►
From YouTube: Layer5 Community Meeting (June 11th, 2021)
Description
Layer5 Community Meeting - June 11th, 2021
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
Welcome
guys
welcome
to
the
community
meeting,
I
guess
we
can
get
started.
We
have
12
people
and,
let's,
let's
let
the
others
join
in,
as
we
start
so
today
is
june
11th,
and
this
is
the
layer,
5
community
meeting
and
if
you
haven't
already
put
down
your
names
on
the
meeting
minutes,
the
link
will
be
on
the
on
the
chat,
and
here
we
have
a
tradition
to.
A
Who
are
joining
this
call
for
the
first
time?
So
I
guess
we
have
a
couple
of
newcomers
today.
So,
let's
start
with
ashish
ashish,
do
you
want
to
introduce
yourself.
A
C
A
D
Sure
hi,
I'm
jean
I'm
a
web
developer
and
I'm
new
to
the
open
source
world,
so
yeah,
I'm
trying
to
see
how
I
can
contribute
to
this
community.
My
tech
is
django.
C
Very
good
what
so
it's
10
a.m,
where
I
am
what
what
is
it
for
you.
D
It's
currently
11
p.m.
Here,.
C
Good
yeah,
the
later
the
hour,
the
more
the
dedication
so
good
yeah,
it's
good
well,
but
just
think
by
the
time
we're
done
today,
we'll
have
known
each
other
for
two
days.
It
would
be.
C
Oh
nice
gene
I'm
nice
to
have
you.
I'm
gonna
chat
at
you
for
a
moment
just
because
what
is
josh
on
yeah
josh
josh
was
on
our
last
call
a
few
minutes
ago.
Well,
josh
could
tell
one
of
the
mesh
mates
he'll
tell
you
it's
well
unfortunate
or
fortunate
for
those
that
say
that
they
are
you
more
front-end
oriented
or
have
a
knack
and
a
focus
toward
those
things
they
get
pounced
on
pretty
quick
because
there's
so
much
need
for
that
type
of
a
focus.
C
There's
a
lot
of
people
who,
just
who
do
other
great
things
but
lack
the
talent
that
you
have
so
so
very
nice
to
have
you
there's
a
lot
of
stuff
to
to
wade
through
and
kind
of
well
like
the
rest
of
us
like
learn
as
you
go
there,
there's
like
to
just
to
help,
explain
kind
of
part
of
the
need
and
part
of
where
you
might
be
able
to
get
engaged
and
where
you
might
be
able
to
own
a
few
things
and
make
some
decisions
and
have
great
experiences
and
all
that
is
I'll,
say
a
couple
of
words
to
to
explain
it
to
you,
but
also
mostly
for
the
benefit
of
everyone
else,
and
that
is
that
there
are
well,
I
guess,
about
three
areas
of
ui,
centered
or
front
end
centered
or
user
experience
oriented
activities.
C
So
one
of
those
is
that
the
community
has
created
five
or
six
websites,
because
there's
a
number
of
different
projects
that
go
on
and
they
you
know
a
lot
of
them,
use
similar
tech
stack,
but
not
all
so
so,
there's
some
so
I
won't
exp
so
I'll
put
a
link
into
the
meeting
minutes.
That
explains
the
tech
stack
of
each
of
the
repositories.
C
This
is
kind
of
a
nice
reference
to
have
so
that's
one.
Another
one
is
that
the
the
biggest
project
here
is
the
mesheri
project,
and
it
has
a
web-based
interface
or
ui,
and
it
is
in
the
process
of
trying
to
feel
more
like
a
spa
or
a
single
page
application.
It's
trying
to
feel
less.
C
You
know
early
2000s
of
sort
of
click,
clicky
and
refreshing,
the
page
more
interactive,
so
it
actually
needs
a
bit
of
an
overhaul.
There's
I
use
who's
on
the
phone.
There's
a
bunch
of
people
on
the
phone
actually
who
are
on
the
phone
on
the
meeting,
who
are
would
be
good
people
to
be
introduced
to
because
they're
kind
of
focused
there.
C
So
it's
a
slightly
different
like
it's.
It's
a
web
ui,
but
it's
it's
not
a
website
and
so
different
designs,
different
different
goals,
different
user
flows
that
happen
and
the
the
third
way
is
there's
a
bunch
of
somewhat
random
and
ad
hoc
designs
that
need
to
be
done
between
trying
to
promote
peep,
the
community
members
and
what
they
do
and
try
to
create
new
logos
and
those
sorts
of
things.
So
so,
anyway,
welcome
nice
to
have
you
if
people
are
really
friendly
with
you
there's
a
reason
why.
D
Thank
you.
That
sounds
really
exciting.
So
much
so
much
is
happening
and
so
much
can.
I
can
contribute
to,
I
hope,
cool.
C
Cool
good
yeah
don't
be
as
an
encouragement
to
you
and
then
to
the
other
newcomers
as
well
as
be
tenacious
or
don't
don't
let
a
lack
of
a
response
from
anyone.
Stop
you
like.
That
would
only
be
because
they're
busy,
you
know
not
not
because
you
shouldn't
be
asking
on
about
something
you
should
so.
D
Nice
yeah
and
I
think
I
think,
everybody's
in
a
different
time
zone,
yeah.
E
A
Meet
you
next
up,
we
have
musa.
F
Recent
graduate
from
with
my
major
in
computer
science
and
actually
I'm
an
open
source
enthusiast
and
an
admin
of
gsoc
project
bench
routes.
It
is
basically
a
monitoring
application,
quite
similar
to
promises
we
can
say
and
but
with
having
some
of
the
different
ideas
like
goals
that
we
think
promises
doesn't
cover
up.
