
one
of
the
measuring
maintainers,
I'm
fortunate
to
be
joined
by
three
others.
Today
I
lose
count
but
we're
somewhere
north
of
10
maintainers
for
measuring
and
so
with.
No
further
ado.
A
Very
good,
very
good,
actually,
no
one's
counting
asuko
but
you're
the
longest
standing
maintainer
on
this
call
like
you've
been
around
the
project
from
I
don't
know
from
the
beginning.
Someone
just
asked
me
earlier:
when
did
the
project
start
and
it's
a
little
bit
deceiving
on
github,
because
it
looks
like
it
started
in
somewhere
in
2018
and
like
the
reality?
A
That's
a
technology
that
mesher
uses
around
linker
d
and
that's
certainly
a
technology
that
mescheri
uses
and
tries
to
help
empower
users
with,
and
then
I
won't
go
through
the
rest
of
the
list.
But
but
there's
a
list,
so
husseina
has
also
been
around
the
project
for
quite
some
time.
Hussein
do
you
want
to
say
hi.
A
D
A
But
we're
also
here
for
you,
which
is
to
say
please
interrupt
as
we
go.
Please
raise
your
ask.
Your
questions
choose
to
laugh
at
corny,
jokes
or
not,
but
certainly
ask
your
questions
and
then
that
way
you
can
guide
the
discussion
without
questions
we'll
probably
step
through
you
know.
All
of
what
kind
of
measure
is
we'll
sort
of
intro
it.
We'll
kind
of
deep
dive
into
some
areas.
A
A
A
There's
a
couple
of
claims
to
fame,
so
to
speak
about
the
project.
One
is
like
about
the
community
a
bit,
and
actually
it
just
randomly
happens
that
we
have
the
collection
of
maintainers
that
we
have
on
right
now.
What's
nice?
Is
that
there's
it's
there's
a
diverse
group
of
individuals
that
are
here
representing
from
independent
contributors,
users,
folks
that
work
at
large
tech
companies
and
folks
that
work
at
very
small
ones
and
so
like?
A
If
we
don't
say
it
enough
during
today's
presentation
jump
into
the
project,
it
is
whether
you're
using
the
project
want
to
contribute
to
it
or
just
be
a
fly
on
the
wall
like
just
hang
out
and
be
a
bystander
come
and
jump
in,
there's
a
lot
of
ways
to
engage
it.
Doesn't
it
isn't
necessarily?
It
doesn't
have
to
be
about
code
to
talk
about
measuring
intro
it
for
just
a
moment.
A
I'll
say
this
that
you
know
how
many
of
you
would
take
a
guess
as
to
how
many
service
meshes
there
are
out
there
like,
there's,
probably
more
than
you
know
of,
and
if
anyone
could
name
well
shoot
more
than
10
I'd
be
really
impressed.
If
you
could
name
over
20
I'll,
send
you
a
t-shirt
or
something
there's
more
than
20.
As
a
matter
of
fact,
we
like
next
year,
we're
gonna,
see
at
least
one
more
one,
more
significant
one,
at
least
so
it's
not
over
like
it's
gonna
get
messier.
A
That's
one
of
the
things
that
mescheri
attempts
to
do
like
part
of
the
genesis.
The
project
was
trying
to
help
people
embrace
diverse
infrastructure
and
be
comfortable
with
any
mesh
that
they
wanted
to
choose,
so
to
really
help
make
that
choice
to
facilitate
that
choice,
help
people
understand
how
service
meshes
work,
help
them
be
confident
and
comfortable
in
operating
them
and
to
give
them
choice
around
the
service
mesh
with
higher
level
tooling
beyond
that
sort
of
horizontal
functionality
about
supporting
what
are
we
at?
A
A
Some
of
those
adapters
get
used
more
than
others,
but
irrespective
meshri
tries
to
provide
a
fairly
uniform
experience
and
one
that
provides
well
provides
some
of
the
things
that
you
see
here
in
this
management
plan,
but
also
provide
things
like
well
configuration
best
practices
like.
Are
you
doing
it
right?
Are
you
configuring,
your
mesh
well
or
not,
in
accordance
with
best
practices,.
A
There
are
a
certain
number
of
best
practices
built
in
to
different
measuring
adapters
and
measure,
will
do
a
static
analysis
and
well
not
just
a
static
analysis,
but
a
runtime
analysis
of
your
environment
and
let
you
know
if
you're
doing
it
right
or
you're
doing
it
wrong,
because
there
are
any
number
of
service
meshes.
There's
a
couple
of
specifications
that
have
come
out
in
the
ecosystem.
