►
From YouTube: Meshery Development Meeting (Feb 3rd, 2021)
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
Hey
great
and
there's
my
warden,
the
the
messiest,
I
guess
is
the
the
messiest
is
probably
the
right.
A
A
All
right,
hey
while
we're,
while
people
are
filtering
in
if
you
have
topics
today,
toss
them
in
and
we'll
we'll
be
off
in
less
than
two
minutes
about
90.
A
A
A
A
A
I
tell
you
what
you
know:
what's
it's
being
the
one
who's
in
responsible
for
trimming
the
recordings
of
these
meetings
I
feel
like
I
can
make
all
kinds
of
horrific
jokes
at
the
start
and
then
we'll
just
trim
those
off
and
we'll
start
the
meeting
so
good
cover
your
ears.
While
I
get
these
jokes.
A
A
A
So
it's
we
should
probably
make
that
a
topic
here.
So
do
do
note
down
any
topics
that
you
might
have
as
we
get
going
here
on
the
topic
of
v05s.
A
A
A
So
we've
talked
about
these
two
artifacts
in
the
mesh
reci
meeting,
so
we've
been
sort
of
restarted
that
meeting
once
every
couple
of
weeks,
and
so
I
believe
both
of
these
docs
have
really
yet
to
be
rewritten,
but
as
we
graduate
to
a
v050,
it's
about
that
time
that
we
start
to
get
particular
about
code
quality.
A
And
we
have
been
right
so
we've
brought
in
some
end-to-end
end-to-end
integration,
testing,
we're
spinning
up
kubernetes
we're
deploying
service
meshes
automatically
each
time
that
we're
doing
a
pr
we've
begun.
Some
of
us
have
begun
to
do
unit
tests
painfully
where,
but
but
it's
a
start,
we've
got
code
cov
plugged.
In
now,
we've
got
cypress
ui
plugged
in
actually
we're
do
we're
taking
stride
we're
taking
steps
giving
it's
by
the
way.
A
Just
as
a
side
note,
it's
important
that
we're
taking
these
steps
and
one
because
when
we
do
get
to
a
1.0
that
you
know,
is
you
like
it
or
not?
That's
sort
of
the
general
signal
that
you
know
a
piece
of
software
is
production
ready
and
so
we'll
need
to
sort
of
harden
our
practices
and
improve
our
coverage
as
we
go.
We've
begun
to
have
different
priorities
of
github
issues,
so
I
think
there's
kind
of
three
priorities
right
now
and
they
probably
need
to
be
updated.
A
Eventually,
there's
a
low
priority,
high
priority
and
urgent,
but
we
may
need
one
that
is
identified
as
like
a
blocker
or
we
may
need
to
sort
of
structure
it
like
priority.
One.
Two
three
four
potentially
and
associate
what
those
mean
a
p1
or
p0.
What
have
you
meaning
it's
a
blocker.
You
can't
release
without
it.
We
also
probably
wouldn't
want
to
release
with
any
p1,
we'll
have
to
define
these
and
that's
kind
of
what
we're
looking
for.
A
What
I'm
looking
for
between
the
master
test
strategy
and
the
build
and
release
plan
is
what
the
what's
the
gating
criteria
for
release,
so
rudolfo,
rodolfo
martinez
vegas,
is
leading
the
ci
calls,
and
I
don't
know
that
he's
able
to
join
this
call
today,
but
for
any
of
you
who
speak
to
him
next,
if
you
would
convey
the
notion
that
part
of
what
we're
looking
for
here
is
a
gating
criteria
for
release.
A
Some
of
the
by
the
way
for
those
of
you
who
have
been
around
the
block
or
or
those
of
you
that
haven't
one
way
to
structure
gating
criteria
like
this
is
in
terms
of
like
how
do
you
define
and
be
comfortable
and
confident
with
code
quality?
Well,
one
of
the
ways
is
to
is
to
say
that
not
only
would
you
say
something
statically
like
oh,
there
are
zero
p
zeros,
there's
zero
blockers.
A
What
you
would
also
say
is
something
that's
a
little
more
dynamic,
you'd
say
for
the
last
two
weeks,
there's
been
a
certain
trailing
percent
like
the
percent
of
open,
p1s
p2.
You
know
these
open
bugs
has
been
decreasing
by
a
certain
percentage
which
basically
just
says
like
it
doesn't
matter
what
their
priority
is.
Just
you
can
characterize
the
quality
of
the
code
overall
as
being
as
only
improving,
because
the
the
rate
of
defects
being
opened
or
the
rate
of
certain
class
of
defects
is
just
trailing
off
and
so
food
for
thought.
A
As
we
go
to
write
those.
If
anyone
has
the
links
to
these
two
docs,
I'm
tossing
the
links,
if
you
would
because,
because
frankly,
some
of
what
I
just
said
is
golden,
so
we
should
put
that
into
the
doc.
A
Okay,
anyway,
we've
got
a
couple
of
other
topics.
I'm
compelled
to
pull
this
one
to
the
top,
though
in
part
just
because,
because
I
see
sako's
smiling
face
at
7
a.m,
so
so.
B
By
the
way,
some
of
you.
A
Yeah
saku,
some
of
you
might
know
sako
some
of
you,
many
of
you
may
not
sako
is
well.
I
don't
know
he's
like
he's
in
the
first
10
contributors
to
the
project,
he's
the
first
g
soccer
to
come
through
the
the
community
and
he's
off
and
has
well
geez.
