►
From YouTube: Meshery Development Meeting (Oct 20th, 2021)
Description
Meshery Development Meeting - October 20th 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
B
Hello,
so
utkarsh
is
not
here
right.
I
think
he
will
be
leaving
the
player
so
if
he
will
be
coming
back,
that
would
be
much
better
by
the
way
I
can
just
show
it.
A
C
E
E
D
B
E
For
the
pattern,
then,
for
separating
out
the
configuration
format,
patterns
and
application
to
be
reusable
across
the
same.
So
today
I
had
a
talk
with
abby
on
the
required
changes,
so
we
had
around
20
minutes
long
talk
and
they
had
has
pointed
me
across
the
changes.
What
has
to
be
made
so
yeah
I'll,
be
making
the
changes
and
updating
the
pr
again
for
the
review.
D
E
Yeah,
the
one
that
donation
has
yeah
yeah-
I
I
actually
I
have
been
in
touch
with
nitish
on
that.
I
haven't
started
working
on
it
as
of
now,
but
yes,
I'll,
be
working
along
with
nutrition,
but.
D
Cool
good,
I
just
wanted
to
check
that
you
were
aware
and
actually
kind
of
call
it
call
it
out
for
a
lot
of
the
about
half
of
the
rest
of
us
that
we
are
where
we're
like
on
the
precipice
of
dot.
Six
like
that,
they
were
just
like
trying
to
tie
off
loose
ends
and
that's
in
part
what
we're
talking
about
today
and
that,
as
we
look
forward
to
dot
seven,
we
might
wanna
well,
it's
not
that
we
might
want
to
like.
D
We
should
have
an
explicit
set
of
discussions
around
things
that
are
going
into
dot,
seven
and
kind
of
what
the
the
themes
are,
or
the
theme
is
but
and-
and
so
as
part
of
that
excuse
me,
I
think
it's
understood
that
there's
a
re-architecture
of
the
ui,
that's
part
of
that
forthcoming
dot,
seven!
So
just
a
healthy
reminder!
D
E
A
Yep,
thank
you
sane,
so
for
the
next
topic
rudraksha,
would
you
like
to
talk
about
updating
the
deployment
of
operator
to
help
from
manifest.
G
G
So,
basically,
it's
about
this
year
in
v
0.6.
We
have
moved
to
helm
for
installing
measuring
and
it
would
be
good
to
establish
uniformity
across
all
deployment
methods,
so
that
I
mean
when
you
install
using
helm
now
so
the
hem
chart
should
install
mystery
server
operator
broker
on
all
those
controllers.
G
But
we
have
a
mechanism
inside
mystery
server
right
now,
which
actually
gets
triggered
using
graphical
subscriptions,
and
when
you
visit
that
operator
page
in
mystery
server
this
settings
page
where
you
can
change
your
kubernetes
context,
you
can
have
the
control
over
there
and
if,
if,
in
that
case,
mystery
server
doesn't
found,
it
doesn't
find
operator
deployed
in
your
cluster.
It
automatically
triggers
its
deployment
and
that
actually
happens
right
now
using
kubernetes
manifest.
G
So
this
pr
is
an
effort
towards
changing
that
from
plane,
manifest
to
himself
now.
Basically,
some
of
the
things
about
this
are
we
don't
actually
have
separate,
charts
or
operator
messing
and
all
that
they
are
right.
Now
the
sub
charts-
and
I
might
be
unaware,
but
the
only
way
I
could
think
of
to
install
only
the
operator
is
fits
the
tar
which
generally
contains
hem,
charts
and
supply
the
part
of
the
chart
to
the
hem
chart
functions
in
biscuit,
so
they
can
install
it.
It's
a
dirty
way,
but
I
don't
know
if
anything
else
exists.
G
This
was
a
little
discussion
I
had
with
darren
as
well
and
he
suggested
a
few
ways.
One
of
them
was
using.
Overwrite
overrides
is
overrides.
If
it
doesn't
discover
mystery
server
in
the
cluster,
it
could
reinstall
it
which
we
might
not
want.
When
mystery
server
is
actually
deployed,
but
not
in
cluster,
and
then
there
was
upgrade,
he
said
we
can.
G
We
can,
rather
than
using
install.
We
can
still
use
overrides
and
use
upgrade
instead,
but
the
problem
with
upgrade
would
be
that
it
expects
mystery
server
to
be
pre-installed
in
the
cluster,
so
that
would
again
fail
so
yeah.
This
was
a
thing
where
we
were
kind
of
having
discussion.
The
current
approach
works.
G
G
C
G
G
G
G
Now
this
would
create
troubles
when
misery
is
deployed
in
cluster,
because
mystery
server
would
not
be
able
to
play
with
the
operator
in
that
case,
because
it
would
try
to
find
one
of
these
releases.
They
don't
exist
because
this
comes
because
in
when
done
using
measuring
cpl,
they
come
under
misery
release,
so
in
that
case
it
would
fail.
G
G
G
G
D
D
So,
to
replay
this
back
to
you,
it's
that
when
measuring
server
connects
to
a
kubernetes
cluster,
it
checks
to
see
if
the
operator
is
deployed.
D
D
Today,
currently,
in
the
current
version
of
mesheri
mesh
redeploy
mesh
reservoir
deploys
the
operator
using
kubernetes,
manifest
as
it
switches
to
use
helm
to
do
that.
Deployment.
D
D
D
Well,
well
I
mean
measuring
ctl
systems
start
when
that
the
plat.
D
I
D
So
measure
server
is
responsible
for
the
genesis
of
measuring
operator
on
docker
host.
You
know
docker-based
deployments,
kubernetes
deployments.
Mesh
reservoir
is
still
responsible
to
make
sure
that
an
operator
is
up
and
running
most
of
the
time
the
operator
would
be
present
and
running
because
it
would
have
because
measuring
ctl
would
have
initiated
that.
G
G
I
just
found
that
error,
which
you
would
actually
get
when
you,
when
misery
server,
is
in
cluster
and
tries
to
deploy
the
adapter
this.
This
is
the
one
the
one
I've
highlighted
and
it's
basically
happening
because
of
the
reason
I
pointed
out
here
that
it
installs
them
as
a
single
chart
here,
but
here
it
is
installed
collectively
as
a
mystery
chart.
D
And
so,
given
that
you
kind
of
have
a
choi
or
we
or
you
have
a
choice
between
different
helm,
verbs
and
different
behaviors
that
you
can
configure.
So
it's
install,
update,
override
and
there's
behavioral
differences
between
those.
G
Yes,
there
is,
but
I
mean
we
can
probably
change
one
of
these
release
names
or
we
can
change
what
it
expects,
but
the
thing
is
that
it
expects
a
charge
like
in
this
case.