A
A
A
Mastery
works
as
a
conformance
tool
in
that
regard,
which
is
you
know,
great
for
helping,
ensure
confidence
around
ongoing
compatibility
to
the
extent
that
you
know
that
you,
as
a
user,
are
interfacing
with
smi
and
its
apis
measury.
So,
as
I
skip
around
slides
here,
let
me
say
this
that
we
talked
about
just
briefly
like
what
a
management
plan
might
be
and
some
of
the
functionality
that
it
might
have
so
like
as
a
random
example.
A
Continuous
delivery,
your
you're
likely
familiar
with
the
fact
that
a
service
mesh
and
the
intelligence
that
it
offers
can
help
facilitate
like
canary
deployments
or
splitting
traffic,
or
doing
any
number
of
things
with
requests.
Any
number
of
things
with
traffic
that
flow
across
the
mesh.
One
of
those
things
can
be
helping
facilitate
continuous
delivery.
A
A
I
call
that
out
as
just
like
a
random
example
of
something
that's
not
fully.
It's
something
that's
provided
for
by
a
service
mesh,
but
not
fully
implemented
by
a
service
mesh,
so
a
management
plane
or
other
tools
like
a
popular
tool
in
the
space,
something
like
flagger
can
help.
Facili
can
use
the
power
of
a
service
mesh
to
perform
these
things.
A
Okay.
So
if
you
take
kind
of
another,
look
at
that
meshery
itself,
it
will,
by
the
way,
I'm
gonna
stop
speaking
here
in
a
minute,
so
you
could
so
not.
Everyone
has
to
listen
to
me
for
so
long
mesherie
will
deploy
in
kubernetes
cluster
or
it
will
deploy
out
of
kubernetes
cluster
on
like
a
docker
host,
either
way
mesherie
and
its
architecture.
It
comes
with.
A
Let's,
let's
go
down
here:
it
comes
with
an
operator,
a
kubernetes
operator,
and
so
when
you
connect
measurey
to
kubernetes,
it
will
automatically
deploy
an
operator
inside
the
operator
is
a
couple
of
controllers
and
one
of
those
is
called
mesh.
Sync
mesh
sync
facilitates
ongoing
discovery
and
synchronization
ongoing
discovery
of
service
meshes
that
you
might
already
have
and
synchronization
of
that
information
back
and
forth
to
measuring
mastery.
Will
it
does
life
cycle
management
of
service
meshes,
so
it
will
provision
and
configure
each
of
these
service
meshes
in
a
green
field
way.
A
A
Another
then
another
component
to
mushrooms
architecture.
Is
this
concept
of
providers
we'll
talk
about
those
another?
Is
the
fact
that
meshri
has
two
apis,
so
it
has
a
rest
api
and
a
graphql
api
mesh
reserver,
the
component
here
and
meshri,
has,
as
a
project,
has
two
of
its
own
clients:
a
command
line,
client
and
a
user
interface
client
a
ui
we'll
get
into
both
of
those
in
a
in
a
bit
as
well.
A
A
So
one
other
there's
just
maybe
two
other
things
that
I
will
try
to
quickly
impart
as
concepts
and
then
I'll
pull
back
and
maybe
we'll
we'll
take
a
look
at
some
of
these,
maybe
we'll
start
to
show
mesh
read
and
kind
of
poke
around
it.
So
two
well:
okay,
fine,
three
other
concepts.
Briefly,
if
you're
at
all
familiar
with
web
assembly
and
the
notion
that
webassembly
filters
can
be
written,
I'm
sorry
webassembly
programs
can
be
written
as
envoy
filters.
A
A
There's
another
concept
of
measuring
that
I
that
I
consider
really
neat
and
and
there's
been
a
lot
of
effort
around
this
recently
and
it's
this
notion
of
like
again
like
are
helping
users
helping
people
understand
the
common
patterns
of
usage
of
a
service
mesh
and
whether
or
not
you're
doing
it
right
whether
or
not.
How
is
your
neighbor
doing
it
like?
What
are
these
architectural
patterns?
How
are
people
deploying
them?
Where
are
they
being
successful?
A
How
do
you
know
if,
if
your
performance
is
doing
well,
how
do
you
know
how
sensitive
to
set
your
circuit
breaker
like
what
are
other
people
doing?
How
many
retries
should
you
have
between
different
services
like
the
answers
to
each
of
these
are
generally
like
a
massive?
It
depends
it
depends
on
your.