I
don't
know
I
feel,
like
you've
been
through
off
to
greener
pastures
and
then
sort
of
off
to
even
greener
pastures
after
that,
and
so
I'm
around.
A
So
sako,
the
one
of
the
things
that
that
you'd
that
you
were
messaging
about
recently
is
like
deployments
of
measuring
and
yeah.
I
thought
I
thought
we'd
I'd
I'd
shut
my
mouth
for
a
minute
and
let
you
talk
a
little
bit
about
like
what
what
metabob
is,
how
measuring
might
be
helpful
or
yeah?
Definitely.
B
So
metabolis
was
was
one
of
my
schoolmate
had
worked
on
his
currently
show
and
we
have
another
ctlc
game
from
school.
So
it
was
part
of
neck
accelerator
where
some
these
researchers.
They
came
idea
that
how
we
can
do
ai
based
code
review,
it
started
with
python
so
using
the
pull
request,
doing
some
nlp
there
and
then
applying
it.
Once
you
have
the
code,
it's
automatically
you
can
sign
in
with.
Basically
it
will
find
the
mistakes
and
so
on
and
we'll
show
you
in
visualize
this
and
give
you
some
suggestions.
B
So
that's
what
it's
doing
at
the
end
and
as
a
moving
forward.
It's
going
promising
because
currently
it's
in
second
accelerator
and
doing
some
funding
with
with
antler
and
currently
in
accelerated
with
alchemist
so
going
through,
and
there
is
a
good
attraction
from
as
well
from
that
perspective,
so
it's
becoming
as
more
inverse
sas
and
I
was
helping
out
more
in
engineering
side
all
this
building
infrastructure
and
for
how
it
can
be
with
best
practice
on
and
so
and
bringing
in
the
devops
side
of
tooling
and
architecting
and
so
yeah.
B
A
And
look
at
that,
we
you
well,
I
hate
to
I
hate
to
give
us
too
much
credit,
but
I
would
say
to
the
extent
that
that
it's
been
a
while
since
you've
been
on
the
call,
but
you
show
up
and
then
all
and
then
just
have
something
to
demo.
That's
that's
how
that's
how
we
do
it?
That's
great!
That's!
Let
me
do
it
so.
B
Yeah,
I
will
have
also
a
couple
comments
regarding
from
devas
devops
perspective
and
yeah.
So
first
thing
it's
it's
in
behind
vpn
we
use
aws
vpn
here
and
if
you're
gonna
go
it's
in.
B
If
you
don't
know
octa
is
a
centralized
tool
where
you
add
all
the
applications
you
have
for
passing
everything
and
and
so
on,
and
I
decided
it
was
a
good
point
for
measuring
to
be
there.
So
again,
I
have
to
work
in
oadc
octa
support,
also
idc,
but
there
is
also
would
be
nice.
I
don't
know
if
there
is
any
plan
for
support
saml,
it's
usually
enterprise
feature,
but
yeah.
B
It's
one
of
the
things
I'm
I
would
say
that
either
it's
an
enterprise
future
cannot
be
if
there
is
any
plan
or
it
can
be
else
added
in
open
source.
There
is
a
way
to
do
that
so,
but
for
now,
once
you
connect
to
vpn
and
you
gonna
just
come
through
and
you
you
see
because
I
logged
in,
but
if
you
haven't
logged
in
it,
should
it's
bringing
you
here
and
the
other
things
I
will
say.
B
B
B
A
B
Yeah,
let's
look
and
reset
so
yeah.
I
have
done
a
couple
things
and
I
think
it
was
a
good
because
every
time
once
there
is
a
qa
manual
tester,
so
they
can
do
before
the
release
and
then
we
can
save
it
because
we
it's
also
you-
you
have
here
istio
as
a
service
mesh
and
would
be
nice
to
have
so
just
here
and
there
I
was
getting.
I
was
just
impressed
with
them
with
in
general,
the
progress
which
is
it
was
a
while
also
and
wasn't
following,
and
it's
really
amazing
jump
to
community
too.
A
Hey
be,
please
be
critical,
please
do
like,
don't
don't
don't
sugarcoat
it
like
give
it
give
it.
B
Yeah,
I
definitely
think
from
real
life
built
perspective.
It's
so
far
going
good.
It's
just
would
be
nice
for
the
group.
I
don't
know
if
it's
support
already
there'll
be
like
group
feature
like
because
I
would
like
to
share
this
result
can
be
some
other
folks.
Let's
say
qa
can
do
and
then
we
can
see.
That's
I'm
not
sure
if
it's
already
there
that
or.
A
It's
not,
but
it
totally
needs
to
be
the
sam,
the
yeah,
the
saml
is
precisely
what
you
would
identify
it,
as
is
like
that's
a
perfect
thing
for
the
providers
to
to
implement
and
to
support
is
like
that.
A
You
see
this
part
like
the
ability
to
share
that
or
not
only
share
the
results,
but
the
ability
also
to
share
your
test
configuration
because
you
know
maybe
you
put
in
certain
parameters
like
you,
ran
it
for
five
minutes
and
you
hammered
you
hammered
it
in
this
way.
The
ability
for
you
to
as
an
individual
user
to
save
that
profile
and
recall
it.
So
you
don't
have
to
type
it
all
back
in
would
be
nice
two.