So
we
want
to
implement
that
now
in
our
project
and
also
like,
I
came
across
mystery
up
means
like
just
two
or
three
days
back
and
tram
inside.
C
Let's
catch
up
like
the
I
can
see
why
yeah
a
cool
project
bench
routes
like
you,
can
see
the
relationship
and
part
of
the
somewhat
similar
focus
or
or
related
focuses,
and
so.
D
F
A
F
Hello,
everyone-
I
am
naman
krithkov
and
I
computer
science
and
engineering
and
I
have
an
interest
in
artificial
intelligence
field,
and
I
like
to
work
in
python
and
I
am
new
to
open
source.
So
he
doesn't
understand
much
the
terms
that
you
all
are
using,
but
I
would
like
to
be
a
good
contributor,
so
I
think
it's
a
great
place
to
start.
Thank
you.
A
I
Yeah
so
yeah,
I'm
personal,
so
I
work
in
a
company,
so
I'm
mostly
as
a
devops
engineer,
but
almost
like
someone
else.
I
like
this
is
what
open
source
is
new
to
me
and
then
I
want
to
contribute
more.
So
there
are
some
place
where
I
did,
but
as
a
document
thing,
but
I
mostly
use
all
the
tools,
like
all
the
things-
and
I
recently
I
was
reading
about
this
easter
open
running
book.
I
So
I
found
this
so
I
I
mean
I
was
going
through
his
profile,
so
I
later
found
that
it's
been
working
on
this
layer
and
measuring
thing.
So
I
thought
like
let
me
introduce
into
the
community
and
then
like.
I
want
to
contribute
more
on
this
book.
C
Persona,
I
guess
I
just
we
have
to
start
off
with
an
apology
if
your
reading
is
still
up
and
running.
I'm
I'm
sorry!
Okay,
if
it's
putting
you
to
sleep
at
night,
then
I'm
sorry.
So
no,
but
that's
fantastic
persona!
C
I
Yeah,
we
not
argo,
I
I
mean,
like
I
do
the
poc
in
the
company
for
the
others.
We
mostly
use,
like
you,
know,
elm
and
then,
like
humanities
and
terraform
provider
itself,
to
deploy
it
yeah.
So
I
do
have
a
few
bit
knowledge
on
python
and
then
I
mean
I'm
more
knowledgeable
but
sometimes
write
telephone
providers
like
using
google
language
so,
but
not
very
frequently.
C
Yeah
curious
by
the
so
by
the
way,
just
so
that
everyone
like,
if
it
wasn't
obvious
from
the
discussion
with
gene
about
the
softer
side
or
the
front-end
side
of
some
things,
and
how
like
it
well,
if
it
wasn't
obvious
that
that's
very
much
so
needed
here.
C
It
should
hopefully,
after
this
it'll,
be
really
obvious
that,
like
there's
so
much
to
do
across
the
projects
and
in
the
community
that
so
this
isn't
a
reflection
on
what
gene
might
do,
but
just
toward
persona
that,
like
so
much
to
do
that
isn't
code
or
that
is
close
to
code,
but
but
not
exactly,
and
so
actually
it's
very
persona
again
like
there's,
there's
a
number
of
well
I'll
abuse
the
term
and
say
devops,
sentry
operation-centric
things
actually
jubrilla
is
on
the
call.
C
While
he
does
have
a
gopher
badge
on
his
a
tattoo
on
his
forehead.
He
also
has
been
focused
on
a
bunch
of
or
on
some
really
interesting,
github
workflows,
which
I'm
hoping
that
he
and
I
might
talk,
give
a
talk
on
at
a
conference.
And
so
point
is
it's
refreshing
to
to
have
an
operations
focused
individual
on
the
call.
Samir
is
on
the
call,
who
is
also
somewhat
operationally
oriented
and.
A
Thank
you
prasanna.
I
guess
that's
all
the
newcomers
we
have
on
the
call
today.
If
I
missed
anybody
else,
would
you
like
to
introduce.
A
Yeah,
I
guess
yeah
that
was
it
so
welcome
again
all
so.
We
do
have
a
newcomers
meeting
every
week.
It
happens
on
thursday
and
yeah.
You
can
join
it
next
week.
I
guess
to
get
started
and
if
every
every
one
of
you
must
have
access
to
the
layer,
5
community
calendar,
if
not,
you
can
just
let
us
know
on
the
chat
and
someone
will
fix
that
up
for
you.
A
H
Yup,
so
this
is
the
design
script.
This
is
about
integrating
pattern,
files
and
machining
data,
so
I
have
I've
been
demoing
pattern
files
for
a
while.
Now
this
is
about
linking
the
data
that
is
produced
by
patent
files
and
the
messaging
data
missing
is
basically
measuring
mechanism
to
basically
collect
all
of
the
data
from
the
queries
cluster.
It
basically
collects
all
of
the
information
that
water
resources
are
different
on
the
cluster
so
that
we
can
make
sense
of
it
in
the
mystery
server
itself.
H
So
this
pic
is
basically
about
that.
How,
when
resources
get
reduced
by
pattern
files,
they
are
basically
when
get
created
by
by
patent
files,
how
we
can
interrelate
those
resources
with
the
data
that
is
being
collected
by
machining.
So
this
is
a
very
small
space.
H
Make
some
changes
to
it,
so
this
will
change.
You
guys
can
check
out
the
update
later
and
let
me
describe
clearly
so
the
idea
is
that
we
will
actually
store
the
name
of
the
pattern
attribute
that
we
have
the
name
space
in
which
the
the
name
says
in
visual
resource
was
deployed.
The
way
that
it
was
created,
for
example,
was,
was
the
patent
file
applied
by
you,
or
was
it
picked
up
from
github,
because
we
we
actually
provide
the
function?