There
isn't
any
charge.
The
chart
doesn't
exist
like
this
release
mystery
operator
release
when
charts
are
actually
installed
in
cluster.
They
are
called
release
and
when
this
is
installed
using
helm,
they
don't
exist.
So
there's
no
way
to
like
specify.
G
Yeah
so
releases
are
like
it's
a
helm
terminology
which.
F
D
So
I
guess
my
very
maybe
the
very
first
question
I
was
asking
is
actually
significant
to
this
conversation,
because,
depending
upon
how
the
logic
by
which
measures
server
is
determining
whether
or
not
the
operator
has
been
provisioned
or
not
sounds
like
is
material
to
this
here,
because
if
like
one
way
is
like
hey,
is:
are
there?
Are
there
custom
resources
defined
for
the
measuring
broker
and
for
mesh
sync
for
the
two
custom
controllers?
D
D
D
So,
okay,
you
know
what
actually
this
is
man.
This
is.
Can
I
catch
up
with
you
after
this
call,
because
there's
this
is
fantastic.
This
is
a
brain
tickler.
I'm
super
interested.
D
I
think
we're
going
to
lose
everybody
else
in
the
on
the
like
this
is
this
is
the
right
thing
to
do
to
bring
this
up
like
this
is
an
exploration
of
the
behavior
of
what
happens
when
you
use
overrides
versus
what
happens
when
you
use
upgrades.
I
know
these
are
things
that
you've
been
exploring.
You've
done
a
lot
of
diligence
here.
This
is
this
is
super
important,
really
important
to
get
right.
We
don't
get
this
right
and
mesh
free
operator
which
includes
mesh
sync,
doesn't
deploy
yeah.
D
G
On
this,
so
I
guess
that
indicator
that
you
just
gave
like
checking
if
the
cr
exists
like
changing
the
way
how
we
check
if
measuring
operator
exists.
C
D
A
Thank
you
rudraksha.
So
let's
go
back
to
abhishek's
topic
for
deleting
multiple
patterns
abhishek.
Whenever
you're
ready.
B
Yeah
so
earlier
we
were
deleting
like
whenever
I
have
to
delete
10
patterns
per
se.
I
would
be
making
10
requests
to
this
server.
Then
server
would
make
a
10
request
to
cloud
and
then
cloud
would
make
10
requests
to
database.
Then
database
delete
and
then
10
10
10
responses
right.
So
it's
it's
good,
but
if
the
server
is
like
a
little
bit
occupied,
then
these
number
of
requests
just
fails
away
and
it
also
causes
like
a
lot
of
network.
B
Latency
like
there
are
network
requests
whenever
we
are
making
requests
to
the
server
or
server
is
making
requests
to
cloud.
So
for
that
reason,
if
you
are
about
to
delete
more
than
like
five
ten
patterns
like
right,
so
we
would
be
getting
errors
like
he
was
getting,
which
was
like
a
lot
of
network
requests
or
something
similar.
B
But
for
that
reason
we
have
okay.
So
that
is
one
more
thing
that
just
ui
just
sometimes
freeze
up
when
more
than
10
five
patterns
it,
although
it
looks
very
clean
and
like
very
great
twice
but
yeah,
so
I
just
deleted
10
buttons
and
10
buttons
were
deleted.
B
So
I
have
what
I
did
is
like
created
a
new
function
to
delete
multiple
patterns
inside
server
and
cloud
that
makes
a
request
to
the
database
itself
like
it
can
delete
any
number
of
patterns,
and
that
would
be
just
one
request
from
the
server
from
the
cloud
to
the
database
so
yeah.
If
I
have
to
delete
10
buttons
now,
so
it
will
just
one
network.
B
B
Delete,
so
are
you
sure
you
want
to
leave
10
buttons
and
if
I
say
yes,
then
those
10
buttons
will
be
deleted
and
10
buttons
deleted.
You
will
get
the
message
right.
Okay,
it's
not
refreshing.
B
Probably
I
have
I
have
to
okay,
it
was
refreshing
earlier
it's
not
refreshing
right
now.
I
don't
know
why,
but
I
will
check
it
once
again,
other
than
that
that
was
just
the
gist
like
all
those
patterns
were
actually
deleted
and
you
can
see
like
no
reports
found
wow.
D
B
I
don't
know
it
needs
a
review
first
of
all,
so
it
can
be
merged
after
that
other
than
that,
I
I
just
see
like
it's,
it's
not
deleting
it
was
actually
earlier.
So
maybe
I
just
put
latest
changes
and
that's
causing
things
to
break,
but
if
you
can
see
the
response,
then
yeah,
it's
saying
like
10
buttons
are
deleted
and
if
I
refresh
it.
B
So
I'm
getting
200
response
from
the
post,
which
is,
I
think,
not
good
right.
No,
no
200
is
great,
adding
new
files
that
it
should
give
201,
but
I'm
making
a
post
request
for
delete
and
200
is
for
like
the
successful
deletion
so
yeah
I
just
deleted
it.
Maybe
this
was
not
working.
I
don't
know
somehow
I
just
clicked
cancel
or
so.
B
So
that's
why
I
just
I
had
different
other
than
that
you
can
see
like
the
patterns
are
getting
deleted,
so
I
deleted
just
20
buttons
with
two
requests
other
than
that.
If
I
could
have
deleted
it
in
the
previous
way,
then
I
could
have
made
20
requests
and
then
server
would
be
making
20
requests
to
the
cloud
and
so
on
so
yeah.
That
was
not
a
pr.
B
I
I'm
just
about
to.
You
know
resolve
the
merge
conflicts
yeah.
It
is
highly
dynamic
place
where
everyone
is
writing
code.
So
that's
why
we
are
getting.
B
Yes,
yes,
yes,
definitely
they
can
delete.
Any
number
of
patterns
like
25
is
like
at
one
time.
We
can
show
25
tables
in
the
in
25
rows
in
a
table
and
all
the
25
patterns
can
be
deleted
at
once
so
yeah
I
have
to
just
upload
25
steps.
D
Is
this
yeah
very
good?
Is
this
bulk
deletion
supported
for
other
content
types
as
well.
D
B
No,
it's
it's
currently
for
patterns
only
because
I
just
want
to
have
like
a
you
know,
a
thumbs
up
or
whatever.
We
call
success
if
it
is
like,
so
we
can
just
copy
and
paste
do
the
other
things
also
like
for
patterns
and
applications
too.
B
So
if
it
needs
some
optimizations,
then
definitely
other
than
that.