The
ability
then
to
share
both
the
results
and
the
test
profile
with
another
user
yeah.
A
B
Yeah,
but
that's
enough
at
this
point,
we
see
enough
value
like
for
before
the
release.
You
do
you
see
exact
time.
What
was
the
parameters
and
what
are
the
test
result
and
you
can
evaluate
whatever
feature
was
added
and
how
the
performance,
how
we
are
doing
with
performance,
so
that's
already
today,
that's
having
saving
that
result
particular
helps
a
lot
just
to
for
that.
B
Another
nice
thing
would
be
have
like
make
sure
you
see.
I
like
run
through
the
the
ui
which,
which
is
also
let's
say,
involved.
Hashicorp
vault
case
is
having
that
one,
and
I
think
it's
pretty
much
the
ui
side
that
it
could
be.
I
liked
also
in
general,
the
world
itself.
They
have
a
good
role
based
access
and
all
this
stuff,
that's
from
my
ideas.
There
would
be
similar,
which
is
already
secure
and
enterprise
kind
of
level,
which
is
even
so
not
yet
but
yeah.
B
That
could
be
one
thing
having
cli
here
and
sometimes
I
think
yeah,
that's!
Oh.
I
also
liked
the
evaluation
based
on
the
based
on,
I
think
nighthawk.
It's
I
haven't
used
that
until
it's
added
new
yeah.
A
Yeah,
it
is,
and
I'm
not
sure
that
it's
it'll
work
for
you
either.
Okay,
which
means
we
probably
need
a
great
yeah.
It
is
new.
Nighthawk
is
a
sub-project
of
the
envoy
project,
so
nighthawk
is
so
so
what's
funny
is
sako
knows
this
area
pretty
well,
the
nighthawk
is
written
in
c,
plus,
plus
sort
of
like
wrk2,
and
so
because
mastery
is
in
go
like
you
know,
wrapping
some
support
around
it
and
stuff.
A
A
Yeah
yeah,
the
one
of
the
things
is
like
to
to
do
well
to
help
people
answer.
Questions
like
hey
when
I
want
to
understand
how
to
best
configure
my
service
mesh
should
I
I've
got
a
certain
requirements,
certain
resiliency
requirements,
for
example-
and
so
you
know
maybe
I
should
go
to
each
of
my
services
and
have
them
have
set
like
two
or
five
or
whatever
number
of
retries
between
each
service.
So
well,
what's
the
optimal
number?
A
Well,
just
like
just
like
answering:
what's
the
performance
of
your
mesh,
it's
like
it
depends
as
we
can.
Nighthawk
will
facilitate
some
optimization
routines
so
that
people
can
run
a
series
of
of
tests
that,
where
the
load
itself
will
be
able
to
put
in
a
controller
that
will
do
an
adaptive
load,
it's
an
adaptive
load
controller,
so
it'll
run
a
test
depending
upon
what
comes
back
from
that
test.
A
Change
the
config
run
another
one
and
have
this
process
automated
such
that
we're
engaging
with
a
couple
well
at
least
one,
hopefully,
two
universities
to
study
to
answer
those
questions.
To
answer
to
find
the
points
of
inflection
for
optimization
around.
B
A
We'll
probably
end
up
like
to
the
extent
that
we
end
up
going
that
deep
with
nighthawk.
What
we
may
end
up
doing
is
at
some
point
it's
less
important
to
people
that
they're
literally
understanding
and
choosing
between
these
three,
and
it
becomes
more
about
the
feature
that
they
get
like.
Did
they
want
to
do
a
long-running
adaptive
load
test?
Okay,
if
they
choose
that,
then
underneath
will
use
nighthawk
if
they
just
wanted
to
run.
You
know
depending
upon
what
they
wanted
to
do
by
the
way.
Sako.
A
And
by
the
way,
you're
all
welcome
to
so
austin.
If
you
don't
mind,
if
you're
able
to
post
a
link
to
the
figma
project
for
the
mastery
ui
people
might
be
able
to
take
a
glance
at
as
a
matter
of
fact,
sako
is
a
great
person
to
take
a
glance
at
the
proposed
ux
for
for
the
performance
profile,
not
the
performance
test
that
he's
looking
at
now,
but
the
profile
the
ability
to
save
your
config
and
recall
it
one
of
the
things
that
I'm
I
don't
know
that
it.
A
It
often
strikes
people
as
being
inability
soccer.
If
you
go
to
the
results
page
and
there's
two
things
that
I
think
are
often
overlooked,
and
it's
it's
our
fault,
not
the
user's
fault.
So
so
one
of
those
is
the
ellipse.
Under
the
the
details
column,
it's
like
the
very
right-most
column
details,
and
it
has
this
horizontal
three
dots,
the
horizontal
line.
A
A
A
A
Without
this
context,
yeah,
it
totally
is
it's
very
important
because,
especially
as
you
go
to
share,
if
you're
sitting
here
and
you're
like
well,
you
want
to
go
share.
These
results
with
whomever
it
becomes
significantly
important
to
the
extent
that
you
ran
yours
in
a
particular
environment
and
yeah
yeah.
B
That's
also
another
question.
I
am
not.
Okay,
probably
it
was
done
old,
one
old
test:
okay,
okay,
yeah
yeah,
that's
the
one
yeah,
that's
the
ones!
I
love
mail,
which
is
saved
from
all
time,
testing
this
gk
cluster.