H
The
functionality
was
recently
added
to
patterns
that
you
can
now,
in
fact,
import
all
of
your
patents
from
your
repository
mystery
will
make
sure
that
reverses
the
entire
repository
and
trap
all
of
the
pattern
files
and
then
apply
it
for
you.
So
that's
also
function,
that's
available,
so
it
will
keep
track.
H
If
your
pattern,
where
your
pattern
file
actually
came
from
it,
would
be
mishma
which
is
actually
a
designer
which,
basically
it
will
be,
it
will
be
a
visual
designer
where
you
can
actually
design
and
then
deploy
your
patterns
right
from
there.
So
this
is
this
is
all
the
attributes
that
we'll
be
keeping
track
of
and
some
other
metadata
that
what
the
pattern
looked
like
and
what
the
home
type
was,
and
then
this
is
the
id,
so
the
previous,
the
previous
page,
the
state
that
is
being
written
here.
This
actually
had.
H
The
idea
have
the
notion
that
we
will
be
packing
all
of
these
information
in
the
kubernetes
annotations
and
then
we'll
pass
and
then
we'll
pass
it
along
to
every
resource
that
parent
file
actually
creates.
For
example,
if
you're
creating
an
application
that
leads
to
creation
of
a
accuracy
service
and
are
gonna
roll
out,
then
we
will
actually
pass
on
those
annotations
to
those
resources
so
that
we
can
keep
track
of.
H
But
today,
after
like
I
and
abhishek,
basically
got
involved
and
basically
were
inclining
towards
actually
passing
along
only
the
id,
which
is
for
the
uuid,
which
will
be
the
uid
and
let
the
let
me
sing
basically
capture
all
of
these
information
in
the
database.
For
us
that
is
so.
We
have
something
like
okay,
I
don't
think
I
have
it
available
here,
but.
H
That
machine
server
would
only
capture
the
uid
in
the
annotation
in
the
community's
annotation.
The
adapters
will
actually
take
and
take
in
those
annotations
that
are,
after
we've,
taken
those
annotations
and
then
we'll
pass
it
to
the
kiwis
cluster,
and
then
those
resources
will
be
created
now
when
mesh
sync
basically
reads
those
resources,
for
example,
it's
reading
the
kvd
service.
It
will
see
that
hey
there
is
an
id.
There
is
a
there.
Is
a
field
called
id
it
will.
It
will
create
a
mesh
sync
object.
H
Machining
object
is
a
generic
representation
which
measures
uses
to
actually
track
all
of
the
metadata
associated
with
the
kubernetes
objects.
So,
basically,
mesh
sync
is
going
to
get
grab
all
of
those
metadata
and
feed
it
in
there
and
now
send
it
send
that
thing
to
the
machine
server
and
our
memory
server
is
basically
keep
track
of
that
particular
resource,
just
like
it
keeps
track
of
other
machining
messaging
objects.
So,
basically,
in
this
way,
what
we
achieve
is
that
meshing
does
its
responsibility.
H
It
plays
its
role,
that
is,
it
identifies
all
of
the
resources
and
then
it
packs
it
pass
it
up
and
sends
it
to
ministry
server.
Now
machine
server
doesn't
have
to
do
anything
else,
other
than
storage,
storing
it
in
the
database
that
that
is
exactly
what
mystery
does
even
today.
So
basically,
this
particular
new
change,
which
mostly
shakes
a
streamlined,
a
lot
of
process
and
actually
basically
made
the
responsibilities
quite
more
clear,
now
mystery
server
just
creates
id
and
that
other
things
are
being
handled
by
the
missing
on
the
other
side.
H
So
this
is
a
quite
a
small
stick.
The
changes
are
described
that
they
are
not
presented
here,
because
we
got
up
all
about
30
minutes
ago.
So
I
didn't
have
time
to
update
this,
but
this
the
changes
that
I
was
describing.
This
will
be
reflected
there
also
very
soon,
but
does
anyone
know-
and
I
hope
I
was
horrible.
C
What
was
that,
did
you?
Were
you
talking,
okay,
bad
joke,
so
did
you
mention
that
there
or
did?
Did
you
mention
a
sub?
A
recent
conversation
that
you
had
had
with
another
maintainer
and
potential
changes
forthcoming.
H
Yeah,
I
and
bishop
were
on
call
for
a
few
hours
at
least,
and
there
we
actually
discussed
this
entire
thing
so,
and
it
was
more
about
how
we
can.
H
We
can
not
replicate
this
thing
everywhere
and
pass
along
these,
this
baggage
to
every
kiwis
resource,
and
rather
what
we
will
do
is
just
tag
an
id
iu
id
so
that
we
can.
We
can
capture
different
instances
of
the
resources
being
created
by
the
mystery
server
by
patterns,
and
let
me
sing,
actually
let
me
think.
Actually
I
get
the
understanding
of
what
the
idea
is
and
then
create
the
packet.
So
basically
that's
what
we
arrived
at,
that
these
metadata
should
not
be
actually
propagated
via
acumen's
ambition.
H
Rather,
mesh
thing
is
just
is
going
to
be
quite
more
intelligent
in
terms
of
attaching
patterns,
metadata.
C
Nice
good,
I
got
that
last
part.
H
I
I'll
actually
probably
was
able
to
describe
it
more
clearly
in
this
way.
I
I
couldn't
get
time
to
beat
this
thing
before
the
ball.
C
Nice,
so,
by
the
way,
I'll
only
does
anybody
need
more
want
more
context
on
this
topic,
or
would
you
rather
just
have
some
time
to
digest
it
onto
your.
J
Own,
I
would
like
to
know
support
more
basic
information
like
what
are
patterns.
What
are
pattern
files,
so
you
can
point
me
to
that.
So
I'm
just
just
think
I'm
a
zero.