I
think
the
pr
is
up
for
review
and
it's
just
about
like
one
much
conflict
like
this.
This
is
the
only
thing
so
it
will
be
like
yeah.
It
will
do
that
so,
okay,
getting
back
to
question
it
can
delete
any
number
of
patterns
like
25,
40
50,
whatever
it
is,
it's
just
about
the
database
has
the
ability
to
delete
like
it
should
have
so
yeah.
D
Nice
any
one
thing
about
the
different
types
of
content
that
are
here.
That's,
I
think
I
think
some
of
you
have
taken
some
steps
on
is
to
avoid
the
copy
paste
to
have
you
know
like
bulk
delete,
is
a
good
example
of
an
operation
that
would
ideally
be
implemented
once
and
is
applicable
to
each
of
these.
I
don't
know
that
the
way
that
our
resources,
our
endpoints
between
graphql
and
rest,
that
I
don't
know
that
that's
easy
to
avoid
you
know
implementing
multiple
times,
which
actually
leads
me
to.
D
B
No,
I
don't
think
so.
It's
it's
fully
managed
by
the
rest
apis.
So
it
there
is
no
graphql
in
the
patterns
page
or
filters
pages
any
in
any
of
the
configuration.
D
As
part
of
the
7
ui
re-architecture,
one
of
the
goals
is
to
ideally
is
to
eradicate
the
use
of
rest
from
this
client,
and
so
such
that,
when
you
delete,
you
know
eight
patterns
that
the
user
would
not
need
to
refresh
to
see
the
fact
that
those
are
gone.
And
then
you
know
have
ten
more
eight
more
loaded.
B
Okay,
okay,
okay,
so
yeah!
Yes,
yes,
yes,
other
than
that,
it
doesn't
need.
We.
We
don't
need
to
reload
it
from
here.
It
will
automatically
after
deleting
10
patterns.
Then
it
will
fetch.
It
will
make
a
new
network
request
to
fetch
like
what
is
the
new
data
and
it
will
just
show
it,
so
we
don't
have
to
refresh
it
for
the
first
time
I
have
to.
I
don't
know,
maybe
I
I
said
no
to
this
one.
B
Whenever
I
was
getting
no,
I
said
yes,
so
I
don't
know.
D
That's
okay,
yeah!
It's
not
not
necessarily
a
big
concern
at
the
in
this
release.
I
mean
it's
a
nice
user
experience.
It
sounds
like
you've
considered
for
it
by
automatically
getting
a
fresh
record
set,
which
is
great
going
forward.
Hopefully
that
would
be
a
subscription
that
you
know
we
would
knit
yes.
B
Other
than
that,
like
I
really
have
a
strong
urge
to
you
know,
use
eggs
yours.
Why?
Because
I'm
really
having
very
much
trouble
managing
requests
and
responses
with
the
fetch,
so
I
just
tried
it
in
this
case
only
I
tried
a
lot
of
different
things
to
you
know
simulate
the
best
user
experience,
but
due
to
some
reason,
it's
due
to
the
way
fetch
is
being
used.
B
We
have
machine
cloud
stopped,
so
maybe
you
know
cloud
is
cloud,
so
it
can
stop
at
some
time
so
it
just.
I
don't
get
the
error
request.
It
just
returns
back
the
it
get
returns
by
500,
but
it
is
not
taken
up
by
fetch.
It
just
says
like
it's:
okay,
so
200
whatever.
So
I.
D
B
Know
if
we
are,
if
there
is
a
problem
with
the
fetch
where
we
have
you
know,
written
fetch
request
the
central
point
from
where
we
are
you
know
making
all
the
api
request
calls
if
there
is
a
problem
or
something
else
for
some
reason.
D
Okay,
good
good,
good,
very,
very
good,
based
on
the
time-
let's,
oh
so,
let's
go
to
the
next.
That's
a
great
suggestion.
I
think
you've
written
some
stuff
about
it.
If
you
haven't
already
dropped
it
into
the
mesher
ui
channel,
please
do
or
if
you
already
have
just
kind
of
call
attention
to
it
again
and
and
yeah,
let's
review
a
pro
a
written
proposal,
so
it
sounds
like
and
I'm
assuming
that
axios
works
complements
dot
the
dot.
Seven
redesign
re-architecture
as
well.
B
B
I
mean
why
can't
why
can't
we
do
it
in
with
a
single
endpoint
deleting
filters
patterns
and
applications
with
a
single
input
right
yeah,
we
can
attach
extra
metadata
about
what
what
kind
of
content
we
are
actually
trying
to
delete
and
at
the
back,
and
we
can
figure
out
and
call
the
appropriate
functions
accordingly.
B
Okay,
that
sounds
good,
but
what
I
saw
is
like
filters
are
managed
on
different
endpoints,
like
the
fetching
of
filters
and
patching
of
patterns
are
actually
different.
After
that
reason,
I
just
thought
like
deleting
it
would
also
be
different
other
than
that.
It
also
looks
great
to
attach
things
on
the
metadata
since
patterns
applications
and
filters
are
somewhere
are
very
much
related
to
each
other,
so
yeah.
D
Each
content
type
is
treated
as
its
own
resource,
and
so
each
resource
has
crud
operations
against
it
and
yet,
to
the
extent
that
there's
reuse
between
like
a
common,
you
know
like
commonality
between
the
way
that
resources
are
deleted.
You
know
great.
The
reality
is
like
some
of
them
will
be
treated
different
from
the
others.
D
It
seems
like
we're.
Gonna
carry
that
forward
to
like,
if
you
scroll
to
the
right
a
little
bit.
I
think
we
have
the
some
of
the
implementation
of
graphql
for
those
same
resources
and
it's
really
lining
up
mostly
the
same
way
that
there'll
be
individual,
endpoints,
graphql
endpoints
for
on
a
per
resource
basis.
D
A
Thank
you,
abhisheka
next
up
mano
has
some
update
on
some
bug
fixes
in
the
with
with
prometheus
so
mara.
Would
you
like
to
okay
thank.
C
C
B
I
know
yeah
so
in
the
last
mystery
mesh
map
call
we
discussed
where
it
was.
There
was
an
issue
with
the
when
we
were
acquiring
permission
and
nelly
also
raised
that
it
was
not
grafana
which
we
need
to
query.
We
need
to
call
it
query,
primitio
or
api.
So
when
I
went
into
single
implementation,
yes,
we
were
using
grafana
to
query
the
json
format,
which
was
throwing
error
because
it
was
trying
to
connect
to
graphana
client
instead
of
the
mutual
client.
B
So
there
here
it
was
trying
to
connect
to
grafana
client
with
a
pro
miniature
url.
