►
From YouTube: Meshery Development Meeting (July 21st, 2021)
Description
Meshery Development Meeting - July 21st, 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
Good
evening,
everybody
or
good
morning,
depending
on
where
you're
from
this
is
the
measuring
development
meeting
today
is
21st
of
july,
and
we
have
a
packed
list
of
topics
to
discuss
today
and
before
we
move
on
to
the
topics
and
the
announcements,
we
have
a
tradition
of
letting
newcomers
introduce
themselves
to
the
community.
A
B
A
Are
there
anyone?
Oh.
A
A
Okay,
what
have
you
been
working
on
like
what
tech
stack.
A
Okay
sounds
good:
when
did
you
join
learfing?
A
I
mean
yesterday,
okay,
sounds
good,
welcome
shivam,
and
we
look
forward
to
see
your
work
here.
A
Thank
you
for
that.
So
you
are
interested
interested
in
going
and
layer.
5
has
a
lot
of
golang
based
projects
and
there
are
excellent.
A
D
A
D
A
Thank
you
for
that
gravish
and
look
forward
to
seeing
your
work.
Oh.
E
E
Yeah
so
hello,
everyone
I
am,
and
I'm
pretty
new
to
this
community
just
and
I
also
attended
the
meeting
on
monday.
So
I
find
it
very
insightful
that
how
this
community
is
being
organizing
this
weekly
meetings-
and
this
is
like
really
new
to
me-
and
I'm
glad
that
I
have
joined
this
community
and
yeah.
I
have
been
learning
golang
and
right
now.
E
I
am
you
know,
starting
starting
to
learn
about
micro
services
like
how
we
can
make
micro
services
in
goliath,
because
I
find
it
very
useful
for
scaling
a
product.
I
think
so
right
now,
I'm
working
on
my
skills
and
that's
all
thank.
A
Thank
you
so
much
for
that.
Intro
yeah
sounds
like
you.
You
will
fit
right
in
here
and
is
there
anyone
else
who
would
like
to
introduce
samsung.
A
So
for
the
newcomers
we
do
have
a
newcomers
meeting
that
that's
on
every
thursday.
A
So
if
you
head
on
to
meet
dot,
clarify
dot,
io
or
and
to
this
link,
you
will
find
all
the
all
the
meetings
in
the
life
community,
so
thursday's
newcomer
call,
is
a
good
place
to
get
started
and
get
oriented
with
all
the
different
projects
in
life
and
finding
areas
to
work
on
or
help
or
get
help
to
troubleshoot
any
troubleshoot.
Any
issues
you
are
having
like
setting
up
the
development
environment
a
lot,
so
that's
a
good
place
to
jump
in.
A
So
if
there
is
not
anyone
else
to
introduce,
we
will
move
on
to
the
announcements
for
today.
So
anita,
would
you
like
to
start.
G
Hello,
everyone
yo
welcome
to
this
meeting,
so
I
probably
didn't
drop
this
since,
like
it
spits
my
mind,
but
I
remember
like
into
this
meeting,
we
set
out
a
chat,
a
topic
on
the
discourse
that
for
those
persons
who
have
topics
to
bring
up
during
our
weekly
general
meetings.
G
So
if
you
have
a
topic
regarding
anything
pertaining
in
this
community
or
you
have
a
suggestion
to
make-
or
you
have
something
you
want
to
share
with
us
in
general
yeah,
so
you
can
just
drop
your
topic
under
this
topic
in
the
discuss
forum
and
them
I'll
include
it
in
the
minutes
for
this
week's
meeting
and
then
subsequently
to
when
the
meetings
come
up.
So
just
comment
under
the
topic
and
I'll
speak
a
topic
immediately,.
H
A
Thank
you
anita,
so
anita
was
describing
about
the
perfect
community
meeting
which
happens
on
every
friday.
So
if
you
have
topics
you
can
submit
it
to
the
meeting
minutes
and
you'll
find
all
those
in
the
in
the
link
I
shared
about
the
mid.latified
at
io
link.
A
Next
up
we
have
abhishek.
Would
you
like
to
talk
about
the
the
new
release.
A
F
Yeah
sure,
so
what
release
are
we
on
so
we're
on?
Are
we
on
37?
Are
we
as
of
this
morning?
Are
we
38
yeah
just
off
half
an
hour
ago?
We
are
38.,
okay
boy.
I
can't
keep
up
with
all
that
you
you
all
are
doing
so
great.
So
just
for
quick
familiarity,
there's
mesherie
is
well,
it's
not
a
1.0,
yet
it
has
functionality
that
people
are
using
in
production.
It
has
functionality
around
performance
management
in
particular
that
has
been
around
for
a
long
time.
F
Some
of
that
is
tried
and
true
there.
It
has
a
few
architectural
components
that
need
to
be
added
between
v0.5
and
1.0,
and
you
can
see
some
of
those
in
the
roadmap.
The
roadmap
is
listed
in
a
couple
of
places.
One
is
in
the
layer.
Five
community
drive,
there's
a
google
doc
under
the
mastery
folder
called
roadmap
and
there's
also
roadmap.md
in
the
github
repo
mesh,
remeshery
repo
and
the
community.
This
project
has
been
it
has
there's
a
there's,
a
lot
to
understand
about
how
releases
are
done.
F
F
So
right
now
you
can
do
some
amount
of
filter
management
or
pattern
management,
even
application
management
in
the
current
release,
even
though
those
things
are
road
mapped
for
v06
or
v07,
architecturally
they're
planned
for
in
advance
and
aspects
of
them
sometimes
are
worked
on
it
in
advance
and
so
yeah
just
some
thoughts
about
the
releases
in
general.
The
particulars
of
this
release
that
just
went
out
there's
well
bendy.
The
thing
that
compelled
the
the
release
for
the
v0538
was
what
again.
A
Yeah,
so
we
apple
has
released
a
new
m1
chips,
so
measuring
ctl
or
mesh
current
the
current
version.
The
current
version
of
measuring,
could
not
be
installed
on
the
on
the
m1
chips.
So
what
this
release
does
is
it?
It
builds
a
new
binary
for
the
apple
silicon
chips,
so
you
can
and
install
it
on
apple
silicon.
I
just
tested
it
now
and
it
it
is
working.
A
So
yeah
5.38
is
support
for
apple
silicon.
A
Aditya
is
also
not
on
the
call.
So
to
sum
up,
what
aditya's
topic
is
like
aditya
has
been
working
on,
adding
support
for
not
not
adding
support
like
he
has
been
testing
a
mastery
on
raspberry
pi.
So
that's
a
valid
use
case
for
running
missouri
and
service
meshes
and
kubernetes,
and
he
has
some
updates
on
that
and
he'll
be
demoing
that
on
the
on
the
fry
on
friday's
community
car
all
right
now,
let's
move
on
to
the
topics
for
today's
and
first
up
we
have
scientin
so
scientist.
I
So,
actually,
I'm
going
to
talk
about
this,
but
yeah
we
are
actually
me
and
rudolph
were
working
on
this,
so
we
are
trying
to
generate
the
json
schema
and
the
definition
for
this
adapter
component.
I
These
are
the
issue
issues
related
to
it.
I
have
so
I
have
solved
this
osm
component,
not
fully
like
one
part
is
left,
but
I
can
show
you
this
and
rudraks
has
also
helped
me
doing
this.
I
Actually,
what
I,
what
we
are
doing
is
that
we
are
creating
a
github
workflow
for
this
for
generating
those
json
schema.
Now,
if
you
want
to
see
the
workflow,
so
actually
this
is
the
workflow
I
have
written.
First,
we
are
going
to
take
the
manifest
files.
I
I
I
And
then
we
have
generated
the
workflow
over
here
now,
after
going
through
this
workflow,
we
actually
got
the
schemas
over
in
the
template
directory.
In
this
workload
we
got
the
all
the
schemas
now
the
task
that
I
was
talking
that
was
remaining.
Actually,
these
are
the
resources
for
the
osm
components.
I
So
actually
we
have
to
do
like
http
route
group,
osm
dot
and
all
these
things
so
that
we
differentiate
between
which
resource
are
for
which
message,
so
that
task
is
left
for
now
for
this
osm
and
if
you
want
to
see
the
workflow
for
this
like
how
did
it
work?