That's
why
I
was
confused,
so
yeah.
B
Yeah,
that's
amazing
that
it's
all
saved
here.
So
that's.
If
you
remember
for
conference,
we
were
doing
without
mesh
and
vismesh.
So
that's
and
yeah.
I
I
don't
know
much
time.
I
couple
things
also
I
would
like
so.
The
fun
thing
would
be
more
also
interesting
for
game.
Four
will
be
ci
cd
part.
It
did,
for
example,
physique
action
if
it
supports
today
already
or
it's
planning
to
support
once
get
access,
it
can
go
and
run
that.
A
Yeah,
it
there's
a
I
so
there's
a
a
couple
of
projects
that
we
want
to
base
around
interactions
with
github
they
haven't
come
forth,
they've
been
described.
I've
asked
a
couple
of
times
for
volunteers
in
the
community
and
we've.
It
hasn't
really
struck
a
chord
with
anyone
just
yet
one
of
the
things
that
you
can
immediately
do
to
them
today,
with
github
actions
is
to
either
use
the
rest
api
to
like
programmatically,
invoke
a
performance
test
or
to
programmatically
provision
a
surface
mesh.
A
As
a
matter
of
fact,
I
mean
it
was
last
thursday
that
one
of
the
contributors
was
demo,
demonstrating
the
fact
that
it
has
the
ability
to
provision
a
mesh
run,
a
test
as
a
matter
of
fact,
it
has
the
ability
to
provision
a
mesh
and
do
a
a
canary
roll
out.
At
the
same
time,
we
that
isn't
that
isn't
released.
B
Just
yet
but
yeah,
I
I
see
yeah.
That
will
be
it's
as
much
so
because
it's
already
installed
in
cluster
it
would
be
nice
to
somehow
integrate
from
protection.
It
can
go
and
run
the
one
in
class
not
bring
every
time
they
measure
yeah.
That's
right.
A
B
Yeah
and
other
thing
I
liked
about
the
validation
regarding
the
best
practices
of,
but
I
think
I
was
sure
if
I
was
doing
right
where.
A
Yeah
there
was
two
there's
kind
of
two
on
for
all
for
almost
all
the
adapters
you
can
verify
whether
or
not
they
are
compliant
with
service
mesh
interface
and
that's
good
that
that's
that
conformance
thing
for
istio.
Specifically,
if
you
go
to
that
adapter
it
that's
weird
that
you
only
if
you
oh
yeah,
I
think,
there's
a
little
bit
of
a
problem
with
that
adapter.
It
looks
like
it's
not
running
at
the
moment.
It's
not
connected.
B
So
yeah
there
is
measure
or
is
some
nsm.
It
will
be
if
it's
more
also.
I
think
the
description
in
in
in
that
sense
that
if
the
first
step
I'm
coming,
then
what
would
which
one
you
should
select
and
yeah,
and
then
I
am
just
curious
for
when
this
mesh
configuration,
as
would
be
interesting
like
to
see
again
again,
that's
a.
B
I
see
the
value
here
regarding
that
to
do
conformance
test
so
yeah,
that's
one
of
the
things
would
be
interesting
to
see
the
result
and
if
there's
any
changes
required-
and
I
will
assume
you
can
apply
automatically-
that
changes
or
how's
that
yeah.
A
That's
good
yeah
to
to
date,
and
this
is
a
little
bit
well.
This
is
a
little
bit
almost
harder
to
describe
because
it
because
the
istio
adapter
in
this
environment
doesn't
seem
to
seem
to
be
having
problems,
but
there's
kind
of
two
validations.
There's
the
one
that
you're
looking
at
for
service
mesh
interface
and
and
it's
yeah
you're
right
like
you'd,
want
to
just
on.
If
you're
you
want
to
ongoing,
confirm
that
istio
continues
to
adhere
to
smi's
specs.
A
So
that's
good,
there's
a
second
set
of
validations
and
it's
really
about
best
practices
like
you've
got
your
istio
environment
misconfigured.
It
today
measuring,
doesn't
have
an
auto
remediation
like
it'll,
just
point
out
that
you're
kind
of
doing
some
things
correctly
or
incorrectly.
A
B
Yeah,
I
would
be
interested
just
to
follow
up
a
bit
also
check
more
on
on
that
yeah
something
valuable
moving
forward
as
well.
I
totally
agree.
I
totally
agree
just
last
few
comments
regarding
there
was
the
help,
because
I
did
it
through
the
helm,
so
I
think,
as
a
just
in
simple
way,
I
needed
to
download
the
measure
itself
to
so
I
think
there
is
a
way
like
to
be
able
to
use.
B
It
is
a
git
helm
package,
so
that
would
be
nice
to
add
some
index
and
not
instead
of
downloading
just
pull
from
the
helm
chart
that
and
in
the
once
it
was
coming
and
port
forwarding.
I
think
we
can
just
just
get
rid
of
the
doing
for
post
level
because
you
can
kill
pod
and
so
on.
It
will
be
nice.
I
mean
just
add
it
simply.
B
A
Yeah,
these
are
great
notes.
It's
it's.
We
need
to
immediately
go
remove
the
fact.
The
suggestion
to
clone
the
repo
is
not
necessary.
There.
There
is
a
helm
chart,
it
is
published
in
helm,
hub
or
artifact.
What
is
it
artifact
hub
or
whatever
that?
Oh.