So
it
was
throwing
an
error
that
no
client
exists.
So
I
replaced
the
code
with
the
api
of
promation
query.
So
we
directly
query
now
permission
for
the
premature
segment
of
the
fetching
matrix
and
we.
D
Here's
one
of
those
smart
alec
comments,
probably
but
that
ends
up
being
somewhat
serious
over
time,
is
that
we
may
need
to
run
some
of
these
query
string
parameters
through
a
bit
of
a
filter,
a
bit
of
a
security
filter
or
a
you
know.
Potentially
a
top-end
filter
like
to
you
know
to
even
very
qui
qualify
that
someone's
specifying
a
time
stamp
where
they
should
be
or
not,
but
but
directionally,
or
at
least
like
of
this
implementation.
Monav.
B
Yeah,
I
think
that
barrick
has
I.
My
server
is
not
yet
up,
but
I
think.
J
I
think
actually,
this
issue
is,
is
different
from
the
one
that
you're
talking
about,
so
the
issue
that
you're
fixing
will
be
so
initially.
If
so,
actually
this
issue
will
be
like
there
is
a
custom
panel
that
is
provided
by
grafana,
and
I
was
trying
to
connect
it
with
the
our
current
ui.
So
so
this
is
actually
an
issue
with
the
ui,
because,
because
the
fields.
J
J
D
I
guess
just
to
clarify
with
this
pr,
though,
the
ability
for
a
user
to
take
any
custom
grafana
board
using
a
grafana
board's
json
description
to
paste
that
json
description
into
meshre's
metrics
settings
that
this
fixes
that
such
that
they
can
query
any
prometheus
metric
that
they
want
to
right.
J
J
So
the
thing
is,
we
need
to
fix
the
custom
value
first
and
if
we
can
display
it,
then
we
can
check
whether
the
custom
panel
works
with
this
code.
Basically
yeah.
The
connection
is
tab
corrected,
but
the
display
is
not
there.
D
Cool
okay,
yeah,
that
makes
sense
yeah.
That's
a
good
qualifying
comment
for
everyone
else
is
not
like
understanding
just
for
context,
for
everyone
else
is,
I
think
what
monik
monav
and
barak
are
saying
are
saying.
Is
that
the
fix
that's
going,
that
this
is
a
good
fix,
there's
another
step
to
be
taken
beyond
this,
but
when
that
other
step
is
taken,
the
goal
is
or
the
the
function
this
functionality
will
be
restored
and
the
functionality
is
that
mesherie
connects
to
right
now,
either
a
grafana
and
a
prometheus.
D
If
a
lot
of
you
might
not
have
noticed,
you
might
have
noticed,
you
might
not
have
noticed
that
if,
if
you
want,
if
you
measuring,
connects
to
a
grafana-
and
it
will
read
in
existing
boards,
that
grafana
has-
and
so
then
you
can
pull
in
metrics-
and
this
is
great.
This
is
a
really
powerful
set
of
capabilities
that
we've
we're
still
collectively
working
on
how
to
expose
to
users
in
a
much
in
a
better
way.
D
D
Grafana
mesri
talks
to
grafana
grafana
in
turn,
talks
to
prometheus.
That
has
all
of
the
metrics
that
have
been
collected
if
the
users
so
desi.
If
the
user
says
well,
you
know.
Rather,
I
would
like
to
have
measury
speak
directly
to
request
metrics
directly
from
prometheus.
D
D
The
current
user
experience
is
here
is
that
they
describe
that
using
grafana
markup,
a
grafana
json
board
descriptor,
and
they
they
paste
that
in
here,
when
they
paste
that
in
meshrey
speaks
interprets
that
forms
a
graphql
query.
I'm
sorry,
a
prom
ql
query
forms
a
query
to
call
prometheus
to
retrieve
the
data
so
more
direct
connection,
and
so
there's
this
was.
This
is
a
not
a
good
user
experience
here.
It's
cool
that
you
can
create
whatever
board
you
want
and
paste
in
well
up
to
the
point
that
that
next
step
is
taken
around
you
know.
K
Leave
one
question
here
so
when
you
paste
the
board
here,
does
it
I
mean
populate
the
boards
on
grafana
graph?
On
a
panel
I
mean
I
mean
the
board.
Is
you
have
busted
your
custom
board,
but
the
ui
want
to
see
the
move.
Does
it
automatically
populate
on
the
next
side.
B
B
Discuss
with
you
hope,
it's
visible
yep,
so
the
first
change.
Basically,
I
made
is
like
with
this
release
panel
over
here,
which
definitely
doesn't
belong
to
dashboard.
As
we
have
talked,
I
shifted
its
detail
to
over
here.
So
in
the
footer
you
could
see,
there
is
the
version
which
currently,
which
version
is
being
run
and
then,
when
we
expanded,
we
could
see
more
details
about
it.
B
We
could
visit
it
and
if
we
are
not
updated,
an
exclamation
mark
over
here
will
be
will
appear
as
stating
that
we,
we
are
not
updated
to
the
latest
version
and
we
could
click
and
go
to
it.
The
second
change
which
I
made
is
the
scroll
bar.
So
if
you
could
see
initially,
the
scroll
wire
was
the
native
score
bar,
which
is
a
block
diagram
with
gray
fill
color.
I
have
updated
that
over.
A
D
A
B
B
D
Yeah,
for
my
part,
I
totally
agree
I'll,
send
a
a
video
recording
of
a
different
style
interaction
for
inspiration.
If
you
go
actually,
if
you
just
go
into
google
docs
you'll
find
on
the
right
hand,
side.
Actually,
this
will
actually
facilitate
us
getting
back
to
the
next
presenter.
D
So
here's
a
google
doc
this
one's
tracking
meeting
minutes
of
various
things,
so
it's
probably
pretty
small
on
a
lot
of
your
uis
or
a
lot
of
your
it's
probably
small
in
your
anyway.
Here's
a
google
doc
you're
doing
stuff
they've
got
there's
an
extra
column
menu
over
here
for
add-ons
and
things
that
I
don't
use
right
now
and
so
for
me,
it's
like
well,
yes,
I
would
like
to
hide
this
panel
and
check
out
how
slick
that
is
whoa.
The
panel
is
totally
gone.
D
You
got
and
there's
just
a
little
overlay
and
pops
out,
like
that's
some
pretty
decent
inspiration
for
kind
of
a
way
to
save
space.
You
can
see,
there's
another,
isn't
it,
let's
see
what
happens
when
these
two
come
next
to
each
other
or
don't
they?
Oh
they
don't
one
chases
the
other
but
yeah.
So
that's
you
know
think
about
that.