Actually,
I
can
show
you
that
also.
I
I
I
F
F
But
I'll
drop
some
links,
maybe
into
chat
so
that
others
can
digest
what's
happening
with
patterns,
and
what
do
we
mean?
There's
is,
and
so
this
is
great.
This
is
great
because
thus
far
it's
really
been
the
issue
adapter
that
has
become
patternized
or
or
supports
patterns
it.
The
work
that's
been
done,
it
looks
like
I
think
other
than
maybe
are
you
in
the
process
of
name
spacing
the
components
or
did
you
already
name
space
them?
F
I
mean
it
sounds
like
other
than
that,
like
it's
pretty,
you
know,
high
quality
work
good.
You
know
good
good
job.
One
thing,
though,
that
does
that
concerns
me
a
little
bit
is
that
we
would
then
we
might
end
up
with
multiple
ways
to
generate
those
configurations
and
by
that
actually
a
few
different
ways.
So
so
one
is
that
currently
the
istio
adapter
and
the
osm
adapter
are
taking
the
approach
of
using
that
service.
Mesh's
ctl,
you
know
it's.
F
It's
command
line,
client
to
generate
the
manifests,
and
if
you
take
that
effort
which
is
unique
to
each
service
mesh
times
eight
more
times,
you
know
times
10
conceptually.
It's
that's
a
lot
of
repeating
ourselves
of
not
in
a
uniform
way
and
it
becomes.
It
becomes
a
lot
to
maintain
and
sustain,
because
inevitably
those
for
as
much
as
people
like
to
think
of
a
cli
as
an
api
or
like
as
a
contract
that
you
can
script
around,
they
aren't
they're
meant
primarily
for
humans.
F
People
do
script
around
clis,
just
like
we
were
just
looking
at,
but
those
contracts
are
not
upheld
nearly
as
tightly
as
something
that's
designed
to
be
an
api
and
is
versioned,
and
so
for
that
reason,
for
the
reason
that
we
don't
want
to
do
ideally
not
things
that
are
one-off
and
specific
to
each
each
mesh
that
it
would
be
probably
advantageous
if
we're
using
a
uniform
approach
to
extracting
json
schema
from
each
type
of
mesh.
F
So
think
about
that
for
a
moment,
while
I
also
say
that
the
manner
in
which
so
outside
of
whether
or
not
we're
using
the
cli
of
each
individual
mesh
or
whether
or
not
we're
pulling
the
manifest
directly,
irrespective
of
that
the
operation
to
do
it
to
perform
that
task,
I
wonder
if
that
shouldn't
be
an
adapter
operation
that
is
invoked
through
a
workflow.
So
what
I
mean
to
say
is
that
it's
we
will
find
use
cases
where
the
adapter
itself
needs
to
perform
that
task
potentially
real
time.
F
Why
would
that
be?
Well,
maybe
someone's
been
running
mystery
for
a
couple
of
weeks,
maybe
a
month,
maybe
a
couple
of
months
eventually
there's
a
new
version
of
open
service
mesh
that
comes
out
with
that
new
version.
There's
new
manifest
those
manifests,
have
either
changed
or
there's
a
new
resource.
So
how
does
the
running
adapter?
F
You
know
go
about
supporting
that
new
version
of
open
service
mesh?
Oh
well,
one
answer
could
be
well,
you
upgrade
mescheri
and
and
its
version
so
that
it
can
support
the
new
version
of
open
service
mesh.
Okay,
that
could
be
an
answer
like
you
have
to
go
out
and
maybe
not
upgrade
all
of
measuring,
but
just
upgrade
the
meshery
open
service
mesh
adapter.
F
Okay,
that's
reasonable
people.
Do
that
kind
of
a
thing
with
software
today,
if
you,
if
you
want
for
your
some
software
to
be
compatible
with
some
other
software,
sometimes
you
have
to
upgrade
one
or
the
other
or
both
all
right,
although
if
the
operation,
the
workflow,
that
we
have
that's
a
build
time,
concern
a
build
time
workflow
if
that's
made
built
into
a
runtime
workflow
in
measuring
I'm
sorry
in
the
adapter,
the
open
service,
mesh,
adapter
or
all
of
the
adapters
one.
F
We
begin
to
have
uniform
capability
across
all
of
them,
both
at
build
time
and
runtime
that
you
can.
Then,
if
we're
interfacing
with
the
kubernetes
manifest
themselves,
the
golang
that
is
written
into
the
adapters
becomes
well
potentially
entirely
uniform
right
right.
We're
dealing
with
manifests
uniformly.
We
extract
the
json
uniformly.
A
workflow
does
run
and
does
do
exactly
what
scienton
you've
just
you've
just
shown,
but
the
the
utilities
or
the
but
sustaining
that
effort
becomes
a
sustaining
becomes
easier
and
the
ability
to
perform
that
operation.
F
K
What
did
you
think
actually.
C
There
would
be
so
definitely
this
would
be
the
better
approach,
but
there
would
be
definitely
like
it
would
be
quite
difficult.
The
reason
is
that,
obviously
we
can't
fetch
a
manifest
on
the
fly.
So
if
a
doctor
takes
the
responsibility
of
doing
these
things
on
its
own,
but
the
primary
difficulty
is
that
conversion
of
which
is
into
reason
schema
version.
Four.
This
is
my
track.
Four.
C
The
second
difficulty
is
that,
if
there
are
not
that
I,
I
think
this
would
be
a
very
rare
situation
that
you're
running
two
services
same
time,
so
basically
figuring
out
what
so
there
are
two
situations.
One
is
when
you
are
actually
grabbing
the
minifigures
from
running
cluster,
basically
getting
crd,
so
that
every
flow
looks
very
similar
in
that
case
figuring
out
that
which
custom
resource
actually
belonged
to
which
service
mesh
or
does
it
even
belong
to
the
service
mentioned
that
would
be
private
equity.
C
F
In
terms
of
like
ascertain,
you
know,
discerning,
which
crds
are
relevant
to
which
service
mesh.
It
should
be
the
case
that
each
of
those
crds
are
named
like
one
that's
the
responsibility
of
the
individual
service
mesh
adapter,
and
so
that
that
adapter
should
be
looking
for
what
I
assume
would
be
name
space
crds
that
you
know
come
from
the
control
plane.
F
It
might
also
be
the
case
that
it,
the
adapters,
don't
need
to
interrogate
kubernetes
and
actively
running
or
actively
described,
custom
resource
definitions,
but
rather
the
same
bundle
that
it
uses
or
the
same
way
in
which
it
retrieves
the
installation
files
I
mean
it
could
use.
It
could
use
that
mechanism
as
well,
which
would
then
for
sure
provide
clarity
over
which
resources
are
assigned
to
what
service
mesh
so
sciatica.
Mostly
it's
a
this
is
a
fantastic
work.
F
We
should
go
use
this.
We
should
pers.
You
know,
you
know
finish
off
the
last
little
bit
of
of
the
feedback
that
you've
gotten
and
and
persist
those
patterns,
those
components
and
then,
as
we
as
as
we
go
forth
and
you
stare
down,
you
give
a
cold
stare
to
doing
this.
F
Eight
more
times
and
then
sustaining
it
over
time,
let's
think
about
some
of
the
the
feedback,
the
the
approach
that
I'm
describing
right
now
and
let's
see,
if
that
that
makes
sense,
the
one
thing
that
would
crush
the
other
thing
that
we
crush
is
pointing
out
is
that
if
this
is
built
into
the
runtime
of
the
adapter
right
now,
the
adapter
will
need
to
make
a
system
we'll
need
to
incorporate
the
the.
What
do
we
call
it?
F
L
Also,
sorry,
for
interrupting
just
a
little
the
point
that
did
you
figure
out
that
versioning
things
sent
them?
Did
you
update
your
workflow
according
to
that
like
service,
mesh
name
and
service
resource
name.
I
J
A
So,
let's
move
on
to
the
next
item
so
drew
jupil
or
rudraksha,
but
any
of
you
like
to
give
an
update
on
the
smi
confirmant
section.