A
It
depends
on
like
a
number
of
things
that
are
specific
to
your
environment,
but
what
the
community
has
focused
on
is
bringing
for
identifying
common
patterns
of
use,
common
features
across
service
meshes
and
then
and
explaining
them,
and
then
making
them
available
as
importable
patterns
that
you
can
load
into
meshri
and
then
have
mesh
redeploy
those
and
then
eventually
then
have
meshri
verify
whether
or
not
you're
in
compliance
or
still
running
infrastruc.
Whether
or
not
your
infrastructure
has
drifted
from
that
pattern.
A
A
More
to
say
about
this
concept
of
patterns
and
trying
to
help
people
get
started
and
quickly
and
then
operate
their
service
mesh
easily
part
of
that
is
plugins
that
are
being
developed
from
mesherie
like
like
the
screenshot.
That's
here
we'll,
probably
take
a
look
at
meshmap
at
some
point
and
show
you
how
you
can
visually
design
you
you
can
import
these
patterns
and
then
modify
them
to
suit
and
and
visually
design.
A
The
only
the
last
thing
I'll
say
is
that
meshri
has
a
sibling
project.
It
actually
has
a
few
sibling
projects,
maybe
only
one
of
which
we'll
speak
to
there's
an
upcoming
open
project
office
hours
in
about
40
minutes
or
so
on,
service
mesh
performance.
It's
a
sibling
project
to
measure
it
is
a
specification
project.
Is
it
is
an
abstraction,
a
surface
mesh
abstraction,
it's
a
standard
way
of
capturing
and
characterizing
the
performance
of
well
most
prominently
of
a
service
mesh,
but
also
subsequently
of
cloud
native
infrastructure
and
how
well
that
infrastructure
is
performing.
A
A
A
Based
on
that
data,
plane
implementation,
then
mesh
re
bubbles
it
up
and
that
might
sound
like
magic
sauce
to
some
of
you,
some
magic,
sprinkles
and
sort
of
hand
wavy
like
oh
it.
You
know.
If,
if
the
surface
mesh
does
it,
then
you
know
mashery
will
and
it's
like.
No,
it's
not
a
hand.
Wavy
thing
like
it's
taken
a
lot
of
work
to
make
that
a
true
statement.
A
Oh,
it's
been
a
long
journey
to
get
there
the
adapters
themselves,
the
message
adapters.
They
have
quite
clearly
gone
through
a
re-architecture.
They
are
in
their
v2
with
like
there's
no
doubt
and
that's
actually
somewhat
similar
for
a
couple
of
other
components
in
measuring
measures
been
around
for,
like
like,
we
were
just
talking
about
how
long
it's
been
around
for
some
of
the
components
are
have
already
been
rewritten,
based
on
learnings,
based
on
changes
based
on
new
service
measures
coming
out
et
cetera.
So
I
think
it's
a
really
good
question
and.
A
D
Surely
so
ask
questions
come
in,
maybe
we
can
do
a
little
bit
of
a
demo
for
measuring.
So,
as
you
can
see,
this
is
the
measuri.
Dashboard
and
mushri
manages
a
lot
of
service
measures,
so
you
can
see
so
I
have
I
have
deployed.
I
have
measurey
steel
and
measuring
d
and
measure
the
app
mesh
adapters
currently
on
my
kubernetes
cluster,
but
I
can
also
have
all
the
other
service
mesh
adapters
in
my
cluster
as
well.
So
measuring
automatically
discovers
the
kubernetes
cluster
it's
in
and
it
also
has
a
service
discovery.
D
So
if,
if
you
bring
measurey
to
a
kubernetes
cluster,
that
already
has
some
service
meshes
installed
in
it,
then
it
will
automatically
discover
them
and
yeah
you.
You
can
just
use
measuring
from
there
itself,
so
mesh
management,
plane
or
even
more
service
mesh
management
platform.
Lets
you
manage
the
lifecycle
of
all
the
different
service
meshes
mentioned
here,
so
you
can.
You
can
go
ahead.
C
D
Install
a
service
mesh,
you
can
even
add
some
sample
applications
on
on
your
mesh
to
just
run
a
couple
of
benchmark
tests
and
to
see
how
different
service
measures,
work
and
contrast
and,
as
we
mentioned
before,
mesherie
also
has
also
validates
validates
the
runnings.
The
running
configuration
of
your
service
meshes
so
smi
conformance
being
smi
being
the
specification
that
server
smashes
adhere
to
so
measuring
machinery
being
the
being
the
service
mesh
agnostic
tool
tool
out
there,
so
it
it
validates
the
validate
smi
performance
of
your
service
meshes.