In
this
I
mean
like
a
lot
of
other
things,
but.
C
Yeah,
sorry
pj,
I'm
sorry,
I
chuckled,
because
I
thought
you
said
just
consider
that
I'm
zero.
You
know
I'm
starting
from
ground
zero
on
this
and-
and
I
thought
you
said
that
you
don't
know
a
lot
of
other
things
which
made
me
chuckle,
because
it's
like,
because
that's
not
true,
and
so
anyway
I
didn't
mean
to
interject.
I
just
wanted
to
explain
why
I
was
laughing
so
bj.
That's
a
great
question.
C
As
a
matter
of
fact,
I
gotta
believe
that
everybody
else
is
most
of
the
people
on
the
call
are
in
about
the
same
position
that
you
are
with
respect
to
understanding
what
a
what
a
pattern
is.
There's
a
couple
of
reasons
for
that
one
is
because
well
we
made
it
up,
so
so
you
know
yeah
and
okay.
When
do
you
time
me
if
you
would
like?
Let
me
let's
cut
me
off
after
five
minutes
and
then
that
way
and
crash,
do
you
wanna?
H
I
I
don't
try,
but
probably
we'll
have
to
fill
in
so
yeah.
Basically
the
idea
of
pattern.
What
pattern
is
right
now
is:
basically
you
get
to
create
some
yarmouths
in
which
you
can
specify.
You
can
basically
specify
that
how
you
want
to
configure
your
service
mesh
or
what
so
controlling
service
mesh
is
one
aspect
of
it.
That
is
a
definitely
you
can't
create.
You
can
provision
a
seo
service
mesh.
You
can
configure
mtls
on
it.
You
can
configure
automatic
site
conjection
on
it.
H
The
other
resource,
the
other
aspect
of
it
is
which
is
not
shown
in
this
particular
budget,
while
is
that
you
can
actually
provision
or
basically
deploy
your
applications
via
patent
price.
That
is,
you
can
specify
that
I
want
to
provision.
I
want
to
create,
deploy
this
particular
application
onto
my
cluster
and
that
application
actually
looks
like
this.
This
is
the
docker
image
that
is
supposed
to
be
required.
You
can,
you
would
be
soon
able
to
even
specify
that
hey.
H
I
want
to
do
candid
rollout,
and
this
is
the
way
that
you
should
be
doing
it.
That
is
pass
only
two
percent
traffic
to
it
then
wait
for
this
much
time
or
run
this
analysis
and
then
promote
it
to
the
promoter
to
ten
percent
traffic.
Wait
for
some
time
so,
basically
stuff
like
that.
So
it's
a
pattern.
Pattern
files
in
general
can
represent
a
lot
of
things.
There
are
multiplayers
are
the
same
way.
H
Okay,
I
think
that
was
this
disturbance
yeah.
You
can
actually
specify
multiple
things
in
patterns,
one
of
the
one
of
it
is
a
configuring
mesh
which
is
I'm
showing
here,
I'm
not
sure
if
we
have
a
one
for
rollout.
This
is
an
older
pattern
file.
This
is
this
may
not
be
valid
for
others,
but
this
this
also
shows
another
aspect
of
patterns,
and
that
is,
you
can
actually
specify
that
you
want
to
deploy
a
application.
H
It's
this,
this
docker
image
run
it
on
this
eastwood
reports,
and
now
you
want
to
do
calorie
on
it.
That
is
if
it's
a
first
instance
do
not
do
gallery.
H
If
you
are,
if
you
are
upgrading
it
to
version
six
versus,
let's
say,
then
what
would
happen
is
that
only
twenty
percent
traffic
would
be
sent
to
the
sent
to
your
new
calendar
instance,
while
the
rest
would
be
sent
to
the
older,
stable
version,
and
then
it
will
be
for
20
seconds
and
then
we'll
send
40
percent
then
wait
for
10
seconds,
and
so
basically
this
is
the
end
of
the
aspect
of
file
and
file.
So
if
that
that
made
sense.
J
H
J
It
application
specific
or
does
every
it
can
be
used
for
more
is,
can
it
be
used
for
more
than
one
say,
for
instance,
a
pattern
file
for
deployment
pattern
file
for
conflict
configuration
is,
is
it
application
specific
or
would
it
be
a
user
specific
or
say
if
you
want
to
deploy
something
onto
a
for
a
particular?
Well
I
mean
we're
talking
about
cloud.
So
there's
no
such
thing
as
deploying
for
a
person
right,
but
does
each
application
have
a
different
pattern
file.
H
So
it
will
sort
of
you
can
stuff
in
all
of
your
deployments
in
in
a
same
pattern
file.
You
can
split
it
up,
it's
a
so
it's
not
a
restriction,
that's
being
imposed
by
pattern
files,
so
you
can
see
that
here
I'm
describing
that
my
name
of
the
application
is
actually
svc.
The
type
is
rollout.
What
you
can
do
is,
you
can
just
add.
The
field
called
scc1
specify
the
type
as
loadout
and
just
basically,
you
can't
even
copy
and
paste
this
thing.
H
So
what
it
will
do
is
it
will
create
an
application
for
you
called
svc1,
and
it
will
have
available
on
this
particular
port
if
this
port
is
actually
available
to
you
so
yeah.
This
is
not
a
restriction,
that's
being
enforced
by
patents.
You
can't
split
it
up
or
keep
it
in
the
same
pattern
as
per
the
need.
G
Yeah
to
give
more
information
to
that,
so
this
pattern
files
are
more,
like
you
know,
from
a
istio
prospective
or
maybe
service
mesh
perspective,
not
from
an
application
perspective.
So
whatever
you
want
to
see
in
your
service
mesh,
you
can
put
them
into
this
pattern
files
and
you
can
not
do
all
of
the
things
what
you
want
to
do
from
the
camel.