A
L
Okay,
so
we
have
discovered
what's
happening,
it's
the
event,
streamer
not
receiving
the
changes
and
not
reflecting
it
on
the
ui
group
is
currently
working
on
fixing
it
hopefully
we'll
be
able
to
fix
it
and
if
not,
I'm
also
working
on
a
little
hack.
M
Good
and
including
that
yeah
we
also
have
glee.
We
would
like
a
review
on
those
updated
google
slides.
We
created
a
flow
of
how
smi
is
working
for
the
call
itself.
One
slide
left
is
basically
the
ci
action
call.
If
I'm
not
wrong,
I
don't
necessarily
think
we
need
that,
but
it's
up
to
you
where
we
want
it.
We
can
finish
that
and
yeah.
That's.
M
A
F
Yeah
yeah,
just
you
know,
I'm
always
anxious.
I
think
about
the
fact
that
there's
a
lot
of
there's
a
lot
of
nice
there's
a
lot
of
people
on
the
call,
many
of
which
I've
had
the
pleasure
of
talking
to
and
some
of
which
I
haven't
yet.
F
I
think
I
think,
a
lot
of
which
kind
of
like
who
are
kind
of
like
me
and
looking
at
stuff
like
this
and
going
that
diagram
looks
pretty
neat
I
what
the
heck
is
going
on,
though
oh
my
gosh,
that's
a
lot
of
lines:
measury,
adapter
library
and
mesh
kit.
What's
mesh
kit,
if,
like
I,
I
imagine,
a
lot
of
us
have
a
ton
of
questions,
so
here's
I'll
make
a
promise
to
all
of
you
and
samir
may
yeah.
F
F
If,
within
a
week
you
don't
have
an
answer,
somebody
hasn't
responded,
somebody
didn't
explain,
didn't
answer
your
question,
then
we're
gonna
have
to
figure
out.
What's
I
don't
know
if
we
have
to
figure
out
what
happens?
I
guess
what
I'm
saying
is.
There's
a
lot
going
on
in
here
and
management
software
is
can
be
well
can
be
quite
stupid
if
it's
done
wrong
and
it's
pretty
easy
to
understand.
F
What's
going
on,
if
it's
fairly
stupid,
we're
really
trying
for
measuring
to
it
like
this
thing's
gonna
manage
a
service
mesh,
that's
emerging
tech,
it's
confusing
as
how,
if
it's
trying
to
make
a
service
mesh
even
easier
to
understand,
that's
a
really
tough
job,
and
if
it's
going
to
do
that
times,
10.
F
F
I'm
hopeful
that
in
a
year's
time
or
less
there's
a
lot
of
people
asking
that
question,
like
literally
googling
what
the
heck
is
a
service
mesh
pattern,
and
then
I
would
it
would
be
really
neat
if
any
and
all
of
you
that
are
on
the
call
right
now,
like
could
field
that
question
could
say:
oh
yeah,
it
like
conceptually.
It's
like
this,
like
that,
you
were
comfortable.
You
felt
comfortable
enough
to
casually
describe
it.
That
would
be
awesome.
F
It
would
be
awesome
for
you
mostly.
It
would
be
awesome.
It
would
be
great
for
the
project
as
well,
but
people
are
trying
to
figure
out
what
are
the
conventions
by
which
you
you?
What
does
a
surface
mesh?
Do?
How
should
I
use
it?
How
do
I
know
if
I'm
using
it?
Well,
that's
what
these
patterns
describe
and
what
we're
trying
to
do
is
take
those
architectural
patterns.
The
way
that
you
say,
if
you
want
to
use
a
service
mesh
to
do
retries,
you
know
retry
a
failed
request.
F
Here
are
the
considerations
that
you
should
contemplate
as
you
configure
your
infrastructure
as
you
configure
your,
what
do
you
call
it
now?
Samir's
got
me
all
distracted
as
you
configure
your
service
mesh
good.
Well,
that's
a
lot
of
things
to
consider
like
again
like
how
do
I
know
if
I'm
doing
it
right
well,
here
here's
a
pattern:
here's
a
sample
pattern.
F
F
The
system
today
that
will
take
those
patterns
and
realize
them
affect
the
infrastructure
like
orchestrate
the
infrastructure
to
apply
them
is,
is
measury
and
that's
why
we're
talking
a
lot
about
patterns
and-
and
so
that's,
why
it'll
also
be
really
powerful
that
people
can
say
well
yeah.
I
feel
good
about
how
we've
configured
retries
in
my
infrastructure.
F
I
didn't
configure
400
of
them,
because
I
know
that
that
might
our
user
might
wait
too
long
if
they're
having
to
sit
and
wait
for
400,
failed
requests
to
retry,
and
I
didn't
configure
zero
because
that
isn't
resilient
enough,
and
so
I
found
the
sweet
spot
and
that's
important,
because
I
started
with
this
pattern,
and
maybe
I
augmented
it-
maybe
I
reasoned
over
it.
But
but
you
know
what
I
can
apply
this
pattern
to
any
mesh
that
I'm
using
and
wow
okay.
This
is
this
is
good
management.
Software
like
mastery
helps
me
do
this.
F
F
Sorry,
I
totally
I'm
sorry,
I
totally
digressed
from
that's
my
performance,
which
is
smi
service.
Mesh
interface
is
an
is
a
specification
and
we're
going
to
be
there's
a
community
meeting
we'll
be
hosting
it
today
in
about
three
hours.
Jubril,
dhruv
and
rudraksh
are
going
to
go
present
on
that
meeting.
They're
going
to
go,
show
off
their
good
work
from
what
they've
been
doing
in
this
community.
F
Part
of
that
is
that
if
you
visit
the
blog
on
this
website,
it
talks
about
how
mesherie
is
the
conformance
tool
that
verifies
the
specification,
smi
specifications,
and
so
these
three
outstanding
gentlemen
are
going
to
go.
Show
this
blog
in
action
and
go
show.
How
mescheri
does
this
validation
and
they're
going
to
explain
the
charts
that
they
were
just
showing
so
good
yep
sounds
like
the
gist
of
it.
Gentlemen.
F
Is
that
you've
got
you've,
got
a
bug
or
two
you've
got
a
hack
in
place
to
you
know
to
to
show
people
that
there
there's
a
pipeline-able
way
to
integrate.
Smi
conformance
into
your
your
ghetto
back,
you
can
buy
a
github
action
and
there's
also
a
way
to
just
run
these
manually
and
what
the
goal
today
isn't
to
give
a
conference
type
talk.
The
goal
today
is
like
get
to
the
point
and
say:
hey:
look.
We
spent
a
lot
of
time
building
out.
You
all
spent
a
lot
of
time.
F
F
You
need
to
empower
them
with
the
right
amount
of
knowledge
to
go,
deploy
the
github
action
or
go
do
run.
These
things
manually
understand
they're,
going
to
start
manually,
which
is
great
because
that
works.
It
should
work
and
that'll.
Give
you
a
little
bit
of
time
to
fix
any
remaining
issues
in
the
action.
A
A
So
so
I
guess
we
will
all
join
today's
smi
meeting
and
see
this
in
action.
Yeah
and
level
is
sharing
the
details
of
the
meeting.
So
moving
on
to
the
next
item,
we
have
abhishek.
Yes,
sheikh,
are
you
on
the
call
yeah?
I
am
on
the
call,
so
I
am
not
ready
with
this
operations
list
yet
so
I
won't
be
able
to
present
this
now.
F
H
Provided
okay,
so
basically
there
will
be
a
resync
operation
which.
A
Which
lets
you
rediscover,
or
basically
sync
again
with
the
kubernetes
cluster.
F
Okay,
good
I'll
take
notes.
I
guess,
as
you
go
like
you,
have
this
written
down
anywhere
the
work
that
you're
doing.
F
Resync,
which
is
a
is
a
database
purge
and
reinitialization
of
of
mesh
sync,
basically
keep
going.
A
A
A
A
F
F
F
F
J
A
F
F
There
is
already
a
connectivity
test
for
the
mesh
re
operator.
If
I
recall
yeah,
there
is
already
one
all
right,
very
good,
okay,
let
me
I'll
I'll
try
to
I'll
iterate
on
this
list
with
you
as
and
let
the
rest
of
the
meeting
go
forth.