A
It's
a
pattern
also
yeah
the
like
get
the
application.
Oh
right,
those
instructions
are
coming
out
of
the
out
of
the
helm.
The
helm
chart
itself
right,
which
that's.
A
B
In
this
case
you
have
to
get
the
pod
name,
you
usually
and
then
adding
performance
here
to
the
port,
but
because
also
in
port
forward
case,
it
support
service
name,
which
is
doesn't
change
which
is
measured
by
default,
and
it
can
just
do
yeah
this
service
measure
directly
to
the
service
level
port
forwarding.
B
A
B
A
B
Sure
I've
gone
close.
A
B
A
Yeah,
this
is
all
great.
This
is
all
great
feedback.
It
becomes
very
painfully
obvious
that
the
majority
of
people
who
are
using
and
developing
on
meshri
are
either
a
just
compiling
and
running
locally
or
b.
I'm
deploying
to
docker
and
much
much
less
of
the
time.
Are
those
users
deploying
to
kubernetes
because
yeah
the
that
flow
is
that
flow
should
be
well
curated
and
it's
not?
The
thing
is,
is
most
of
the
work
has
been
done.
It's
mostly
about
documentation,
just
cleaning
it
up,
making
sure
it's
it's
smooth,
but
that
yeah.
B
I
just
can
see
like
going
to
webs
and
every
time
I
see
it's
really
well
documented.
It's
really
ui
and
everything.
I
keep
me
impressing
and
I
follow
also
in
the
slack
the
documentation
and
design
salt
going
for
there.
That's
all
amazing,
it's
just
the
feedback,
it's
purely
from,
let's
say
in
enterprise
level
requirement.
B
If
you
are
coming
without
small
engineering
background.
Let's
say
if
you
put
somebody
in
create
him
result
who
doesn't
understand
queens
and
what
we
need
couples
seeing
manual
testing
to
do
that.
That's
the
requirements
coming
for
there,
so,
but
definitely
otherwise.
All
the
stuff
from
engineering
perspective
totally
makes
sense
and
really
amazing
job,
and
I
definitely
see
the
value
and
it's
easy
to
sell,
also
in
in
special
ways
startups
coming
we're
moving
fast.
The
releases.
A
It's
getting
faster,
yeah,
the
the
there's,
a
bunch
of
people
on
the
call
who
are
being
quiet,
but
it's
getting
faster
because
of
them.
Yeah
it's
been
there's
I'm
two
other
things
to
point
out
to
you.
One
is,
I
know
it's
funny,
it's
funny
that
I'm
gonna
point
this
out,
but
it's
actually
because
I
know
that
you
already
know
that
this
is
a
capability.
A
B
That
that
one
I
I
did
and
from
ui
perspective,
I
was
looking
so
and
it
isn't
impressive
that
yeah,
because
I
think
you
can
select
all
of
them.
Yeah,
that's
and
compare
it
yeah,
that's
yeah!
B
It's
just
a
bit
more
for
ui
perspective,
just
to
understand
which
one
is
going
to
what
yeah,
but
because
I
think
there
is
a
different
colors,
so
yeah
that
that's
all,
because
there
are,
for
example,
three
test
cases
I
selected.
I
would
assume
it's
how
how
are
we
gonna
be
for
different?
Okay,
that's
the
one
yeah
that
that
would.
A
A
Hell,
I
don't
know,
I
can't
read
this
thing.
I
I
don't
know
what
it's
saying
yeah.
That's
it
that's
a
joke,
but
but
but
yeah,
no
it.
It
is
a
challenge
to
present
so
much
info
all
in
one
thing,
and
it
clearly
needs
some
love
and
some
help
it
we
shouldn't
have
to
like
we
can
go
put
up
documentation.
That
explains
how
to
interpret
it,
but
that's
the
wrong
answer.
A
The
right
answer
is
fix
the
user
experience
so
that
it's
a
more
intuitive
and
self-ex
self-explanatory
and,
to
the
extent
that
you
can't
get
it
super
intuitive,
then
provide
helpers,
provide
a
small
guide
like
a
small
tour,
maybe
the
first
time
that
someone
goes
to
use
it.
So
you
can
say,
look
you've
got
just
like
sako
just
said:
you
have
three
tests
that
you're
comparing
so
there's
these
tick
marks
on
the
x
axis
on
the
horizontal
axis,
the
first
segment
that
those
those
one
two
three
four
you
know
six
or
seven
dots.
A
Those
are
all
from
your
first
test
in
the
middle
column,
from
your
third
test
and,
and
then
the
color
coding
is
like
is
the
the
percentile.
A
Like
I,
I
see
I
see
that
you're
saying
that
if
you
go
to
that
teal
color,
the
qps
on
the
third
yeah
on
the
third
test,
like
I
got
a
higher
queries
per
second
on
the
highest
on
that
one,
okay
and
but
what
does
that
mean
like?
Is
that
good
or
or
is
that
bad
or
where's?
A
More
than
that
saku
it's
been
our
fantasy,
or
at
least
my
fantasy,
and
maybe
the
goals
of
some
others
is
that
so
by
the
way
for
everyone
on
the
call.
I
hope
that
this
feedback
from
sako
is
like
water
on
parched
soil,
and
I
suspect
many
of
us
just
don't
get
enough
user
feedback.
So
I
just
this
is
soccer.
I
really
appreciate
this.
It.