Maybe
you
can
do
it
from
here
that.
C
Is
in
part
true,
although
gurav
is
gonna,
is
going
to
expand
on
what
samir
said
and
he's
gonna
show
us
how
pattern
files
can
be
used
for
applications,
which
is
the
line
that
it
actually
ends
up
being
both
it
ends
up
being,
quite
you
know,
yeah,
let
me
yep
it
isn't
being
quite
deep
on
applications
as
well
garage,
a
quick.
K
C
C
In
essence,
to
answer
jubilee's
question:
when
meshi
is
presented
with
a
pattern
file
it
it
creates
a
a
series
of
workflows
to
I
don't
we
don't
use
the
term
around
these
parts
much
but
to
it,
creates
a
series
of
workflows
to
orchestrate
the
instantiation
of
your
infrastructure,
interfacing,
like
primarily
with
kubernetes,
and
so
it
creates
a
directed
acyclic
graph,
a
dag
that
the
dag's
more
responsibility
is
more
or
less
to
understand
and
identify
the
sync,
the
sequence
by
which
the
various
components
of
the
various
pieces
of
infrastructure
and
configuration
for
that
infrastructure
needs
to
happen
within
that
pattern.
C
And
so
it
passes
it.
So
so
it
passes
along
to
measure
adapters
and
yeah.
So,
let's
do
something
else,
because
I'm
sensitive
to
the
fact
that
some
of
us
on
the
call
are
still
asking
the
very
first
question,
which
is
what's
a
pattern
again
like
not
they're.
Not
we're,
not
curious
about
the
the
technical
details
of
like
how
it
is
that
measuring
you
know,
digests
them
and
what
it
does
in
terms
of
in
terms
of
like
there's
a
lot
to
that.
C
But
to
first
start
over
here
I
think,
is
sort
of
vendor
for
five
minutes.
C
You,
like
mr
patel,
if
he's
on
he's
one
of
the
few
people
that
feels
at
leisure
to
just
cut
me
off,
which
is
which
is
great,
so,
okay,
so
for
everyone
here,
you're
you've
got
some
level
of
familiarity
with
mescheri
as
a
service
mesh
manager.
One
of
the
things
that
mesh
re
does.
It
seems
it's
kind
of
a
prominent
feature
at
the
moment,
but
it's
kind
of
a
side.
Note
it's
a
really
powerful
side.
C
Note,
and
the
side
note
is
that
of
the
things
that
meshri
does
it's
able
to
do
them
across
any
service
mesh?
That's
part
of
the
goal,
part
of
the
belief
of
the
project.
The
kind
of
vision
is
that,
even
though
certain
service
meshes
will
enjoy
more
popularity
and
more
use
that
the
world
won't
be
made
up
of
just
one
like
that's
a
that's
a
fact
today
and
will
be
for
a
long
time
so
of
the
things
that
mastery
does
it
need.
C
These
capabilities
need
to
be
written
in
such
a
way
that
they
broadly,
it
won't
always
be
the
case
for
every
single
little
thing,
but
broadly
measuring
needs
to
be
able
to
do
the
same
things
across
all
the
meshes
in
part.
It
never
will
because
each
service
mesh
isn't
made
equal,
they
are
made
similar,
but
they
do
different
things,
and
so
measury
isn't
doesn't
orchestrate
in
the
lowest
common
denominator
fashion.
It
tries
to
embrace
the
and
the
speci
proprietary
capabilities
of
each
service
mesh
and
empower
people
to
use
them
in
an
easier
way.
C
Harness
the
capabilities
of
a
mesh
is
through
a
pattern
files
like
a
simple
set
of
a
simple
set
of
yaml.
That
says,
maybe
you'd
like
to
use
a
service
mesh
to
define
timeouts.
So
if
you've
ever
visited
a
website
and
you're
saying
you're,
seeing
the
spinner
on
the
website
and
you're
waiting
for
it
to
come
back
and
like
you're
wondering,
should
I
refresh
my
page,
should
I
give
it
another
30
seconds
and
waste
my
life?
C
Should
I
just
leave?
Should
I
like
well,
thankfully,
like
the
further
that
we
go
into
the
future
here,
the
more
intelligent
infrastructure
like
service
meshes
will
be
pervasive
and
the
more
that
people
will
be
using
things
like
timeouts
to
intelligently
make.
You
know,
prevent
you
from
having
to
wait
when
a
web
page
is
or
isn't
going
to
come
back
like,
hopefully
the
infrastructure
answers
that
question
for
you
and
does
the
right
thing.
C
But
how
do
the
people
that
are
running
that
infrastructure
know
when
a
timeout
should
be
set
or
not,
and
how,
for
how
long
and
like
that?
Actually,
that
answer
actually
changes
and
it
changes
sometimes
real
time
based
upon
what
services
are
running
and
based
on
a
lot
of
other
signals,
and
so
service
measures
are
quite
complex,
and
so
so
is
running
this
infrastructure
to
help
make
it
easier
and
to
help
people
and
be
comfortable
and
answer
the
question
whether
or
not
they're
doing
it
right.
C
It's
nice
when
you're
able
to
ask
and
have
answered
like,
am
I
doing
this
right?
How
are
other
people
doing
it?
What's
the
common
pattern
for
configuring,
a
timeout
or
configuring
any
of
this
stuff
and
sometimes
you're?
Looking
for
a
template,
sometimes
you're,
like
look
I'm
comfortable
with
the
template,
we've
customized
it
I'm
comfortable
with
the
pattern
we've
customized
it
we
have
our
own
we've
tweaked
it
now.
I
want
to
make
sure
that
our
infrastructure
runs
in
this
way
or
that
it
dynamically
responds
and
in
accordance
with
some
parameters
that
you've
defined.