F
I
know
nitish?
I
know
you'll
be
using
some
of
these
soon.
I
know
those
that
worked
on
mesh
sync
have
separate
operations,
I
believe
for
deploying
measuring
operator,
but
you
know
think
about
this.
Those
that
are
working
on
measuring
receipt.
F
F
Well,
a
pre-flight
check
might
or
not
pre-flight,
but
a
health
check
rather
might
be
an
invocation
of
this
connectivity
test
and
the
same
thing
could
be
said
for
all
of
the
adapters,
so
anyone
who's
used
mesherie
in
the
ui
you've
seen
those
material
ui
calls
on
these
adapter
chips.
You
can
click
on
the
chip
and
it'll.
Do
a
quick
like
ping
test,
it'll
say:
oh
mesh:
we
can
talk
to
kubernetes
measure,
it
can
talk
to
its
adapters.
You
can
click
the
button
a
whole
bunch
of
times
in
a
row
and
get
these
these
notifications.
F
A
A
A
M
N
Okay,
so
I
hope
you
guys
are
able
to
see
the
screen
so
actually
recently,
I
was
working
on
a
new
feature
for
mystery
called
connection
wizard
and
if
you
want
to
get
the
context
of
what
it
is
about
and
what
strength,
what
it's
trying
to
do,
you
can
go
to
the
issue
that's
2434
with,
but
actually
like
the
shape
that
it's
trying
to
take
right.
Now,
it's
a
little
bit
different
from
what
is
being
described
here,
but
still
it
should
give
you
a
context
of
what
it
is
trying
to
do.
B
N
So
yeah,
so
the
main
idea
of
this
is
that
you
should
be
having
like
for
the
first
time
user.
There
should
be
a
way
for
him
to.
We
should
be
actually
navigating
him
to
configure
the
whole
setup
of
mystery
so
that
he
doesn't
have
to
go
to
the
settings
page,
and
you
know
it
might.
He
might
be
lost
here
or
there.
So
that
is
something
that
we
are
trying
to
achieve
with
the
configuration
I
mean
connection
with
that.
N
So
actually,
like
one
point
that
I
want
to
discuss
beforehand,
is
that
the
name
was
actually
before
it
was
named
as
configuration
wizard
in
the
sense
that
we
are
trying
to
configure
things
in
the
misery.
But
right
now
it
has
been
renamed
to
connection
wizard
because,
like
when
you
think
about
this
whole
of
what
it's
actually
trying
to
do,
it
is
actually
trying
to
establish
a
connection
with
the
messaging
components.
It
is
not.
It
does
not
have
the
capability
to
provision
and
de-provision
something.
N
N
I
mean
to
set
up
the
connection,
so
in
the
first
step
we
have
a
connection
to
the
kubernetes
cluster,
so
first
of
all
like
when
we,
when
the
user
comes
to
lands
on
this
page,
like
the
ui
will
check
whether
the
cluster
has
been
configured
or
not,
and
it
will
try
to
ping
the
kubernetes
cluster
and
if
it
succeeds
like
the
screen
that
is
shown
right
here
will
be
displayed
to
the
user
and
it
would
actually
get
the
all
the
information
from
the
message
server
and
it
will
display
over
here,
like
you
can
see
the
current
context
and
kubernetes
version
like
this
is
actually
similar
to
what
we
have
in
the
settings
page
right
now
and
yeah.
N
If
you
want
like,
we
can
actually
disclose
this
connection
and
we
can
upload
a
new
config
file
when,
if
the
user
wants
to
do
so,
so
the
one
thing
that
we
should
keep
in
mind
is
that
it's
not
like
it
is
only
deep.
It's
not
it's
only
making
the
connection
or
breaking
the
connection,
it's
not
provisioning
or
de-provisioning
something.
So
that
is
something
we
have
to
keep
in
mind
and
yeah.
You
could
actually
pin
like
these
tips.
Are
we
see
everywhere?
N
N
So
moving
on
to
the
next
step,
we
have
measure
operator,
which
is
exactly
the
same
thing
like
it
tries
to
connect
to
the
machinery
operator
and
if
it
is
successful,
then
it
displays
this
page
and
if
it
is
not,
then
actually,
if
it
is
not,
then
it
kind
of
becomes
like
a
bug,
because
if
it
is
not
configured
then
like
from
what
I
understand
like
whenever
you
run
messages
here
system
start,
it
actually
does
all
these
things
in
the
background,
and
if
it
is
not
like,
if
the
message
operator
is
not
yet
running,
then
it
should
not
technically
open
the
ui
right,
like
I'm,
not
totally
sure
on
this,
but
that
is
something
that
is
not
yet
being
attended
to
right
now.
N
So
if
the
connection
is
not
yet
there,
then
we
have
to
think
about
what
we
should
actually
be
doing.
In
that
case
and
yeah
you
have
the
status
of
these
things
like
my
my
sink
and
nats
are
actually
not
connected
right
now,
so
they
are
in
disabled
state,
but
my
operator
is
connected,
so
when
I
try
to
ping
it
right
now,
there
is
an
issue
with
that.
I
guess
it's
because
since
missing
and
nats
are
not
being
connected,
those
are
not
connected
right.
N
So
when
I
try
to
bring
it,
the
response
won't
come
back
and
it
gets
stuck.
So
that
is
something
we
have
to
like
I'll
have
to
change
that
behavior.
N
So
yeah,
coming
to
the
service,
which
is
like
there
has
been
a
lot
of
thinking,
process
involved
in
this
thing,
because
when
you
think
about
service
mess
like
it
can
be
subtract
like
yeah,
so
basically,
what
we're
trying
to
implement
right
now
is
that
we
should
actually
see
like
we
will
ping
the
adapter
like
whether
the
adapters
are
available
or
not,
and
if
an
adapter
for
a
particular
service
mesh
is
available,
we'll
show
the
like
we'll
colorize.
N
This
thing
and
make
it
display,
as
it
is
connected
to
the
adapter,
so
basically
it's
it.
What
it
means
here
is
that
it's
a
connection
to
the
adapter,
whether
it
is
successful
or
not
and
yeah.
If
it
is
not,
then
you
can
provide
the
url
in
context
and
it's
basically
it's
basically
like
establishing
the
connection
to
the
adapter.
So
these
things
when
we
see
that,
like
you
know
like
we
would
notice
that
it
is
exactly
the
same
thing
as
we
are
having
in
the
settings
page
right
now.
N
But
the
thing
is
that
when
we
like,
when
misery
grows
right
like
at
that
time,
we
would
be
intending
to
have
different
things
in
the
setting
space,
such
as
like.
We
would
actually
be
trying
to
have
very
complicated
things
in
the
settings
page
rather
than
the
things
that
we're
having
right
now,
such
as
you
know,
managing
multiple
service
messages
and
all
those
kind
of.
C
N
Yep
so
eventually
like
the
functionalities
that
are
being
done
over
here,
like
will
be
transferred
to
like
this.
These
things
would
be
configured
in
the
connection
wizard,
so
this
would
become
something
of
a
different
nature,
yeah.
N
So
that's
how
that's,
where
it's
going
right
now
and
even
in
the
dashboard,
when
you
look
at
it
like
rather
than
like,
when
you
think
about
it,
like
the
dashboards,
would
actually
be
describing
about
what
the
what
measure
is
about
and
how,
how
it
is
managing
multiple
service
meshes
and
all
those
kind
of
things,
but
rather
it's
trying
to
show
the
connection
and
this
kind
of
stuff
which
in
the
future,
would
be
better
if
you
could
put
it
in
the
connection
wizard
like
have
us,
that's
what
it's
trying
to
assume
now
like
when
you
look
at
it
these.
N
N
Yep
and
the
service
mesh
ui
right
now,
like
it's
not
fully
complete
because,
like
we
should,
the
thought
was
that
we
should
actually
show
all
the
adapters
that
are
available.
I
mean
all
the
services
service
messages
that
maturity
machine
could
eventually
get
connected
to,
but
right
now
we
are
only
showing
three,
so
the
idea
is
to
you
know,
have
them
in
a
vertical,
corrosive
kind
of
thing
and
in
the
right
side
pane.