A
It
you're
shining
a
light
on
some
things
that
are
long
overdue,
one
of
the
not
only
the
ux,
the
ui,
the
interpretation,
the
are
those
important
but
there's
been
a
long-standing
goal.
We
have
open
to
us
use
of
the
the
cncf's
cluster,
their
cluster
of
physical
servers,
and
I
know
sako
knows
this,
but
so
for
the
rest
of
us.
We
can
go
now
that
measures
like
in
a
in
a
much
better
place
than
it
was
a
year
ago.
A
It's
open
to
us
to
go
over
and
run
a
series
of
tests
and
use
that
hardware
to
produce
some
additional
insights.
Some
insights
that
we
can
come
back
and
and
expose
to
people
here
and
help
them
and
to
be
able
to
say
the
the
fact
that
sako's
test
results
from
a
year
ago
are
still
present
and
available
to
him.
A
So
on
that,
so
I
can't
help
it
but
the
the
soccer,
if
you,
if
you
visit
the
website,
smp
hyphen
spec
dot,
io
smp
I've
inspected
io,
it
is
service,
mesh
performance
spec.
A
B
Yeah,
I'm
gonna
have
a
more
deeper
look
and
I
think
one
more
thing
would
be
nice.
I
don't
know
it
would
be
support
visit
slack
integration
once
you
have
the
if
there
is
any
scheduled
way
of
running
that,
let's
say
week
friday,
it
runs
and
pushes
to
select
as
a
notification,
or
it
could
be
also
kind
of
alerting
and
so
on.
Again,
it's
more
about
moving
forward
stuff.
I
am
talking
that
would
be
interesting
to
which
is
one
of
the
most
all
the
companies
will
use
slack
so
would
be
if
there
is
an
integration.
A
To
run
to
initiate
a
performance
test
from
slack
too,
like
yeah
or
to
just
show
you
the
graph
of
the
results
like
in
slack
yeah,
like
that's
one
of
those
things
under
your
user
yeah
that
yeah
we
should.
Let
me
just
stop
and
say:
yes,.
B
Yeah
and
one
one
more
thing
like
I
think
it
can,
it
has
potential
to
the
pleasures
like
something
like
status,
page
yeah,
which
is
pinging.
If
the
english,
whatever
is
down
it,
could
also
be
used
because
it's
inside
posted
in
cluster,
it
can,
I
don't
know,
run
test
every
one
minute
and
10
50
minutes,
because
it
has
a
potential
already
all
in
place
and
once
if
it's
down
and
so
on,
it
can
also
be
there.
So
it
could
be
performance
plus
health
check
functionality.
It
can
be.
A
A
Yeah,
haven't
you
having
a
having
lead
technology
strategy
at
solarwinds
at
a
monitoring
company
it?
It
doesn't
that
that
fact
doesn't
escape
me
that
yeah,
the
observability
and
monitoring
is
like
they're.
A
Our
some
of
the
contributors
have
been
asking
around
roadmap,
and
just
not
even
some
of
the
particulars
just
kind
of
you
know,
vision
and
like
what
what
are
some
of
those
goals
when
we
just
talked
about
some
of
them
it
for
my
part
and
I'm
just
one
voice
of
others,
there's
a
number
of
other
maintainers
on
this
call,
and
for
my
part,
it
scares
me
to
step
it
like.
That
is
part
of
the
goal.
That's
absolutely
part
of
the
goal.
A
A
It
scares
me
a
bit
that
the
actually
there
was
a
demo
last
thursday
from
from
utkarsh
one
of
the
contributors
where
he
really
opened
up
the
ability
to
take
a
service
mesh
pattern
and
which,
which
is
just
a
simple
yaml
file,
very
simple.
It
describes
the
mesh
and
its
config
that
you
want
to
provision
the
like
the
performance
test.
You
want
to
run
the
in
this
case.
A
It
isn't
isn't
too
far
of
a
step
away,
and
I
won't
name
specific
other
projects,
but
very
large
projects
that
so
but
yeah
you're
right,
like
hey,
you're
there
and
you're
doing
these
things
anyway,
yep,
of
course
like
you
should
it
should
be.
It
should
do
the
things
that
you
just
said:
the
github
integration,
the
slack
integration,
the
saml
stuff.
Yes,
the
helm,
stuff
just
needs
to
be
cleared
up
and
it's
kind
of
silly
that
it
hasn't
been.
There
was
one
other
that
he
said.
B
Yeah,
I
think
in
general,
just
the
scheduled
one
and
learning
one
that
once
getting
integration,
that's,
but
as
of
today,
just
without
any
this
comment
that
is
already
enough
value.
It's
insoluble,
let's
say
point
it's
stable,
robust
and
it's
it's
worked
perfectly
and.
A
I
think
yeah,
it's
it's
about
time
to
it's
about
ready
to
go
into
the
cncf
as
a
matter
of
fact
like
in
part,
because
you'll
you'll
see
one
other
thing
that
veneth
who's
on
the
call
and
there's
some
others
that
are
working
on
this
as
well.
A
project
called
mesh
map.
Actually,
I
think
we
talked
about
the
visual
topology
has
is
also
behind
the
scenes
yet
to
be
exposed
it
will.
It
will
be
a
chiali-like
in
nature,
all
right,
so
one
which
will
be
really
helpful.
A
One
thing
that
you
just
touched
on
about
scheduling
is
is
very
much
so
a
need.