C
Now
I'm
getting
too
abstract.
Part
of
the
point
here
is
that
we
spend
a
lot
of
time
in
the
cncf
and
within
there
we're
trying
to
curate
a
set
of
these.
What
these
patterns
are
part
of
doing
so
is
because
well
I'll,
say
we
like
collectively,
are
writing
a
book
service
mesh
patterns.
C
So
this
is
the
this
is
like
the
next
book
to
be
produced
and
it'll
probably
be
two,
this
one
being
foundational
patterns
and
the
other
one
being
advanced
right
now,
there's
about
60
patterns
that
we
think
will
be
defined
and
when
they
are
defined
they'll
be
published
publicly.
Well,
not
only
in
the
book,
but
just
on
github
as
well.
C
Mescheri
will
is
the
tool
that
will
be
used
to
implement
each
of
these
patterns,
so
you
can
so
that's
what
utkarsh
is
showing
is
that
you
can
feed
measure
these
patterns
and
it
will
let
you
experiment
with
them.
Let
you
run
your
infrastructure
in
that
way,
and
the
patterns
need
to
be
quite
expressive.
They
need
to
be
able
to
describe
what
workloads
it
is
that
you
want
to
onboard
onto
the
mesh.
What
service
mesh
you
want,
what
version
and
how
you
want
it
configured
and
boy.
That's
that's!
C
So
yeah
the
other
things
to
look
at
as
everyone
sort
of
digests.
This
is
that,
if
I
can
stop
talking,
soham
might
show
the
fact
that
the
intention
is
for
measuring
to
visualize
these
and
what
a
pattern
looks
like.
So,
if
you
looked
at
like
what
is
the
circuit
breaking
pattern,
look
like
in
my
environment
like
this
is
a
mock-up,
but
like
here's,
what
my
environment
looks
like
it's,
these
interconnected
set
of
nodes
and-
and
I
want
to
define
when
a
circuit
should
break
like
when
the
infrastructure
should
do
something.
C
I
want
to
apply
this
pattern
and
so
measuring
it
outside
of
just
yaml
mesh
intends
to
visualize
these
things
and
so
yeah
I'll
stop
there.
I
expect
like,
for
my
part,
I
expect.
G
So
lee
one
question:
so
would
we
be
able
to
deploy
an
application
with
this
patent
files.
L
L
So
yeah
here
what
the
basic
ui
of
the
machine
application
machine
applications
would
look
like,
so
the
user
would
be
able
to
manage
and
make
changes
in
their
applications
through
this
and
submit.
What
was
your
question
about
the
basically
about
the
application
application
management.
C
This
is
great,
so
samir
just
this
is
proof
that
mesher
will
treat
applications
as
a
first-class
citizen.
But
in
answer
to
your
other
question,
what's
the
difference
between
how
you
describe
an
application
using
a
pattern
file
and
how
you
describe
an
application
using
like
a
kubernetes
manifest
or
a
helm,
chart.
G
C
C
Do
you
wanna,
do
you
wanna
say
like
one
of
the
advantages
is
that
you
can
describe
your
service
mesh,
its
configuration,
the
the
workloads
that
you
want
onboarded
on
there
or
to
be
running
on
there.
C
You
know
and
like
the
fact
that
you
might
want
to
run
a
performance
test
at
the
same
time
as
well
like
while
you're
performing
a
canary,
and
maybe
the
strategy
of
the
canary
is
not
duration
based,
but
maybe
performance
based
in
terms
of
how
you
roll
out
that
canary
that
you
can
describe
all
that
in
one
file.
That's
that's
one
advantage.
G
But
it
also,
I
can
do
in
the
manifest
with
using,
like
you
know,
breakers,
like
three
high
fans
and
then
write
another
manifest
with.
Like
you
know
the
crd
of
this
istio,
then
I
can
also
explain
like
maintain
my
canary
deployment
information.
There.
C
G
C
Or
something
okay,
you
also
couldn't
describe
a
performance
test
in
there
either
and
the
fact
that
maybe
you'd
want
for
meshri
to
generate
some
load
or
meshwii
to
replay
traffic
that
it
had
captured
of
your
used
like
real
traffic
captured
from
your
users
that
you'd
like
to
replay
as
part
of
the
canarying
of
your
next
release,
like
none
of
the
messages
do
that
or
to
generate
load
like
maybe
just
generate
synthetic
load,
to
measure
it
and
then
have
that
be
part
of
the
gating
criteria
as
to
whether
or
not
the
canary
continues
okay
and
to
potentially
describe
that,
like
I
said
like
all
in
one
file
potentially
or
you
can
break
it
out
into
multiple
files
for
from
an
application's
perspective,
messaging,
isn't
necessarily
intending
to
like
the
app
like,
isn't
really
trying
to
bring
a
bunch
of
special
sauce
to
the
management
of
applications.
C
Just
yet
like
so
you
can
interchange,
you
don't
need
to
rewrite
your.
You
can
take
your
kubernetes
manifest
and
load
them
here,
and
meschery
will
deal
with
converting
them
from
manifest
to
pattern
file
format.
And
so
then
it
will
also
let
you
retrieve
them
in
pattern.
File
format
or
back
in
manifest
format,
yeah.
C
Interfacing
with
mystery
is
not
intended
necessarily
to
be
the
source
of
truth
for
well
for
a
lot
of
things,
including
the
definition
of
your
application.
So
so
we
would
bi-directionally
interface
with
github,
to
the
extent
that
you
ended
up,
making
a
configuration
change
to
your
application
here,
that
it
would
help
facilitate
a
pull
request
or
just
simply
do
a
push
or
what
have
you
to?
C
I
keep
saying
github
but
to
your
source
control
the
so
these
are.