You
show
the
same
data
plane
that
we're
showing
over
here
and
it
would
actually
be.
N
N
That's
what
and
I
have
not
yet
implemented
it
because,
like
I,
was
working
on
these
two
things
getting
these
two
things
done
and
yes,
that
is
something
I'll
be
working
on
in
the
future,
and
the
same
thing
would
be
for
the
matrix
as
well,
and
one
more
challenge
with
the
matrix
thing
is
that
metrics,
like
graphing
and
promoters.
These
things
are
actually
in
the
context
of
a
particular
service
mesh.
So
in
those
cases
like
we
still
have,
we
are
not
yet
there
like.
N
We
are
not
sure
as
to
what
we
should
be
doing
for
that
case
and
yeah.
So
that's
where
we
are,
but
the
aim
right
now
is
to
have
the
have
the
same
thing
that
we
are
having
right
now
in
the
service
mix
like
we
have
a
carousel
over
there
and
this
right
side,
a
data
plane
that
shows
the
status
of
the
connection
for
graphene
and
promotes.
N
A
Yeah,
I
I'm
sure,
like
there
were
a
couple
of
questions
on
that,
like
what
would
someone
like
to
talk
about
like
oh
about
the
configuration
why
we
need
a
configuration
reset
in
the
first
place
like.
I
A
Yeah
like
so
the
thing
you
mentioned
on
the
chat
lee
like.
F
Yeah
yeah,
so
so
originally
so,
hey
here's,
here's
our
situation,
people
go
to
use
measuring
and
they
find
that
mesh
sync
isn't
connected,
nor
the
adapter
didn't
the
operator
didn't
deploy
or
that
they
can't
talk
to
an
adapter
or
that
they
can
having
trouble
getting
messages
to
talk
to
the
kubernetes
instance
that
they
wanted
to
talk
to
and
like
any
number
of
underlying
infrastructure
issues
get
in
the
way
of
all
of
that
bad
credentials.
F
Lack
of
network
connectivity
container
fell
over
blah
blah
blah
blah
blah
blah.
Just
like
goes
on
all
kinds
of
things.
Your
cube
api
isn't
exposed,
doesn't
even
isn't
even
reachable
like
you're
using
aks.
You
don't
have
the
right,
like
there's
all
kinds
of
things,
and
so
we're
trying
to
help
guarantee
that
p
and
so
like,
like
the
answer
can
be
today.
The
answer
is
great:
hey
you
know
these
things.
Are
you
can
work
through
them?
Some
of
them?
F
You
can
just
go
to
the
settings
area
and
there's
an
advanced
set
of
all
of
the
configuration
that
you
can
do.
You
can
do
from
here
good
and
to
the
layman
or
to
the
new
user.
It's
like.
Okay,
so
you
mean
I've
got
to
get
grafana
and
prometheus
connected
one
or
more
kubernetes
clusters
pointed
to
different
contexts.
F
What
if
I,
what
if
I
have
10
clusters-
and
I
have
five
contexts
or
I'm
sorry,
10
contacts
in
my
cube,
ctl,
my
cube
config,
but
I
actually
have
multiple
cube
configs
like
it's
going
to
go
on
and
get
pretty
hairy
and
great
measuring's
going
to
deal
with
those
things
measure
is
going
to.
Let
people
get
do
highly
configurable
deployments.
F
What
we're,
but
in
the
meantime,
people
who
are
just
you
know,
deploying
mini
cube
and
try
to
connect
mesherie
like
fumble
and
something
happens.
It's
like,
oh,
my
goodness,
and
then
they
reach
out
in
the
slack
and
they
say
hey.
Can
you
what's
going
on?
Can
you
help
and
it's
like?
F
Oh
goodness
boy,
maybe
it's
a
bug,
maybe
it's
not
maybe
it's
an
environment
thing
and
so
giving
them
tooling,
like
a
diagnostics
tool
to
say:
hey,
can
you
send
in
the
diagnostics
report,
we'll
take
a
quick
look
and
determine
briefly,
you
know
quickly
what
what
the
issue
is
great.
Okay,
that
will
be
helpful.
F
What
would
be
even
more
helpful
is
to
not
say:
can
you
run
a
diagnostics
report
that
members
of
the
community
can
help
you
with,
but
is
rather
to
say
you
know
what
here's
a
diagnostics
tool
for
you,
here's
an
opinionated
walkthrough
we
used
to.
We
were
and
by
the
way,
there's
a
whole
write
up
on
this
there's
a
whole
google
doc
and
are
a
couple
of
them
and
and
there's
mock-ups
and
other
things
on
what
was
being
referred
to
as
the
configuration
wizard
just
this
week.
F
This
was
been
renamed
from
configuration
wizard
to
connectivity
wizard,
because
more
of
the
configuration
and
the
advanced
configuration
will
be
done
in
the
settings
area.
So
nitish,
if
you,
you
might
have
already
gone
there.
But
can
you
go
to
settings
briefly
in
case,
and
so
what
the
connectivity
wizard
will
be
is
an
opinionated
guide
such
that
whenever
someone
says
I
can't
connect
to
my
kubernetes
or
I
can't
measure
the
mystery
operator
isn't
deployed,
or
it's
not
talking
to
me
like
great.
F
Can
you
go
to
the
activity?
Wizard
are
all
three
of
those
are
all
four
of
those
steps
lit
up
green?
Yes,
yes,
good,
you're
cool,
you
know
great,
then
like
there's,
no
question
you're
good,
no
one
of
them's,
red,
okay
and
then,
like
here's,
the
it's
supposed
to
facilitate
and
grease.
Do
we
facilitate
people
having
successful
initial
deployments
that
are
simple
in
nature
and
grease
those
wheels?
Eventually,
the
settings
area
is
going
to
go
quite
deep
with
the
concept
environments.
F
You
know
40
different
kubernetes
clusters
and
what
operators
are
running
on,
which
ones
and
blah
blah,
and
just
like
it's
going
to
go
on
for
a
while
you've
got,
and
this
is
I'm
not
making
up
these
numbers.
These
aren't
like
grossly
large
numbers
in
any
respect
like
a
40
cluster
environment.
Sure,
no,
no,
that's
like
just
not
very
big,
with
five
istio
service
meshes
or
10
across
the
40.
That's
also,
quite
not
that
large.
F
It
is
early
days
for
people
having
tons
of
service
meshes,
but
eventually
that
won't
be
the
case
like
there's
been
a
trend
for
people
to
downsize
the
size
of
their
clusters,
they're
in
increasing
the
number
of
clusters
that
they
have
also
they're
in
increasing
the
number
of
service
mesh
deployments
that
they
have
like
it's
coming
service,
mesh,
sprawl,
yeah,
it's
a
thing,
it'll
happen.
It's
coming
to
unify
them.
A
tool
like
measure
will
really
help
point.
F
Is
it's
going
to
get
complex
the
configuration
if
you've
ever
used
a
software
installer
where
you
install
the
software
you
go
next
next,
you
know
quick,
install
custom.
Install
the
settings
area
is
the
custom
install
custom
configuration
the
connectivity,
wizard
of
which
there'll
be
an
icon
in
the
nav
bar
that
probably
looks
like
merlin's
hat
or
a
magic
wand,
or
a
wand
with
a
gear,
a
gear
with
a
wand
over
top,
so
some
something
some
quick
wizard.
F
That
is
a
forced
step
through
in
a
sequential
manner
to
say
if
all
three
of
the
all
four
or
three
or
these
light
up
green
you're
good.
If
not,
here's
simple,
simplified
troubleshooting
steps
to
take
you
through
and
guess
what?
If
you
follow
this
next
next
next
in
the
connectivity,
wizard
you'll
have
a
successful.
If
you
go
into
the
settings
area
and
you
do
advanced
configuration,
you
might
just
shoot
yourself
in
the
foot
and
watch
out
at
least
over
here.
F
That's
not
going
to
be
the
case,
so
samir,
a
lot
of
that
was
probably
aimed
toward
you
and
if
that
doesn't
make
sense
or
if
that,
if
you
still
think
like.