D
D
We
also
have
these
concepts
of
service
mesh
configurations,
so
measuring
patterns,
as
we
pointed
out
earlier,
offer
offer
a
offer
a
really
configurable
way
for
users
to
bring
in
their
own
service
mesh
configurations
and
apply
them
onto
their
service
measure
directly.
We
also
have
a
concept
of
web
assembly
filters,
so
users
can
bring
in
their
own
filters
and
use
it
to
extend
the
capabilities
of
service
meshes
like,
for
example,
studio
has
an
envoy
control
plane.
D
Another
really
great
capability
of
these
service
meshes
is:
is
the
performance
management
capability
of
measuring?
So
what
you
can
see
here
is
a
set
of
automated
tests
that
we
run
to
benchmark
different
service
meshes
using
measurey.
So
if
you
go
go
here
and
take
a
look
at
some
of
these
results,
you
can
see
how
how
different
service
meshes
or
you
can
even
go
ahead
and
compare
different
service
meshes
and
how
they
perform
under
different
environments.
And
what
all
these
statistics
are.
We
have.
D
We
have
a
load
generator
within
measuring
that
runs
prince
this
benchmark
test
against
deployed
service
measures,
service
mesh
on
the
application,
and
you
can
collect
metrics.
You
can
also
go
ahead
and
add
your
own
metrics
collect
your
own
metrics
from
prometheus
and
have
a
grafana
dashboard
to
display
those
metrics
as
well
and
yep.
A
C
So
for
bringing
up
mystery
initially
we
need
to
do
this.
I
have
the
measuring
up
and
running,
but
I'll
quickly
show
what
I
have
done.
I
have
started
if
you,
if
you
don't,
have
the
repo
locally,
you
can
start
with
this
command.
It's
a
cli
command
which
is
listed
in
the
machine
documentation,
quick
start
guide,
so
I'm
running
meshri
on
kubernetes
platform,
so
we
can
check
the
status
of
mastery
by
using
system
status
command.
C
So
machine
is
up
and
running
and
it
is
at
this
endpoint.
So
let
me
go
there
and
access
this,
so
it
would
take
me
to
the
same
ui.
What
naveendu
was
showing,
so
it
detected
that
I
am
running
a
kubernetes
cluster,
so
I
also
have
these
service
meshes
installed.
So
let
me
quickly
show
the
performance
test
that
I
have
already
run
so
this
these
can
be
triggered
from
the
ui
as
well
as
from
the
cli.
So
let's
first
check
the
syntax
of
the
commands
that
we
can
use
with
performance.
C
A
And
part
of
what
hussein
is
going
through
here
as
well
relates
to
mesherie's
sibling
project
service.
Mesh
performance.
Remember
recall
that
that's
a
specification
project
that
it
well.
I
won't
explain
it
again,
but
suffice
to
say
part
of
you
can
see
part
of
mesher's
implementation
of
service
mesh
performance
as
a
project
here
in
terms
of
how
it
defines
performance
profiles,
and
so
there's
some
ongoing
research
being
done
in
the
service
mesh
performance
project.
A
C
C
A
C
Had
to
show
one
more
part,
which
is
mystery,
ctrl
system
login,
so
we'll
be
using
two
providers,
whether
machine
provider
or
a
nun
provider.
So
if
you
use
machine
provider,
whatever
test,
we
run
would
be
persisted
and
will
be
able
to
view
them
again
when
you
log
in
again
so
I
have
already
logged
in,
but
I
would
show
the
command,
which
would
allow
us
to
login
to
a
particular
provider.
C
C
C
A
A
A
A
It
looks
like
it's
not
necessarily
speaking
to
kubernetes
fully
so
we're
gonna
have
to
wait
just
a
moment,
but
there's
a
couple
of
ways
in
which
you
can
configure
your
service
mesh,
so
one
of
those
is
to
do
so
using
a
form-based
system.
So
this
concept
that
we
were
talking
about
patterns
earlier
you
can
import
these
patterns
and
apply
these
patterns.
A
A
The
untitled
designs
or
patterns
that
aren't
really
there,
I'm
just
going
to
grab
this
one
it's
about.
I
think
I
think
it's
it
probably
I
mean
if
its
name
is
correct.
It
describes
a
deployment
of
book
info
the
sample
app
from
istio.
That
also
has
a
circuit,
breaker
configured
and
so
the
way
that
mescheri
stores
and
applies.
This
configuration
is
in
a
yaml-based
format.
This
is
the
pattern
file
format
we're
looking
at.
A
For
that
specific
version,
you
can
then
go
through
and
use
this
configurator
to
define
your
in
this
case
your
authorization
policy
particular
rules
about
where
messages
messages
are
coming
from
and
going
to,
and
so
you
can
use
this
form
based
flow
to
to
configure
your
mesh
if
we,
as
you
do
so
you'll
end
up
seeing
that
you'll
end
up
seeing
your
configuration
I'm
created
over
here.