If
you
would
like
that,
we
don't
need
for
them
to
be
yeah,
we're
wasting
a
bunch
of
space,
so
who's
next.
L
Yeah-
actually
it's
not
much.
It's
just
note
that,
after
after
we
make
the
big
release
of
mystery,
which
is
0.6.0,
there
is
a
pr
that
that
is,
that
is
trying
to
introduce
multi-cluster
support
in
mystery.
A
bit
of
mystery
still
supports
multiples,
so
you
can
configure
it
to
talk
to
more
than
one
cluster
at
a
time.
L
So
now
this
particular
pr
is
attempting
to
support
communication
with
multiple
clusters
at
the
same
time,
but
we
are
holding
on
to
this
vr
so
that
so
that
the
bugs
in
that
vr,
which
probably
a
few
of
you,
would
report
back,
it
doesn't
doesn't
blocks
the
version
0.6.0
release.
So
the
thing
the
the
note
that
I
want
to
give
in
here
is
that
I
have
written
down
some
instructions
in
slack
and
I've.
L
I
wanted
to
place
it
in
here,
but
they
were
quite
huge,
so
I
haven't
pasted
them
here.
If
you
would
click
on
the
instructions
link
on
the
mystery
message,
development
meeting
minutes,
you
would
see
us
thread
thread
in
the
thread.
I
have
mentioned
the
instructions
how
you
can
actually
test
out
this
new,
a
new
feature
in
your
mystery
deployment.
L
So
yeah,
please
test
it
out
and
report
back
the
bug
that
you
face
so
that
I
can
fix
them
because
it's
critical
it
has
a
huge
surface
area,
so
there
would
be
bugs.
I
mean
I
don't
write
perfect
cold.
So
please
give
some
feedbacks
in
report
bugs
that's
pretty
much.
It.
L
Yeah
right
now,
no,
we
have
written
them
down
only
in
slack
I'll.
What
says,
should
I
place
them
here?
Because,
because
this
is
this
is
persistent
I
mean.
Should
I
paste
the
instructions
here?
Would
that
help
or
room.
D
Yeah
yeah,
I
think
so
you
know
as
a
matter
of
fact,
aditya
or
andor,
michael
and
others,
so
by
the
way,
with
just
a
brief
history
when,
when
mr
gefeler,
one
of
the
messaging
maintainers
joined
our
small
gang-
and
he
grew
it
significantly
with
mesh
kit
and
a
bunch
of
other
things
that
one
of
the
things
that
he
you
know
acknowledged
is
that
we've
got
a
lot
of
tribal
knowledge
that
is
ephemerally
put
into
slack.
And
so
it's
taken
us
a
long
time.
D
But
we
brought
out
two
other
forms,
two
other
mediums,
email
and
the
disqus
forum,
both
of
which
are
persistent
and
we're
not
using
the
email,
a
bunch
think
about
this.
A
little
bit
like
hey.
Anyone
like
this
is
a
great
actually.
D
So
I
think
an
immediate
answer
is
like
this
is
a
great
thing
to
put
on
the
discus
forum
like
as
a
hey
preparation
for
usually,
I
think
we
think
of
a
discussion
forum
as
like
user-facing
type
questions,
but
but
that's
not
doesn't
have
to
be
the
case
at
all,
and
so
that's
a
good
one,
but
also
if
the
rest
of
you
would
think
about
your
prep.
Like
there's
a
there's,
a.
D
I
think
it's
developers,
or
it's
dev
group
at
meshri.io.
If
you
go
to
mesri.io
subscribe,
there's
a
number
of
mailing
lists
there
and
so
there's
a
maintainer's
mailing
list.
There's
a
there's,
a
contributor's
mailing
list
and
a
user's
mailing
list.
The
contributors
mailing
list
might
be
a
good.
You
know
another
good
place
for
this,
and
so
it
could
be
that
karsh
you
took
advantage
of
both
you
made
a
post
on
discus
and
then
you
sent
out
an
email
that
briefly
describes
it
and
has
a
link
to
discuss
food
for
thought.
D
I
Yeah
navindo:
could
you
open
up
the
pr.
I
So
here's
a
brief,
so
my
task
was
to
create
a
documentation
page
for
system
check
command
on
mischief
ctl,
since
there
was
no
page
regarding
it.
So
also,
I
had
to
create
a
guide
which
will
help
users
to
what
you
can
see
how
to
perform
system
checks
using
measuring
ctl
and
along
with
some
faqs
in
it.
I
If
you
could
see
like
there
will
be
mentioned
some
kinds
of
deployment
checks,
what
are
these
kinds
and
how
to
perform
them
and
then
a
separate
set
of
faq,
ranging
from
pre-flight
adapter
checks
full
and
even
what
are
the
require?
Some
of
the
requirements
of
measuring
what
kubernetes
version?
I
need
what
version
of
cube,
cdl
and
much
more.
This
is
still
in
progress,
but
for
an
hour
ago
I
updated
them
so
yeah,
that's
much
of
it.
So
any
feedback
of
this.
A
G
A
Yeah,
like
the
plan
here
is
to
like
we
will
be
pointing
more
users
to
this
command,
to
ensure
that
their
environment
is
ready,
and
I
think
pure
is
working
on
a
command
that
would
create
a
diagnostic
bundle
that
with
which
they
can
directly
open
up
an
issue.
So
this
is
pretty
important
and
like
we
will
see
more
changes
in
this
talk
in
the
future
as
well.
So.
D
I
mean
yeah
part
of
the
part
of
the
problem
with
showing
output
or
showing
a
screen
by
the
way.
If
we're
going
to
show
output
and
textual
format,
we
probably
wouldn't
want
a
screenshot.
We
want
it
to
be
part
of
the
problem
with
doing
that
is
like
it's
going
to
change,
and
so
that,
but
that's
fine,
because
aditya
is
here
and
it
helps
keep
this
stuff
fresh.
D
This
is
fantastic
man
I
mean
like
like
this
would
be
highly
referenced
to
be
something
that
a
lot
of
users
come
by
and
they're
like.
Oh,
you
know
I
couldn't
I
couldn't
get
it
deployed.
Like
oh
look,
did
you
check
out
this
guide
and
then
hopefully
the
system
check
itself
really
kind
of
gets
them
through.
D
What's
going
on,
they
may
need
to
refer
back
to
this
guide,
though,
when
a
particular
check
fails
and
piyush
who's,
not
on
the
call
has
made
sure
that,
through
mesh
kit
or
through
the
error
codes,
there
are
specific
error
codes
for
each
type
of
failed
check,
of
which
we
you
know,
of
which
we
would
do
really
well
to
describe
in
full
within
your
guide
like,
and
if
the
the
guide
ends
up
potentially
being
renamed
to
like
troubleshooting
measuring
deployments
yeah,
you
can
imagine
it
being
fairly
highly
referenced
and
within
here
it
might
be
that
what
you
have
now
becomes,
or
the
bulk
of
what
you
have
for
system
check,
becomes
a
section
of
the
guide.