Oh
hey,
that's
confusing
that
there's
a
simplified
view,
one
that
forces
an
opinionated
walkthrough
versus
what
I
described
for
settings.
We
should
definitely
talk
because
we
don't
want
to
make
it
confusing.
F
F
Yeah,
I
think
that
yeah.
Yes,
I
think
I
think
I
think
so
I
would
say
I
think
that
you
could
actually
say.
I
think
yes,
is
the
answer
to
that
question.
If
you
ask
that
question
of
anything
of
any
aspect
of
what
you're
configuring
would
it
be
easier
if
it
was
in
a
visual
topology
yeah,
it
probably
is
like
there's
a
couple
of
questions
that
having
a
tabular
format
is
actually
it's
probably
easier
like,
and
these
are
questions
like.
F
I
don't
know
when
you
visually
represent
it:
it's
pretty
intuitive
and
snaps
your
attention
to
it
right
away
as
well,
but
it
would
be
something
like
who's,
my
top
talker,
which
services
that
of
the
ones
that
I'm
running,
which
ones
are
receiving
the
most
traffic
like.
If
you
have
that
in
a
tabular
format,
you
can
just
sort
the
columns
and
see
oh
here's,
my
top
five
sources
of
traffic
or
destinations
for
traffic.
Great,
that's
very
okay!
That
works
well
for
a
tabular
thing.
F
A
Let
me
just
interrupt.
We
are
out
of
time
for
this
meeting,
so
I
like,
as
you
might
have
guessed,
we
are
still
going
to
go
a
bit
longer,
so
if
everyone
anyone
wants
to
leave
this
call,
please
feel
free
to
do
so
and
don't
be
compelled
to
stay
here.
So
thank
you.
Everyone
for
joining
I'll,
see
you
next
week
and
in
the
other
meetings.
That's
happening
this
week.
F
Great
no
back
to
the
there's
a
couple
of
other
folks.
I
think
that
we're
gonna
demo,
some
things
it'd
be
great
to
try
to
get
them
squeezed
in.
A
Sure
so,
thank
you,
nadesha,
like
really
looking
forward
to
have
this
released
and
yeah
so
moving
on.
We
have
utkar
utkar.
Would
you
like
to.
O
F
C
Okay,
so,
basically
for
the
context
recently,
there
was
a
requirement
that
we
noticed
so
is
to
facilitate
a
few
features
to
facilitate
a
few
infrastructure
to
say,
and
that
was
that
earlier
mystery
server
as
well
as
mission
cloud
used
to
assume
that
the
redirect
keyword
where
I
would
look
like
slash
api,
slash
user
slash
token.
So,
basically,.
G
C
Was
the
sort
of
part
coded
rpr
was
recently
raised?
I
actually
raised
that
via
which,
which
adds
this
particular
change
to
machine.
That
is
basically
now
you
can
and
you
can
provide
a
environmental
variable
to
mystery
server.
Why?
I
measure
ct,
of
course,
I'm
actually
demoing
it
on
mystery
server
directly,
because
the
image,
basically
this
vr
hasn't
been
moistured.
So
there
is
no
image
to
run
in
the
cluster,
so
I'm
demoing
it
on
the
mystery
server.
So
you
can
actually
given
a
custom
custom
callback
url
to
machine
server
and
it
will.
C
It
will
make
sure
that
it
passes
on
this
particular
redirect
url
callback,
you
order
it
to
the
machine
provider,
and
it
will
also
make
sure
that
it
at
least
basically
redirects
you
to
the
url
that
you
are
actually
passing
so
it
will.
It
should
actually
serve
the
use
case
that
we
had
earlier.
That
is,
let's
say
mystery
is
running
behind
a
proxy
and
you
are
redirecting
it
via
a
prefix.
So
you
have
something
like
this:
you
can.
Actually
you
can.
C
What
you
can
do
is
when
you
spin
up
mystery,
wirelessly
ctl,
you
can
use
the
same
environmental
variable.
Instead
of
running
make
a
meat
run
fast
dip.
This
is
actually
a
make
recipe.
What
you
can
do
is
you
can
run
mystery,
c2
system
start,
and
it
will
pass
on
this
particular
environment
available
to
the
manifest
or
the
docker
compose,
whatever
platform
you're
using
so
for
the
for
the
sake
of
demo.
Let's
do
it.
C
Okay,
now
I'm
going
to
start
my
mystery
server
and
I'm
actually
running
industry
cloud
also
locally,
because
obviously
machine
cloud
hasn't
like
the
pr
has
been
merged
on
the
machine
cloud.
Also,
so
that's
where
both
of
the
systems
are
running
locally
on.
C
So
now,
if
I
would
go
to
the
I'll
go
to
localhost
9.81,
it
redirected
me
to
this
particular
page,
because
actually
I
already
choose
the
provider
once
I'm
signing
in
with
github,
and
it
will
redirect
me
to
localhost
colon
8080
seo.
This
was
supposed
to
be
a
proxy
url,
and
that
is
basically
whatever
you
actually
provided
in
there
just
because
I
provided
local
hospital
on
8080
stu.
It
actually
directed
me
to
this
particular
url
and
actually
I'm
running
a
node
server,
which
is
actually
capturing
it.
C
And
this
is
basically
the
response
for
for
the
like
how
it
would
be
used
or
how
it
would
change
the
things
so,
for
example,
use
case
where
the
callback,
the
url,
is
not
exactly
who's
cool
9081
starts
something
we
can
actually
not
start
providing
a
custom
quality
model,
for
I
think
some.
You
also
have
a
like
a
similar
issue.
That
is,
if
the
machine
was
running
behind
a
proxy
and
it
was,
the
traffic
was
getting
redirected
based
on
a
traffic
circus.
C
So
now
what
can
happen
is
this
can
just
be
http
called
that
particular
endpoint.
C
One
caveat
is
that
you
will
have
to
make
sure
that
this
always
end
with,
unless
until
there
is
a
custom
redirection
that
happens
in
during
the
proxy,
which
can
definitely
happen.
So
if,
if,
if
it's
a
simple
prefix
based
proxy,
that
is
you
just
see
that
hey?
If
the
request
is
coming
to
slash
mystery,
slash
something
just
related
to
slash
aps
users,
and
then
that's
fine
other
than
that.
You
will
actually
have
to
basically
add
that
at
this
particular
suffix.
C
The
reason
is
that
mystery
actually
expects
the
end
point.
The
handler
actually
expects
the
token
to
be
received
at
this
particular
endpoint.
So
this
is
this
is
a
minor
caveat
that
you
can
control
the
entire
thing
except
the
suffix,
so
yeah,
that's
pretty
much.
C
A
C
C
No
not
yet
only
manifest,
and
so
basically
measuring
ctl
supports
it.
So
anything
that
mystery
ctl
does
you
can
do
it
from
there?
Him
is
something
that
yeah
I
have
to
either
I'll.
Add
it
in
the
same
pr
or
I'll
create
a
new
one.
So,
no
not
yet
only.
C
No,
not
the
mesh
point
was
actually
so.
What
you
should
be
able
to
do
is
just
run
this.
This
was
not
supposed
to
be
actually
a
user
specific
setting
or
something
so
yeah
yeah.
You
should
be
able
to
do
this,
and
this
particular
environmental
variable
would
be
actually
passed
on
to
the
docker
compose
or
the
manifest.
Whatever
platform
is
basically
the
context.
A
Yes,
nobody
has
questions,
so
you
don't
think
it's
really
good,
please
giving
the
thumbs
up.
F
F
F
F
Cool
anything
else.
A
B
So
so
the
issue
which
suhadi
and
me
are
working
so
as
as
you
already
explained
about
the
like
that
mystery
ctl
system
start
command,
it
should
not
overwrite
that
mystery.yaml
file,
but
but
we
have
few
queries
on
that
before
implementing
the
changes
basically
like
what
is
happening
is
con
when,
whenever
we
we
use
config.yml
and
it
overrides
in
the
master
dot
yaml,
so
that
is
expected
behavior.
B
But
as
you
mentioned
that,
if
power
users
want
to
change
in
the
docker
compose.yaml,
it
should
not
overwrite
the.