This
is
a
beta
feature,
so
the
ux
is
in
need
of
improvement.
So
you
know
all
feedback
is,
is
a
welcome
here.
A
Meshery
itself,
by
the
way,
is
currently
in
its
v.
Sorry
v.5
release.
So
it's
about
halfway
to
its
1.0,
so
there's
there's
a
few
areas
still
being
worked
through
that
same
concept
of
like
configuring,
your
service
mesh
and
whatever
version
of
that
mesh
that
you
want
to
work
with
applies
here
in
this
in
mesh
map.
This
is
a
plug-in
to
mesh
re.
It's
a
visual
designer
in
which
you
can
go
about
configuring,
your
mesh.
So
if
you,
if
you
have
a
particular
application
or
service
that
you're
running.
A
Really
it's
just
a
some
random
example
of
taking
like
a
particular
service,
getting
it
adjoined
to
something
like
a
circuit
breaker.
If
you're
going
to
deploy
this
on
on
kuma
and
kind
of
visually,
designing
out
the
fact
that
you're
pulling
in
and
visually
design
what
it
is
you
want
for
your
mesh
to
do
so.
A
Maybe
these
are
like
three
versions
of
the
same
service
and
you're
wanting
to
have
a
circuit
breaker
from
the
first
part
of
that
that
service
over
to
the
remainder
of
these,
that
you
want
to
have
a
circuit
breaker
over
and
so
on,
or
whatever
that
configuration
is.
I
should
probably
have
a
better
example
just
off
hand,
but
the
point
is
like
there's
a
lot
of
ways
to
configure
each
and
every
mesh
and
each
and
every
mesh
at
a
particular
version.
So
someone
had
asked
a
great
question
earlier
about.
A
You
know:
mesher's
support
for
different
types
of
data
planes,
different
cni's,
different
data,
plane,
implementations,
and
it's
like
looking
at
this.
You
might
begin
to
get
a
sense
of
you
more
or
less
what
I
was.
The
answer
that
I
was
giving,
which
is
like
there's
a
lot
of
different
service
service,
meshes
supported
at
different
versions,
and
we
don't
nearly
have
enough
maintainer
fingers
to
have
typed
out
all
that
code.
So
the
support
for
these
different
meshes
is
now
meshri
self
generates
these.
A
A
Let
me
have
a
visualizer,
so
I
need
to
refresh
my
environment,
but
if
I
had
had
a
service
mesh
deployed,
we
would
see
its
deployment
here.
We
can
go
about
interacting
with
it
and
so
there's
kind
of
a
difference
between
these
two
modes.
This
is
more
of
a
view
mode
where
you
would
observe.
What's
going
on
on
your
active
infrastructure
and
then
over
here,
where
you're
designing
patterns,
maybe
you're
working
with
filters
as
well
webassembly
filters
you
you
do
the
design
here
and
then
watch
it
run
over
there,
so
good.
A
A
A
Meshery
is
opinionated
about
bringing
back
certain
measurements
measuring
certain
things
well
in
the
next
release,
measure
will
continue
to
be
opinionated,
but
it'll
also
be
fairly
configurable
in
terms
of
on
what
metrics
or
what
telemetric
signals.
It
is
that
you
want
to
look
at
whether
that's
logs
or
metrics,
or
traces
meshri
actively
lets
you
expose
the
logs
of
any
of
your
service
mesh
components
and
more
than
just
your
service.
A
So
people
ask
silly
questions
all
the
time
I
do,
and
I
think
it's
in
part
because
of
that
culture
and
the
way
that
we're
collectively
learning
and
uplifting
each
other
and
creating
some
pretty
cool
things.
I
think
measury
turns
out
to
be
the
most
popular
linux
foundation,
internship
program
or
the
project
that
has
the
most
popular
internship
program
more
than
more
than
kubernetes,
even
so
asuka
and
piyush
things
that
we
kind
of
missed
out
on
or
maybe
didn't.
Let
people
know
about.
A
A
Thank
you
very
much
for
getting
up
either
very
early
or
in
zuko's
case,
just
taking
a
having
a
casual
day,
I
suppose,
and
joining
and
sharing
today,
so
we'll
have
the
service
mesh
performance
project,
yeah
service,
mesh
performance
project
hours
coming
up
so
there'll
be
a
few
folks
from
intel,
maybe
a
red
hatter
and
some
hashicorp
folks
joining
as
well.
So.