D
It's
like
here's,
how
you
can
troubleshoot
using
system
check,
here's
the
faqs
that
you
have
like
here's,
if
you're
in
docker
and
here's,
if
you're
in
kubernetes
or
here
you
know,
but
that
all
can
evolve
over
time,
and
so
this
is
great
like
for
michael.
D
Can
you
scroll
down
the
page
if
you
would,
for
my
part,
yeah
like
I
still
don't,
I
think
we
have
a
google
doc
that
the
design
spec
for
system
check
that
has
a
list
of
what
all
of
the
checks
are
would
be
great
if
we
copy
and
paste
that
list
into
here,
and
then
that
list
actually
provides
you
with
good
structure
to
fill
in
saying
well,
this
is
what
this
check
does,
and
this
is
why
it's
needed,
and
this
was
the
why
it
might
possibly
fail,
and
this
is
what
you
probably
want
to
do
if
it
does
fail
and
those
are,
that
would
be
awesome.
I
Okay,
I
just
mentioned
them
in
layman
terms,
so
that
users
can
easily
understand
them.
D
D
I'll
say
this
that
instead
of
four
basic
tasks-
or
you
might
say,
for
basic
tasks
and
then
give
an
example
of
like
performance
management
generally,
that
that
category
of
functionality
works
without
kubernetes,
but
for
any
life
cycle
management
of
service
meshes
themselves,
like
you,
gotta
that
you're
going
to
need
to
kubernetes,
so
performance
management
works
without
kubernetes,
but
kind
of
the
rest
needs
kubernetes.
D
H
H
Okay,
so
initially
the
issue
that
I
observed
in
this
doc
doc
page
was
basically
that
the.
D
H
The
the
page
link
itself
was
present
in
this
registered
reading
section
like
I
found
that
a
bit
redundant,
because
we
are
on
the
page
itself
and
the
there
was
a
link
for
it
in
the
solution
section
itself
so
yeah
I
created
a
pr
around
it,
and
then
I
just
suggested
that
maybe
we
could
make
a
reusable
component
for
the
rest
reading
section,
because
this
particular
section
was
repeated
in
many
pages
like,
for
example,
it
was
in
this
field
also
so
like
this
was
this
section
was
in
like
10
to
11,
page
doc,
pages
pages
so
yeah,
so
I
created
our
usable
component
for
it.
H
Yeah,
so
this
is
the
component
which
was
created
and
this
co.
This
was
replaced
in
other
files
as
well.
So
you
could
see
that
network
connection.
Is
there
here
yeah?
So
this
this
decreases
the
code
size
in
other
pages
as
well,
and
the
other
things
which
I
changed
in
almost
every
file
was
that
I
felt
this
contributing
get
flow
and
these
contributing
ui
and
contributing
guidelines
which
were
having
a
parent
page,
which
was
contributing.
So
I
change.
I
change
the
link
of
every
page
because
it
will
it
will.
H
It
will
indicate
that
contributing
will
be
the
parent
that
this
link
is
the
parent
link
for
this.
This
link
so
yeah.
I
changed
that
as
well
yeah,
the
that
the
pr
was
merged.
I
just
wanted
to
show
that
these
were
the
changes
that
I
made.
D
D
H
You're,
if
you're
open
it's
a
parent
page,
because
these
like
and
like
when
we
come
on
this
page
like
this,
so
that
there's
like,
if
we
consider
it
as
our
tree,
then
like
this
page,
the
backwards.
Yes.
D
No,
no,
no
yeah,
no,
no,
no
yep!
You
don't
even
have
to
explain
it
makes
intuitive
sense.
I
agree,
but
for
my
part
I
agree
yeah,
that's
great!
I'm
just
trying
to
be
brief
and
yeah.
Thank
you.
So
much
for
the
suggested
reading,
like
yeah,
hey,
where
aditya
can
tell
you
all
about
not
repeating
ourselves.
He
did
a
ton
of
work
within
mastery,
ctl
command
guide
to
make
it
highly
dynamic,
and
so
what
you're
doing
with
the
include
template
for
suggested?
D
Reading,
beautiful,
like
yeah,
basically
at
the
bottom
of
every
page,
we
should
probably
have
suggested
reading
like
that
suggested
reading,
include,
should
probably
be
in
in
one
of
the
layouts
like
you
might
want
to
add
it
to
a
layout.
Now,
the
and
you
you
will
should
and
will
have
success
using
liquid
templating
using
built-in
jekyll.
D
H
D
H
Think
dynamically
it
is
like
decided
like
which
place
we
have
page.
We
are
on
so
like
we
don't
have
to
like
pass
the
variables
as
such,
but
like
in
in
this
code,
like
we
have
like
in
this.
H
This
sorted
reading
variable
like
this
sorted,
the
all
the
pages
and,
like
this
item
type
like
when
we
have
a
page
or
type
as
as
contributing
and
page
or
an
item
to
category,
so
it
it
sorted,
sort
all
the
pages
and
like
filter
out
all
the
pages
itself.
According
to
the
page,
we
are
on.
D
Okay,
so
the
maybe
the
only
quality
feedback
that
I
have
for
my
part
that
I
have
for
you
is
this
looks
like
a
darn
fine
job,
probably
worthy
of
including
this
snippet.
This
include
in
one
or
more
of
the
layouts
so
that
it
becomes
universal
to
most
of
the
pages,
not
all
the
pages,
and
that's
why
it
would
it's
only
applicable
to
some
of
the
layouts
or
one
of
the
layouts.
H
Okay,
okay
and
yeah-
I
I
just
wanted
to
like
discuss
one
other
thing
which
was
like
in
in
this.
In
all
these
pages
like
in
this
toc,
basically
table
of
content
like
we
have
a
reputation
of
links
right
like
like.
If
we
are
on
this
this
page
also
like
this
restituting,
we
have
here
as
well
and
then
contributing
to
mastery
docs
in
here,
then
we
have
here
as
well,
so.
D
You
know
what
let
me
just
wait.
Sometime
totally,
I
didn't
get
a
chance
to
comment
on
the
comment
on
your
comment
on
my
comments,
so
I
didn't
get
a
chance
to
comment
back
on
the
comment.
I'm
just
trying
to
say
comment
as
many
times,
and
that
makes
total
sense.