I
mean
the
changes,
okay.
So
the
the
question
here
is
like
what
are
those
changes
like?
How
do
we
start
implementing
this?
Because
in
the
conflict.yaml
we
have
like
few
parameters
which
gets
overwrite
in
the
mystery
file.
B
So
I
mean
there:
if
we
see
the
mystery.yaml,
we
have
ports
volumes.
I
mean
what
all
changes
we
are
expecting
that
which
which
power
users
can
change
and
yeah.
And
the
second
question
is
about
like
what,
if
config.ml
overrides,
suppose,
I'm
I'm
gonna
change
tag
in
the
mr.yaml
and
if
I
run
that
command
system
start
config.ml
will
overwrite
that
change.
So
is
it
allowed
or
not?.
B
Yeah,
so
the
mystery
ctl,
so
basically
concept.yaml
overrides
the
adapters
channel
and
version
parameters
right.
So
what
if
power
user
changes
that
value
directly
in
the
mainstream.cml
as
he
has
and
config.tml
will
overwrite?
That
is
it
allowed
or
not?.
A
Yeah,
I
think
so,
like
the
so,
let
me
try
yeah.
A
Yeah,
so
I
ideally
like
we
want
the
user
to
configure
the
whole
aspects
of
their
measuring
deployment
through
their
config.yml
file.
A
So
for
like
these,
these
should
be
enough
for
a
normal
user,
but
we
also
have
let
them
have
the
ability
to
modify
things
in
their
in
the
in
the
in
the
docker
compose
file
or
the
kubernetes
manifest.
So
this,
the
config.yaml
will
be
the
source
of
truth
for
everything
that
is
set
here
and
for
things
that
are
overwritten
here,
like.
G
A
If
the
you,
so,
what
measure
recital
does
right
now
is
to
rewrite
the
entire
docker
compose
file.
Basically,
it
just
downloads
the
docker
compose
file
from
github,
but
it
doesn't
persist
it
it.
Doesn't
it
basically
rewrites
the
entire
thing
from
scratch,
so
instead,
what
we
should
be
doing
is
like
rewrite
the
stuff
that
is
written
in
the
config.yaml
file,
but
we
can
persist
things
the
other
things
like
if
they
force
for
any
reason,
change
these
things.
Then.
A
A
A
Yeah
basically
read
in
the
config
for
the
docker
compose
file
into
a
struct
and
then
change
only
the
values
that
that
are
in
the
in
the
mesh
conflict.
The
config.tml
file.
B
Okay,
okay,
yeah,
fine
I'll
I'll.
We
will
work
on
this
november.
A
Okay
thanks:
anyone
else
has
comments
or
different
opinions
about.
A
F
I
think
one
thing
that
I'll
be
like
if
I
think
that
it's
a
it's
like
step
is
like
stepping
in
a
pile
of
poop
of
manure.
Like
it's
not
easy,
it's
not
exactly
what
we
want
to
spend
a
lot
of
time
in
is
allowing
people
to
have
their
cake
and
eat
it
too,
which
is
to
customize.
F
Certain
aspects
like
provide
endless
support
for
whatever
customizations
that
they're
going
to
want
to
put
into
their
kubernetes
manifests
and
into
their
docker
compose
files,
because
that's
a
lot
and
like
us,
like
measuring
ctl,
trying
to
take
all
that
on
and
internalize
it
and
guarantee
that
those
customizations
won't
be
overridden
is
difficult.
F
We
it
it
is
convenient
for
mastery
ctl
to
facilitate
the
deployments
to
deploy
itself
and
the
way
that
it
deploys
itself
is
by
using
platform-specific,
manifests
either
kubernetes
or
docker
the
manifest
that
we
that
mastery
itself
generates
people
do
want
to
customize.
That's
in
part,
why
people
like
helm
because
they
can
we're
customized,
they
use
values.yaml
to
provide
environments
or
customize
to
provide
environment
specific
configuration.
F
F
From
context
to
context,
and
then
and
then
all
of
you,
then
neha
and
and
all
of
you
will
be
in
a
really
good
position
to
to
be
more
fervent
in
your
opinionated
stance-
that
whatever
that
stance
is
like,
you
know
you,
some
users
want
to
deploy
on
a
certain
type
of
network.
So
in
this
manifest
this
compose
file
that
we're
looking
at
they
want
to
be
able
to
specify
network
equal
host
network
driver
equal
this.
They
want
to
configure
those
things.
F
It's
like
look,
that's
totally
legitimate,
like
yeah,
that
that's
that's
what's
required
in
their
environment.
Oh
my
goodness.
We
can't
consider,
for
we
can't
like
create
measuring
ctl
commands
for
all
those
things
or
so
how
do
we?
But
we
do
need
to
dictate
in
some
cases
like
if
one
of
the
use
cases,
if
someone
if
we
release
another
adapter
like
for
app
mesh
out
of
aws,
for
example-
and
someone
goes
to
upgrade
well,
how
do
we
make
sure
that
they
get
that
adapter?
If
we
don't?
F
On
facilitating
upgrades
and
getting
people
new
features
automatically
automatically
without
while
at
the
same
time
trying
to
balance
honoring
any
custom
configurations
they
have
so
I
don't
know.
I
think
there
can
be
multiple
right
answers.
Yeah.
B
Hey
hilly
yeah,
so
actually
what
you
said
makes
sense,
and
here
I
that's
what
I
was
also
thinking
about
like
what,
if
user
provides
his
yaml
file
based
on
the
platform
and
and
we
use
instead,
that
file
instead
of
checking
the
docker
compose
and
and
simply
we
can
use
that
file
and
overwrite
it.
Instead
of
checking
which
parameter
is
changed
right
and
providing
options.
F
And
I
think
that,
like
it's
a
reasonable
like
it's
like,
there
are
multiple
wrong
answers
and
multiple
right
answers,
or
what
have
you
like
it?
I
they
almost
all
become
correct
answers
if
we're
just
not
lying
to
the.
If
we're
being
forthright
with
the
user
and
saying
hey,
don't
don't
touch
that
area
like
we're
going
to
overwrite
that,
like
you
know
or
like
as
long
as
we're
you
know,
yeah,
then
then
they
can
say.
Okay,
now
I
get
it
like.
Okay,
I
will
always
use
the
flag.
F
Hyphen
hyphen
skip
reset
and
I'll,
just
like
bake
that
into
my
alias,
because
I
never
want
measuring
ctl
to
overwrite
my
context,
because
I've
totally
customized
the
heck
out
of
it
and
so
like
like
part
of
it,
is
just
like
hey
at
some
point
we
have
to
have.
If
the
user
wants
to
be
smart
and
and
do
smart,
complex
things.
Okay,
then
they're
gonna
have
to
read
this
doc
because
because
that's
the
cost
of
them
doing
highly
customized
things
and
so
and
that's
totally
legitimate.
That's
okay!
Like
we
there's,
you
can
ask
that
of.
B
F
And
so
yeah
it
sounded
of
of
of
what
I
or
I
didn't,
listen
to
the
whole
conversation,
but
just
for
my
two
senses
like
I
think
that
all
of
you
have
sound
heads
on
your
shoulders
and
what
you,
just
I
just
heard
each
of
you
describe
was
like
those
things
make
sense,
and
I
think
you
get
the
problem
statement
at
hand
and
it's
kind
of
we
sort
of
have
one
foot
in
being
authoritative
for
their
manifest
and
one
foot
in
like
oh,
you
can
customize
it
and
to
start
with,
if
we
just
make
clear
where
that
line
is
we'll
be
in
a
defensible
position
with
the
user.
F
B
Oh
okay,
yeah
yeah,
I
mean
it's
still
like
will
be
clear
because
I
see
there
are
a
lot
of
like
changes
which
can
be
done
by
power
users
like
changing
core,
changing
volume,
power,
changing
adapter,
name,
changing
and
like
there
are
a
lot
of
things.
B
So
that
is,
I
mean
the
challenge
and
and
all
these
results
in
one
file
doc
and
if
you're
going
to
do
that,
and
so
I
think
we
need
some
design
here
and
like
we
need
to
think
and
maybe
I'll
think,
and
we
can
just
come
on
some
design
and
then
we
can
go
ahead
and
implement
this.