As
a
matter
of
fact,
a
workflow
engine
in
general
that
has
a
scheduler
is
very
much
needed
as
a
component
of
measuring
there's
a
bit
of
an
analysis,
that's
begun
about
which
workflow
engine
to
use
and
to
integrate
with
so
to
in
part
to
answer
adina's
question
about
canary
releases.
The
demo
that
was
done
last
week
was
with
argo
cd,
although
to
the
user,
they
didn't
even
know
that
argo
cd
was
being
used.
A
It
was
just
a
simple
pattern
file
you
apply
it
and,
by
the
way,
there'll
be
a
demo
of
that
at
the
cncf
sig
network
tomorrow,
1
pm
central
we'll
do
a
demo
of
that
that
whole
flow-
and
this
is
an
integration
with
oam,
with
open
application
model,
so
mastery
and
oem
mesh
is
becoming
oem
compatible.
I
know
that
some
of
this
stuff
is
new
on
people,
but
we
just
don't
have
that
much
time
to
explain
at
all.
A
So
where
there's
an
open
analysis,
that's
begun
about
what
workflow
engine
to
integrate
with
in
temesri
and
ideally
to
embed
and
treat
as
a
library.
To
me,
this
one
looks
pretty
hot
and
I'm
fairly
like
what
we
fairly
much
like
what
we
would
want.
Argo
cd
is
sort
of
the
second
most
as
a
sub
capability
of
what
argo
cd
as
a
general
workflow
engine
is
capable
of.
Flagger
is
a
specific
canary,
rollout
thing.
Flagger
great
will
integrate
or
what
have
you,
but
it
will
dismiss
that
as
like.
A
That's
not
a
full-blown
workflow
engine
and
there's
all
kinds
of
workflow
that
we
want
to
facilitate
like
a
slack
notification
or
like,
like
a
schedule.
Hey,
I
want
to
schedule
this
test,
there's
also
a
policy
engine
that
we're
in
need
of
so
the
founder.
A
So
I
have
a
long
history
with
opa
and
styrofolks
and
helping
them
go
into
the
cncf
and
the
founder
of
of
caverno
just
reached
out
to
me
a
couple
of
days
ago
wanting
to
catch
up
he's
a
good
friend.
He
used
to
deliver
my
product
at
cisco,
and
so
anyway,
it's
a
we
need
to
evaluate,
but
we
need
both
flow
engine
and
a
policy
engine.
That's
actually
one
thing
I
want
to
talk
to
everybody
about
today
was
we're
just
having
a
discussion
with
potential,
so
the
oh
adina
is
this
workflow
engine?
A
A
So
this
call
didn't
quite
go
as
I
was
expecting
there's
a
couple
of
things
to
talk
about
the
the
feedback
from
sacco
is
very
much
so
important.
I
think
it's
very
important
for
everyone
here
to
hear
that
directly
from
well,
both
from
a
contributor
and
a
user
like
sacco,
the
the
last
thing
I
wanted
to
there's
two
two
things
so
I
I
asked
I
caught
made
a
call
for
volunteers
on
get
nighthawk
last
night
and
I've
yet
to
be
able
to
go
through
all
the
messages
in
slack.
A
So
if
I'll
be
finding
I'll
be
finding,
if,
if
someone
didn't
already
volunteer
we'll
we'll,
I
will
find
you.
A
A
Outcast
systems
is
a
low
code
offering
so
they
I
think
this
probably
tells
you
all
that
you
would
need
to
know
is
that
they
have
a
designer
a
low
code
designer
you
have
activities
you
drag
and
drop
them
in.
You
build
out
a
workflow
of
like
doing
I
don't
know
twitter,
a
sentiment,
analysis
or
whatever
you're
doing.
I
don't
know
they're
they're,
empowering
I.t
with
I
don't
know
ticket
create.
A
You
know
whatever
flows
that
they
have,
but
we
spent
a
long
time
with
them
talking
about
how
they
are
thinking
very
well
toward
measuring
and
some
of
the
use
cases
that
they
have
and
and
a
lot
of
their
use
and
what
we
had
shown
them
on.
The
call
is
something
that
a
number
of
you
have
seen.
A
But
that
we
haven't
spent
a
lot
of
time
on
and
we
seem
to
perpetually
lose
time
to
to
talk
about,
and
I
wanted
to
make
sure
that
this
slide
was.
A
Was
shared
and
anirud
are
two
people
who
come
to
mind
that
have
asked
about
sort
of
vision
and
road
map,
and
you
know
forthcoming
things,
and
where
can
I
optimize
my
time?
How
can
I
help
what
do
we?
You
know
like
what
you
know.
A
What,
where
are
we,
where
are
we
sort
of
generally
headed
part
of
the
answer,
is
for
those
individuals
to
inform
that
and
to
you
know
affect
that,
and
but
from
a
very
from
a
very
high
level,
like
hey
here's,
the
perspective
of
what
that
playing
field
looks
like
or
what
a
full
a
fully
capable
service
mesh
management
plane
might
look,
like,
I
say,
fully
capable,
maybe
maybe
there's
additional
things.
A
Are-
and
this
slide
doesn't
quite
build
out
very
well.
The
core
element
here
is
like
we
said
in
the
past.
Mesh
sync
is
extraordinarily
important
to
the
project.
It's
it's
the
operator
measuring
operator.