That's
a
great
you
did
a
good
job
yep
we
should
do.
We
should
merge
your
pr
to
remove
the
duplicative
table
of
contents.
H
Yeah
yeah,
so
this
I
just
made
a
single
line
change,
because
the
toc
toc
component
was
included
like
two
times
in
the
in
the
page
layout
as
well
and
as
well
in
the
default
layout,
and
the
page
layout
inherits
default
like
page
layout
healthy.
H
Let
me
show
you
yeah.
So
page
layout
includes
the
default
layout.
So
when
this
particular
talk
table
of
content
was
included
in
page
and
as
well
as
it
was
included
in
included
in
default
as
well.
So
I
removed
this.
I
just
made
a
single
line
change
in
in
which
toc
was
removed
in
page,
so
that
only
the
poc,
which
was
present
and
default
layout
only
that
would
be
seen
yeah.
So
this
this
guy
is
not
merged.
Yet
yeah.
D
D
Schweiz,
do
you
mind
we
need
to
get
some
updated
docs
for
the
app
mesh
adapter.
So
by
the
way,
all
of
you
that
are
on
the
call
there's
a
gentleman
theodore
salvo,
he
joined
slack
a
couple
days
ago,
he's
from
aws.
He
focuses
he's
a.
D
D
You're
all
welcome
to
come.
Let
me
know
if
you're
going
to
be
up
at
that
hour,
that
would
be
4
p.m.
Pacific
and
and
we'll
get
you
the
info
they're
just
we'd
spoken
to
them
like
a
year
and
a
half
ago,
but
mesri
is
in
a
much
different
place
now
and
so
they're,
just
I
think,
they're
interested
to
see
what's
going
on
and
so
it'll
be
a
good
time
to
show
off
the
new
now
beta
adapter
for
app
mesh.
D
D
A
That's
all
for
today,
we'll
see
you
next
week
and
if
you
are
hanging
back
to
the
nerd
party,
feel
free
to
do
so
and
we'll
be
talking
about
helm
and
mystery
operators
bye.
Everyone
bye,
bye,.
D
D
D
Yeah,
no,
no,
I
think
so.
No,
I
was
having
a
hard
time.
I
think.
Do
you
in
your
mind,
is
the
potential
crux
of
the
current
issue
potentially
fixed
by
changing
the
logic,
the
the
the
determination
by
which
mesh
reserver
figures
out
if
the
operator
is
deployed
or
not.
G
It
would
be
fixed,
partly
the
other
issue
would
be
like
when
users
would
actually
try
to
delete
it
from
the
ui.
They'll
have
trouble
like
when
mystery
server
would
try
to
interact
with
it
like
because
it
doesn't
matter
if,
in
that
case,
it's
installed
or
not,
it's
just
an
operation,
and
that
would
happen
so
it
would
fail.
D
Yeah
yeah
wait.
No,
I
mean
you
get
it
yeah
on
that
that
piece
that
you
didn't
fully
describe
the
situation
like
so
if
yeah
users
that
go
to
the
ui,
if
measuring,
if
the
operator,
if
okay,
if
the
op,
if
mastery
server
is
connected
to
kubernetes,
but
for
whatever
reason
the
operator
is
not
installed.
D
That
won't
be
a
problem
right.
The
people
will
click
to
install
it
and
you'll
run
okay.
If
it
is
installed.
Let's
say
that
you
know
here
shortly:
people
are
going
to
be
able
to
connect
mesh
reserver
to
four
or
five
kubernetes
clusters,
so
they're
in
a
situation
where
they've
got
like
five
of
them,
they're
tearing
down
one
of
the
clusters,
or
rather
they
don't
want
to
run
service
mesh
stuff
on
it.
So
they
don't
want
mescheri
talking
to
that
cluster
anymore.
G
Yes,
this
pr
right
now,
doesn't
I
mean
when
you
switch
context,
it
would
delete
all
of
those
crs
and
all
that
stuff,
okay
and
I
actually
got
a
little
little
lost.
D
In
what
I
just
said,
yeah,
oh
yeah,
no,
you
actually
that's
good.
The
beh,
that's
yeah
I'll
be
more
concise,
which
is
yeah
when
a
user
switches
contexts
when
he,
when
I'm
sorry
when
a
user
is
in
mesh
for
ui
and
they
take
a
kubernetes
chip
and
disconnect
it.
They
delete
the
connection.
D
G
G
D
And
that's
because
I
recognize
that
you're
talking
about
helm's
special
construct
of
a
release
so,
but
but
the
reason
that
the
two
really
helm
releases
would
be
considered
different
is
because
we're
using
like
we
still
only
have
one
set
of.
We
only
have
one
chart
with
many
subcharts
currently
right.
G
D
D
Have
it
deploying
the
full
helm
chart
you're,
so
you're
selectively
using
those
sub
charts
to
deploy
just
to
the
subcharts,
but
wouldn't
it
isn't
it
true?
You're
on
the
witness
stand
now
like?
Isn't
it
true
that
you
could
just
use
overrides
to
you'd,
still
install
the
same
high-level
helm,
chart
that
has
the
server
and
all
the
adapters
and
all
the
stuff
you
don't
want,
but
you
would
use
an
override
to
specify
that
you
only
wanted
the
operator
and
that
that
deployment
wouldn't
be
in
a
separate
release,
helm
release
but
would
be
the
same
deployment.
D
D
D
Track
it
talks
about
tracking
the
kubernetes
cluster
that
mesh
reservoir
is
installed
on
and
distinguishing
that
from
all
of
the
other
kubernetes
clusters
that
might
be
under
management,
and
if
we
track
that,
then
you
can
then
treat
the
release
the
help
you,
then
you
can
have
different
logic
by
which
you're
managing
helm
releases,
your
helm,
chart
install
uninstall,
because
you
know
right
like
that.
That
would
be
one
way
to
circumvent
your
issue.
D
Right
because
your
issue
right,
thank
you,
bark
cause
your
issue
right
now,
rather,
does
that
make
sense
or
no.
D
But
when
he
comes
back,
okay,.
C
D
D
Okay,
yeah
so
a
little
while
ago,
a
little
while
ag.
Can
you
still
hear
me.
D
Navindy
and
I
spoke
about
how
to
augment
mesh
config,
so
that-
and
I
don't
know
that
this
is
appropriate
to
to
do
what,
like
I
don't
know
that
we
should
go
off
and
do
this
in
measuring
ctl.
This
is
just
a
design
proposal
at
the
time.
Maybe
it
makes
sense.