A
The
item
we
missed
was
just
demo
on
the
measuring
ctl
token
command.
So
would
you
like
to
demo
it
right
now
or.
E
E
I
missed
the
joke:
okay,
so
basically
the
token
command.
I
have
added
a
token
command
for
the
thread
operations
on
tokens
inside
mesh
config,
so
the
first
command
is
messy.
Ctl
system
token
create
so
right
now,
as
you
can
see,
the
token
salary
is
empty.
So
if
I
do
not
provide,
if
I
don't
provide
the
path
by
the
flag,
the
default
path
is
all
got
this,
so
I
can
create.
If
I
create
this
default
token,
it
would
be
added
here
with
this
default
path.
E
E
E
E
Another
command
is
so
we
have
a
list
command
which
will
basically
list
all
the
available
tokens
which
in
this
case
are
defaulted
for
two
another
one
is
the
delete
command,
so
I
can
give
it
so
basically
it
deleted
it
from
here.
Another
one
is
the
the
last
one
is,
I
guess
the
this
is
the
last
one,
so
we
have
create
delete
list
set
yeah.
Of
course,
the
set
command,
so
user
can
set
a
token
for
a
specific
context.
E
So,
if
no,
if
the
context
flag
is
not
given,
it
will
set
the
token
for
the
current
context.
So
right
now
I,
if
I
just
do
this,
it
will
set
the
token
for
the
current
context
as
the
name
I
provided
it.
E
So
there
is
this
issue
I
ran
into
so
the
thing
is
that
if
I
do
not
specify
the
token
name
here
and
if
I
just
comment
it
out
or
just
let's
say
that
user
deletes
or
removes
this
token
name
line,
and
if
I
try
to
login
the
thing
that
happens
is
that
I'll
just
show
you
what
happens
if
the
user
tries
to
log
in
so
the
login?
What
the
login
does
is.
It
goes
into
the
current
context.
It
looks
for
the
token
name.
It
then
it
writes.
E
The
custom
path
is
slash
home,
slash,
dot,
message,
slash
the
default
bus
or
dot
json.
So
it
writes
the
token
there
and
if
I
do
not
provide
the
token
name,
it
will
basically
just
do
wait,
I'm
still
initiating
login,
okay,
so
yeah,
so
it
it
gives
out
an
empty
token
name
here
so
which
will
basically
we
won't
be.
Even
we
won't
be
able
to
delete
it
because
there
is
no
name,
it
basically
doesn't
have
any
name.
E
So
one
thing
I
was
I
would
recommend
is
if
you
can
have
an
additional
check
inside
of
login,
that
if
the
token
name
is
not
specified,
we
can
go
ahead
and
maybe
tell
the
user
that
you
don't
have
a
token
name
or
maybe
set
that
open
name
for
this
specific
context.
Otherwise
we're
going
to
have
these
empty
token
names
here.
F
Something
that
I
think
we're
starting
to
see
happen
in
mastery
ctl,
and
that
is
like
this
we're
taking
thing
the
user
experience
kind
of
one
step
further
and
I
think
you're
you're
advancing
this
agenda,
and
that
is
like
when
you
know
we
we
can
tell
the
user
I
get
I
get
yeah
anyway,
we
can
give
the
user
what
they
want.
F
Like
sometimes
you'll
execute
a
command
using
a
cli,
you
misspell
the
command
and-
and
it
says
I
don't
know
what
you're
talking
about
it's
kind
of
what
it
responds
with
like
air
other
command.
Other
clies
you'll
execute
it
and
it'll
return
back
and
say:
oh
did
you
mean
this
other
word?
That
is
a
really
close
spelling
you're
like
oh
yeah,
that's
it
good!
Okay,
thanks!
F
That
was
helpful
and
I
think
we're
in
that
same
context
like
we're,
taking
the
next
step
and
we're
beginning
to
prompt
people
like
when
it's
become
when
it's
like,
really
obvious,
hey.
If
you
have
this
token,
that
is
a
white
space
for
a
name.
Maybe
that's
kind
of
an
illegal
thing
and
shouldn't
be
allowed
that
I'll
give
an
example,
and
I
don't
think
that
we
should
go
necessarily
go.
F
Do
what
I'm
about
to
say,
but
but
I'll
give
an
example
of
of
ideally
the
frame
of
mind
that
we're
using
as
we're
designing
commands,
and
that
is,
if
someone
says
token
list
and
the
go
the
golang
procedure
goes
to
the
function
goes
to
get
the
list
and
it
identifies
the
fact
that
there's
an
illegal
token
name
like
a
white
space.
Well,
why
would
that
be?
How
is
that
possible?
Well,
someone
probably
manually
edited
their
config
file.
Okay.
Well,
that's
a
so!
We
can
tell
the
user.
F
F
I
don't
think
that
we
should
go
through
all
that
trouble
for
the
thing
I
just
said,
but
I
wanted
to
call
it
out
as
I
don't.
I
don't
think
that
we
should
shy
away
from
prompting
people
for
some
things.
I
don't
think
that
we
should
go
do
what
I
just
said,
which
is
like
fix
all
the
things
for
them,
but
there's
a
certain
set
of
comment
or
pitfalls
that
will
be.
I
thought
I
thought
navendi
was
commenting
on
what
I
was
saying.
I
was
like
oh
geez,
okay,
I
guess
not.
F
I
don't
know
you
know
geez,
no,
there's
a
certain
set
of
pitfalls
that
people
will
run
into
and
there's
a
subset
of
those
that
will
be
common
pitfalls,
that
we
might
want
to
take
an
extra
moment
to
sort
of
try
to
help
them
address
the
issue,
prompt
them
with
something,
and
so
that
that's
great
and
I
think,
are
you
using
prompt
ui
in
a
couple
of
places.
F
Cool
and
that
like
when
you
think
about
it
from
that
perspective,
it's
like,
oh,
you
know
like
okay,
like,
like,
I
think
most
of
you
who
most
of
us
most
of
you
who
work
on
mastery
ctl
thank
you
neha.
It
becomes
pretty
evident
or
it
becomes
pretty
easy.
F
I
think
did
we
merge
your
token
pr
already
and
you're
working
on
unit
tests
separately.
E
F
If
it
was
under
experimental,
then
we
should
just
go
merge
because
it's
it's
it's
pretty
good.
It's
been
through
a
few
reviews
you
know
like
since
it
isn't.
I
don't.
F
There
aren't
many
people
that
are
just
randomly
discovering
new
commands,
so
I
would
ask
you
know
husseina
and
nivendu
if
it
would
be
okay
to
go
ahead
and
merge
the
token
pr
understanding
that
I
think
right
now
it
it
does,
does
it
or
does
it
not
have
an
adverse
effect
on
how
tokens
are
referenced
currently
like,
like,
I
guess,
what
I'm
trying
to
say
is
is
does
this?
Is
this
a
one-time
shift
and
it'll
break?
If,
if
this
isn't
done
well,
is
it
gonna
break
other
things
or
is
this
just
augment?
A
F
F
It's
like
it's
like
another
another
15
seconds
down
the
drain,
like
everyone
has
to
reach
for
their
mute
button
like
what
I
don't
know,
and
then
10
people
say
yeah,
and
then
you
say
what
I
couldn't
you
can
or
you
can.
You
know
somebody's
gonna
pay,
two
dollars
for
yeah.
F
A
Yeah,
thank
you
all
for
joining
and
staying
up
late.
So
I
guess
I'll
see
you
on
the
next
week's
meeting
and
maybe
I'll
see
you
on
the
community
meeting
on
friday
as
well.
A
Just
a
quick
note:
we
have
a
new
tomorrow
for
service
mesh
performance,
which
is
another
layer,
5
project
that
has
been
recently
adopted
into
the
cncs.
So
I
think
it's
in
it's
the
same
time
tomorrow
same
time
as
the
machine
development
meeting,
so
tomorrow
will
be
the
first
official
meeting
after
we
have
where
we
have
been
adopted
into
the
sessions
here.
So
it's
on
the
community
character.
So
if
you
are
interested,
please
join
that
with
that
I
beat
you
farewell
bye.