That's
deploys
synchronizes
information
from
kubernetes
and
is
the
lifeblood
or
pumps
in
a
bunch
of
info
about.
A
A
Slack
github,
you
know,
like
integrations
sure,
a
schedule
sure
to
basically
be
invo
those
workflows
to
be
invoked
and
be
called
from
policies
that
have
been
evaluated
so
data
from
mesh
think
or
other
that
are
being
reasoned
over
and
policies
being
evaluated
depending
upon
the
evaluation
of
that
policy
that
a
given
workflow
might
kick
off
or
not
like
you
know,
could
or
could
or
couldn't
depends
on.
A
What's
going
on
that,
some
of
what
we've
talked
about
with
get
with
nighthawk
and
the
a
bit
of
machine
learning
that
we're
looking
to
do
there
can
have
an
effect
on
policy
or
some
of
those
policies
can
be
machine,
learned
things
meshrey's
ability
to
this
whole
thing
is
measury
right,
but
so
we
shouldn't
use
the
logo
here,
but
just
measures
ability
to
do
multi-mesh
life
cycle
and
configure
management
stuff
is
part
of
that
flow
because
it
might
be
part
of
the
flow
might
be
to
go
affect
the
config
of
a
mesh
part
of
the
flow
might
be
able
to
go,
insert
a
filter
into
a
mesh
or
into
a
site.
A
Awesome,
a
lot
of
things
observability
and
also
the
ability
to
take
action
to
like
invoke
a
performance
test
or
invoke
a
conformance
test
or
potentially
to
at
some
point.
There
will
also
be
a
whole
set
of
designer
features
that,
where
you'll
be
able
to
design
your
topology
and
then
just
tell
mastery
to
to
realize
that
or
you
know
to
go,
make
it
so
when
it
does.
Rather
than
encouraging
ad
hoc
administration
of
your
infrastructure,
we'll
integrate
with
github
for
change
management
to
integrate
with
a
git
ops
based
deployment
flow.
A
A
B
A
We're
just
about
out
of
time.
I
know
there
were
a
couple
of
things
to
call
out
so
navindu
is
here
and
has
been
iterating
over
designs
to
enhance
free
ctl,
to
support
deployments
to
whether
you
want
to
use
helm
to
deploy
mastery
or
just
measuring
ctl,
to
deploy
measuring
into
kubernetes
there's
a
set
of
designs
and
a
number
of
open
issues
for
anyone.
Who's
interested
in
working
on
mastery
ctl,
please
see.
Novendu
anirud
has
been
focused
there
heavily
as
well
see
novendo.
A
C
C
A
Yeah
good
one
is
that
good,
there's,
there's
more
there's
more
to
do
on
meshmap
and
so
having
you
having
you
and
and
or
others
engaged
beauty.
That
makes
me
chuckle
makes
me
happy.
The
second
reason
why
I
chuckle
vineeth.
Do
you
want
to
talk
about?
Why
I'm
chuckling
about
that.
D
Yeah,
we
have
been
wondering
about
that
since
a
few
months
right
now
so.
C
Yeah,
I
actually
looked
at
that.
The
main
like
cytoscape,
has
functionality
for
adding
animations
on
edges
and
the
density
can
also
be
decreased
on
increase
based
on
the
code,
but
the
javascript
library
does
not
have
something
like
that,
like
I
couldn't
find
anything
which
was
inside
escape
chase,
but
maybe
we
can
still
do
it
if
we
get
the
notes
and
the
edges.
C
D
Yes,
the
js
library
doesn't
have
these
features.
What
cytoscape
has
so
that's
what
we
are.
We
are
thinking
about
how
to
include
how
to
add
traffic
capability
capabilities
into
the
I
mean
animation,
capabilities
into
the
cytoscape
js
library.
A
B
Yeah,
it's
a
atm
and
it
has
been
a
while
thanks
so
much
for
having
me
and
I
will
continue
because
more
invest
and
from
user
perspective
just
move
forward.
I
will
come
be
happy
to
share
the
results
and
we'll
also
follow
up
on
a
couple
stuff,
and
that's
definitely
nice
to
see
mrs
ctl.
The
documentation
and
posting.
A
Those
are
the
kindest
compliments,
because
sako
knew
it
knows
what
it
was
like
when
it
was.
You
know,
barely
barely
anything
so.
B
A
A
I
mean
the
thing
is:
is
like
it's
not
it's
the
reason.
This
is
the
last
thing
I'll
say
and
I'll
shut
up
and
it's
it
really
needs
to,
because
so
many
people
have
given
so
much
work
to
mastery
to
smp
to
get
nighthawk
to
all
the
projects
and
for
anyone
who
comes
over
to
evaluate
yeah,
it's
possible
that
they
wouldn't
hit
the
layer,
5
io
site,
it's
possible.
B
B
I
don't
have
any
comments
on
that
regard,
that
there
wasn't
any
confusion
because
sometimes
presenting
too
much
information.
It's
also
not
good
from
user.
It's
becoming
hard
to
follow,
but
I
I
haven't
felt
going
to
website
and
plus
in
general.
That's
also
good
comment
for
the
product,
I
think
just
from
value
perspective,
because
there
is
parameters
and
so
on
just
showing
a
lot
of
stuff.
That's
also
from
something
it's
not
good.
Sometimes
you
just
have
to
find
the
right
balance
to
to
make
it
more
actionable
for
the
user.
The
ones
value
generates.