These
are
helpful
things
I
think
those
are
compelling
things,
but,
but
one
of
the
things
that's
really
compelling
about
it,
I
think,
is
that
you
can
visually
configure
your
application
in
the
in
mesh
map
in
in
an
extension
to
measure
so.
C
C
M
C
But
but
rather
through
through
demonstration-
I
don't
know
if
you
guys
wanna.
M
Cool,
so
talking
about
pattern
files
in
measure
ui,
you
can
upload
your
pattern
files
over
to
a
dedicated
page.
M
This
page
facilitates
you
to
edit
your
pattern,
files
and
upload
them
view
them,
basically
the
thread
operations,
but
inside
mesh
map
we
have
a
designer
pane
in
the
design
of
you,
which
lets
you
see
all
of
the
patterns
that
you
have
uploaded
and
once
you
have
uploaded
a
pattern
file,
you
can
just
visualize
your
pattern,
files
inside
your
canvas
and
then
you
can
basically
play
with
it
like
draw
edges,
and
then
you
can
add
the
protocol
and
the
port
number
for
the
edge.
M
It
basically
does
all
that.
Maybe
if
you
want
to
create
copies
of
as
you,
you
can
do
that
also,
and
the
other
thing
is
that,
if
you
don't
want
it,
you
can
just
reset
everything,
and
it
just
goes
back
now.
If
you
are
satisfied
with
your
design.
M
For
example,
if
you
have
such
a
design
and
you're
satisfied
with
it,
you
can
either
save
it
as
a
new
pattern
file
which
would
be
shown
here
if
you
click
save,
and
the
other
thing
which
was
talked
about
earlier
by
kash
is
the
deploy
feature
which
is
here
that,
if
you
want,
you
can
directly
apply
or
deploy
your
design
using
this
button
over
here,
so
pardon
files
is
what
a
mesh
map
is
built
upon.
I
guess,
if
I'm
not
wrong,
that's
for
the
demo
right
now
coming
to
the.
M
Updates
so
last
time
we
saw
that
we
could
create
our
edges
and
etc.
So,
this
time
I'm
going
to
present
a
slightly
different
thing,
which
is
a
filter.
M
M
Now
we
can
also
drag
and
drop
filters
over
a
node
and
then,
as
soon
as
we
drop
filters
over
our
node,
we
get
a
tooltip
which
prompts
us
to
basically
configure
our
filter
and
also
yeah.
This
is
something
that
I
started
working
on
very
recently,
so
it
has
a
couple
of
bucks.
M
For
example,
there
should
be
an
svg
which
is
currently
not
loading,
and
also
we
also
I
had
to
implement
something
called
as
json
form
schema,
which
basically
creates
this
form
automatically
by
identifying
the
filter
that
I
am
picking
up
and
by
identifying
the
node
that
I
am
dropping
it
on
which
I
will
work
upon
right
now,
but
yeah
right
now.
You
can
drag
and
drop
filters
upon
nodes,
so
that
was
the.
G
G
Yeah,
so
I'm
really
sorry
if
I'm
asking
so
many
questions,
so
I
just
try
to
understand
now.
As
you
told
we
can
deploy
an
application
with
a
pattern
file.
So
once
I
do
a
deployment
with
the
pattern
file,
if
I
go
to
cube
ctl
describe,
will
I
be
able
to
see
that
from
the
give
ctl
command
prompt
or
should
I
go
to
ist
or
ctl
or
where
should
I
or
should
I
go
to
mystery
cpl
to
see
the
configurations?
G
M
G
M
Create
features
which
will
let
you
do
everything
over
mesh
map.
You
will
not
need
to
go
anywhere
else
than
bash.
Now
you
you'll
be
able
to
do
everything
that
you
can
on
mesh
map,
be
it
cube,
ctl
deploying
application,
be
it
visualizing
designing.
That
is
what
we
intend
to
do
in
a
stretch
goal.
I
I
M
Yeah,
if
the
like,
for
now,
I
guess,
created
the
endpoints
for
applying
the
patterns
and
for
now
the
application
pattern
that
gets
deployed
is
currently.
What
should
I
say.
M
Restricted
in
a
sense
lee
I
I
guess
you
can
help
me,
explain
it
a
little
bit
but.
C
Yeah,
I
think
the
answer
I
think
if
I
read
past
what
prasanna
was
asking,
it's
sort
of
that
there's
there's
two
modes
to
mesh
map,
so
the
mode
that
sohum
is
kind
of
walking
us
through
is
the
designer
mode
in
which
you're
really
not
looking
at
live
infrastructure
but
you're.
C
I
C
And
then
you
can
like,
like
you,
can
tell
like
this
is
an
early
project,
but
you
know
you
so
you
can
save
it
or
or
apply
it
and
have
it.
You
know,
and
that,
but
there's
a
second
mode,
and
I
don't
know
that
soham's
system
is
set
up
to
be
able
to
show
this
very
well
at
the
moment.
But
but
if
you
can
show
the
details,
pain,
there's
sort
of
the
left-hand
side,
we
have
this
designer
pain.
Where
you
go
into.
C
You
know,
designing
your
you
know
doing
what
we
just
said
on
the
right-hand
side
is
more
of
what
you
might
be
used
to
from
other
visual
topology
tools.
So
this
mode
is
the
visualizer
mode
and
it
would
be
showing
you
what
is
actively
running
and
you
know
metrics
on
on
your
current
deployment,
and
so
you
just
switch
between
read-only
mode.
So
to
speak.
You
know
visualizer
mode
and
designer
mode,
and
there
isn't
this.
It
doesn't
have
data
loaded,
but
mesh
sync.