D
You
know
ongoing
ion,
but
but
one
of
the
things
that
we
had
were
forced
to
try
to
figure
out
inside
of
this
proposal
was
to
to
create
a
qualification
about
the
the
types
of
kubernetes
clusters
that
mesherie
is
interacting
with
and
we
used
we
just
made
up
some
terminology.
It
was
better
than
george
or
I
used
rebecca
this
morning
for
something
else.
We
actually
so
we
actually
gave
these
things
a
name.
D
The
meschery
kubernetes
cluster
is
the
one
in
which
measure
server
is
deployed
and
messaging
operator
is
deployed
as
well,
because
the
operator
gets
university
deployed
across
everything,
and
so,
since
you
have
a
fork
in
the
road
here
about
how
to
treat
helm
deployments
of
the
operator
different
between
when
mastery
server
is
deployed
and
when
it
isn't,
if
we
began
to
this
is
one
of
the
solutions
is
to
just
track.
D
An
attribute
of
that
kubernetes
connection
and
to
understand
is
measuring
server
running
in
this
cluster.
The.
If
so,
go
this
way.
If
not
go
that
way,.
F
D
D
Like
it's
funny,
I
had
I
went
round
and
round
on
with
abhishek
kumar
in
this
in
the
past.
One
of
the
things
that
what
hussain
is
pointing
out
is
like
hey
shouldn't.
We
just
have
isn't
it
less
complex
when
you
just
have
a
single
manager
for
a
component,
you
know
a
single
manager
controls
that
thing
and
so
right
now,
for
the
most
part,
it's
mestre
ctl
is
the
controller
of
mesh
measuring
and
all
of
its
components.
D
F
D
The
reason
the
reason
that
measuring
that
measure
ctrl
is
deploying
the
operator
is
because
we,
because
we
are
so
desirous
of
the
user,
experience
being
that
by
the
time
somebody
gets
in
to
the
ui
that
the
operator
is
up
and
mesh
sync
is
reporting,
hopefully,
and
that's
a
great
experience,
because
it
opened
up
the
dashboard
for
the
first
time.
It's
like
oh
wow.
This
is
an
intelligent
system
like
it
understands
what's
going
on
here
and
if
not
it's
like
okay,
they
go
in
or
you
know,
and
if
not
it's
like.
D
So
rudraksh
the
thing
that
we
just
described
about
treating
the
clusters
differently:
oops,
we
just
lost
them
okay
anyway,
I
don't
think
it's
going
to
work
out
for
him
today.
Unfortunately,
he's
going
to
be
out
tomorrow,
his
dad's
having
surgery
and
he'll
be
out
for
a
couple
of
days,
which
is
going
to
be
unfortunate
because
it's
going
to
impact
the
dot
6
release.
This
isn't
the
only
way
to
skin
this
cat,
so
husseina
just
presented
one
that
would
fix
the
issue,
there's
no
more
fork
in
the
road
there's
just
one
road.
D
D
Just
yeah,
I
think
three
three
solutions
were
are
presented,
husseina
pointed
out
one
which
is
hey,
stop
having
mastery
ctl
and
mesh
reservoir
managing
operator
deployments.
Just
have
one
of
them.
Do
it
then
there's
no!
D
That
reduces
the
conflict.
There
is
no
conflict.
You
have
one
manager,
okay,
so
rudraksha
rudraksha,
I'm
gonna
record
my
voice
and
send
it
to
you,
because
I
don't
have
confidence
that
you're
gonna.
Your
connection
will
last.
D
And
okay
cool
well
I'll
catch.
F
D
Yeah,
nice
nice
to
see
you
hussaina
by
the
way,
by
the
way,
hussain,
the
gal,
whose
first
name
starts
with
the
d,
and
I
think
it
was
deepani.
F
G
D
Okay,
yeah!
No
thanks,
yeah
wait,
wait
one
other
quick
question
on
the
the
those
pesky
sdks
like
like
in
some
respects.
It
would
be
nice
to
drop
support
for
deploying
mastery
on
docker,
but
we've
put
so
much
into
it
and
it
works
mostly
okay.
Now,
anyway,
I
guess
it's
still
an
open
consideration
for
us
about.
F
Yeah,
I
think
people
who
are
familiar
with
docker
technologies.
I
think
it
would
still
be
an
option
to
try
out,
but
then
again,
anyways
kubernetes
is
also
created.
D
Right
the
thing
things
like
docker
desktop
comes
with
kubernetes,
so
so
that's
good
like
it
is:
it's
feasible.
I've
spoken
to
docker
cto
a
couple
of
times
and
they've
looked
at
measuring
and
one
of
the
ways
in
which
people
can
view
measuring
and
they
get
value
from
sharing
is
well.
D
Is
this
if
service
meshes
become
or
I'll
say
it
like
this
kubernetes
became
ubiquitous,
and
so,
if
you're,
developing
cloud
native
applications-
and
you
want
to
have
every
developer
needs
to
have
their
own
kubernetes
cluster,
like
oh
great
docker,
desktop,
is
a
great
solution
for
that.
You
get
a
local
one.
D
D
Okay,
well
wow
nice
to
see
you.
It
was
thanks
for
staying
up
late
too.
It
was
nice
to
see
you
on
the
on
at
your
cubecon
talk.
F
Thank
you.
Actually,
I
was,
I
mean
I
had
the
mystery
up
and
running
and
all
that,
but
I
believe
I
get
these
issues
with
sharing
screens
with
google
meat
as
well
as
I
I
was
not
sure
what
platform
it
was.
So
it
was
not
letting
me
share
my
screen.
D
No,
but
and
then
also
like
any
of
that,
any
of
us
not
being
as
smooth
as
we
possibly
could
like.
That's
probably,
you
know,
like
my
fault,
for
not
getting
us
coordinated
and
like
what
are
we
even
doing
and
like
can
we,
you
know
test
out
the
system
and
like
yeah
again
all
that
was
so
late
coming
that
that
that's
on
me
so.
D
Yeah
there
was
a
couple
of
people
that
jumped
in
to
the
project
now
based
on
having
joined
yeah.
F
Maybe
his
name
is
start
with
d
or
okay
yeah.
D
Yeah,
I
think
you're
right.
It's
like
I
think
his
first
name
is
deepak
and
then
he
calls
himself,
mr
wick,
or
something
like
that.
Oh
okay,.
E
D
F
D
Well,
fair
enough,
I'm
gonna!
I'm
gonna
go
get
this.
D
F
G
D
Sure,
well,
you
know
if
they're
gonna
give
you
the
holiday,
like
you
know,
yeah!
Oh
that's
cool!
I
that
you
know
that
reminds
me.
I've
got
to
go,
get
some
candy
for
the
trick-or-treaters
for
the.