C
I
A
C
C
Yes,
is
the
answer
to
not
only
in
mesh
map
in
that
plug-in,
but
also
just
in
your
pattern
files,
if
you
want
to
that,
you
you'd,
be
able
to
describe
your
auth
and
in
your
auth,
z,
config,
like
you'd,
be
able
to
describe,
and
if
that
has
to
deal
with
a
filter
like
non-white
filter
that
that
that
also
is
describable
inside
of
a
pattern
file
which
means
that
inherently
it's
visualizable.
C
Oh,
my
gosh,
that's
not
a
word,
but
you
can
visualize
that
in
in
mesh
map
as
well,
and
so
such
that
when
you
drag
and
drop
actually
I
don't
think
you
showed
filters
in
the
demo,
but
when
you
drag
and
drop
a
filter
over
like
let's
say
that
you're
configuring,
your
service
mesh,
you
drop
over
a
filter
and
that
filter
deals
with
authorization
and
authentication
that
of
what
soham
had
shown
was
that
there
was
immediately.
C
There
was
a
pop-up
to
to
further
configure
that
particular
entity
on
the
map
and
so
yeah,
whatever
the
and
so
mesh
map
itself
is
entirely
ignorant
of
what
what
things
are
possible
to
configure
and
how
does
it
know
whether
or
not
to
show
you
a
pop-up
of
the
port
and
protocol
and
other
stuff?
It's
like.
Well
it
doesn't.
It
just
asks
the
it
just
looks
to
the
pattern
file
it
just
it
just
looks
to
mesh
sync.
C
So
I
know
that
there's
some,
I
think
I
think-
and
I
know
that
there's
some
blanks
here-
some
some
blank
spots
in
some
people's
minds
as
to
how
this
goes.
It's
kind
of
a
lot
to
explain
it's.
It's
a
work
in
progress
and
and
as
people
show
point
like
one-off
things
that
are
being
worked
on.
It's
it's,
not
the
full
story,
and
so
but
the
questions
are
fantastic.
Please
keep
the
questions.
Coming,
did
the
so.
A
Yeah,
as
we
are
running
out
of
time
lee
like
did
you
want
to
talk
about
changing
the
format
of
the
meeting
yeah
yeah.
C
Hey
we
we
meet
at
least
once
a
day
on
various
topics,
we're
going
to
kick
off
the
first
web
assembly
focused
meeting
on
monday
and
if,
by
the
way,
if
you
want
to
know
what
all
the
meetings
are,
you
can
just
go
to
meet.layer5.io
and
it
will
bring
you
to
this
calendar
and
tell
you
some
info
about
each
of
the
meetings.
What's
going
on
so
there's
a
so
samir.
C
Very
much
so
like
your
use
cases
on
the
agenda,
I
think
it's
the
first
thing
on
the
agenda
for
and
if
you
can
make
this
oops,
it's
got
the
wrong
link
or
I
clicked
the
wrong
one.
But
if
you
can
make
the
meeting,
please
do,
if
not
we'll
be
capturing
your
use
case
yeah.
It
is
the
wrong
link.
C
Oh
good,
nice,
what
was
I
trying
to
say?
Yes,
that
we
run
these
meetings
and
the
meetings
have
some
amount
of
consistency
to
them?
We
make
bad
jokes.
We
welcome
new
people,
we
do
demos,
we
we,
you
know
there's
like,
but
I
wonder
if
like,
but
this
isn't
this.
This
is
your
meeting
as
much
as
anyone
else's,
and
so
maybe
it's
really
boring.
Maybe
it's
too
complex,
maybe
it's
too
simple.
C
Maybe
we
should
incorporate
like
standing
sections
standing
topics
in
the
meeting
like
a
short
tutorial
on
something
and
not
given
from
me
necessarily
but
from
any
of
you
or
I
don't
know
like
you
know.
Just
I
wanted
to
raise
up
that
question.
Have
people
think
about
if
we
should
change
how
these
get
run.
C
C
There
have
been
a
lot
of
people
working
on
meshri
ctl,
the
command
line,
client,
it's
been
fantastic
and
there's
a
spreadsheet
that
tracks
all
the
feature:
functions
of
meshrectl,
the
command
line,
client,
and
so
it
had
been
decided
not
too
long
ago
that
every
meeting
we
should
spend
just
you
know
five
minutes
to
look
at
that
spreadsheet,
so
that
we're
we
know
that
we're
tracking
the
whole
all
of
those
changes
very
well,
and
so
that's
just
an
example
of
what
I
meant
by
you
know
having
some
recurring
topics.
C
Potentially
I
like
in
this
one
like
on
this
one.
Should
we
be
using
this
one
to
hear
presentations
from
external
projects
like
and
we've
done,
a
bunch
of
them.
In
the
past
we
had
caverno
recently
bringing
someone
bring
someone
from
argo
or
from
from
other
related
projects
that
we
interface
with.
I
know
the
I
have
known
the
prometheus
co-founders,
just
as
an
example,
because
prometheus
was
on
here
for
five
years
now,
but
like
like
there
isn't
a
project
maintainer
that
I
don't
that
that
we
don't
have
a
relationship
with.
C
A
Thanks
lee
yeah
like
we
can
get
back
on
that
on
slack
like
if
you
have
suggestions
or
if
you
have
any
more
ideas
to
what
we
just
mentioned
like
reach
out
on
slack
with
that,
I
guess
we
are
over
time
with
our
meeting.
So
we
have
missed
a
couple
of
items.
A
So,
if
you
are
looking
to
contribute
to
measuring,
then
we
we
have
a
recording
profit
on
the
last
week's
call.
So
it's
in
the
meeting
minutes
and
we
we
missed,
also
missed
on
jubril's
progress,
update
on
the
github
action,
so
I
guess
jubril
will
be
talking
about
it
on
slack.