►
From YouTube: Meshery Development Meeting (Feb 24th, 2021)
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
No,
no,
no!
Nobody,
just
you
guys,
just
making
me
look
like
the
narcissist
here
just
at
when
I'm
just
it's
just
out
of
friendliness.
That's
all
there
is
that's
a
good
fine.
Looking
young
man
right
there.
A
All
right,
very
good,
everyone
feeling
warmly
embraced
freshly
harassed,
hopefully
this
morning.
Well,
it's
wednesday
the
24th.
We
are
what
we
got.
We've
got.
Some
topics
lined
up
we're
like
yeah,
close
to
the
dot
five
release.
A
As
a
matter
of
fact,
unless
there's
significant
pushback
on
this
call,
I
will
I'll
push
for
us
to
make
it
this
week
to
like
acknowledge,
maybe
there's
a
bug
or
two,
but
that
we're
young
in
our
releases
and
that
we
need
to
get
it
out
there
there's
just
a
lot
inside
that
release,
so
a
lot
to
get
into
people's
hands.
A
Okay
call
for
topics
before
we
get
rolling
heavily
here,
if,
if
you're
on
the
call
slap
your
name
in
on
the
list,
otherwise
it
doesn't
count.
Even
though
you
were
here.
C
So
abhishek
said
there
was
something
regarding
the
image
hub
set
up.
I'm
not
too
sure
exactly
what
it
was
about,
but
I
thought
he
you
and
him
were
going
to
discuss
so
yeah.
A
I'll
just
put
your
name
down
next
to
that
we'll
just
good
yeah.
No
thanks
for
bringing
it
up.
It
should
be
on
the
list.
A
I
think
I
recall
actually
there's
one
to
touch
on
with
with
michael
and
maybe
some
others
and
it's
about.
A
A
A
She
may
not
have
been
able
to
join
the
call
just
yet,
but
she'd
been
okay,
good,
so
she's
not
she's,
not
here,
yet
all
right,
fair
enough.
What
do
we
have
first,
so
a
little
bit
about
what
I
was
just
articulating
any.
This
is
probably
our
most
timely
and
critical
item
to
chat
through
so
so
we
you
know,
I'm
not
sure.
I
don't
quite
recall
what
call
we
were
in.
What
meeting
we
were
in
where
we're
kind
of
discussing.
A
How
do
we
begin
to
harden
releases?
How
do
we
up
the
number
of
unit
tests
that
we've
got?
How
do
we
up
the
number
of
integration
tests
that
we
have
we've
you
know
rodolfo
has
is
here
he's
leading
the
continuous
integration
meeting
that
working
group
that
meets
bi-weekly?
A
So
I
believe,
I
believe
a
meeting
is
tomorrow.
We'll
ask
him
for
an
update
in
friday's
meeting
in
the
community
meeting.
One
of
the
things
that
that
has
been
a
call
to
action
in
there
has
been
well
a
test
strategy
and
a
test
plan,
a
document
that
would
highlight
how
we
approach,
testing
and
philosophy
some
philosophy,
some
tooling
around
that
and
then
specifically
an
itemized
test
plan,
test
cases
that
we
would
intend
to
verify
for
regression.
Each
time
that
we
go
through
a
release.
A
So
there's
some
of
you
who
are
on
the
call
who
have
a
lot
of
experience
with
that
kind
of
a
thing,
so
this
is
entirely
open
to
your
influence.
There's
some
of
you
who
are
on
the
call
who
don't
know
what
I'm
talking
about
and
it's
an
excellent
opportunity
for
you
to
learn,
or
you
have
a
sense
of
what
I'm
talking
about,
but
don't
have
any
suggestions
as
to
the
particulars
like
what
should
be
getting
gating
criteria
for
when
a
release
goes
out.
When
should
we
allow
ourselves
to
override
that?
A
How
do
we
maintain
a
good
implicit
contract
with
the
user
with
those
that
might
be
consuming
our
apis?
A
A
Okay,
wow
so
pulling
it
back
to
I'm
I'm
proposing
that
we
do
a
5
release
this
week.
For
me,
it
is,
it
is
time
we
have
got
like
six
different,
significant
architectural
components
going
into
this
release.
It
was
like
time
a
while
ago,
we
needed
to
begin
to
get
some
of
these
components
out.
A
Think
about
this
for
a
little
bit,
if
you
would
we'll
we
can
come
back
to
this
topic
if
there
are
really
what
we're
trying
to
assess
here
is
whether
or
not
there's
any
egregious
sort
of
outstanding
broken
pieces
you
know
of
of
measuring
before
we
can
go
announce
to
the
world
that
they
should
come
check
it
out.
A
A
Maybe
I'm
you
know,
exaggerating
that
a
little
bit
with
the
word
starved,
but
but
we
need
the
feedback
and
you
do
as
well.
I
know
it
feels
really
good
to
see
people
using
your
software
and
they
are
there's
been
so
so
husseyna
saved
the
day
on
monday,
the
istio
con
is
going
on
this
week.
On
monday,
we
did
a
workshop
on
istio.
We
used
meshri
to
teach
people
about
istio.
A
A
D
Yeah,
basically,
a
mystery
ctrl
system
config
for
all
the
providers.
They
use
different
type
of
scripts
to
generate
the
cube,
config
dot
ml
file,
so
in
case
of
azure
and
eks.
These
are
two
simple
steps
which
azure
and
aws
cli
are
readily
having,
so
they
they
give
out
the
config
in
one
single
command,
but
in
case
of
gke
it's
a
series
of
steps.
First,
you
get
the
service
account
name
then
get
some
tokens
and
then
get
the
csr,
and
so
it's
set
of
steps
bit
complex.
D
So
what
we
want
to
do
going
forward
is
the
very
first
issue
we
encounter
with
this
type
of
design
is
that
the
os
specific
parts
were
not
being
considered
so
wherever
we
put
cubeconfig.yml,
that
was
failing
if
users
were
running
messy
ctl
on
a
windows
based
system.
So
I
started
out
with
that.
But
then
later
I
had
moved
some
code
to
the
config.go
and
I
think
we
should
overall
move
away
from
the
scripts
concept
and
write
everything
in
the
go
line.
So
there
is
mesh
sheet
available
which
start.go
and
other
cities
are
using.
A
Along
those
same
lines,
docker
is,
is
writing
docker
compose
they're,
rewriting
it
in
golang,
so
if
you're
unfamiliar-
and
it
may
not
be
obvious
that
you
know
docker
and
docker,
engine
golang
in
written
and
goling
docker
compose
is
actually
a
separate
binary.
It's
been
written
in
python
and
is
being
rewritten
in
golang.
A
A
So
along
those
same
lines
as
well
as
to
what
hussain
is
saying,
is
to
move
away
from
that
and
use
that
golang,
the
docker
compose
goaling
package
natively
to
be
able
to
more
to
one
eliminate
a
dependency
that
the
user
has
to
bring
a
prereq
that
they
have
to
have
installed
on
the
system.
You
just
bring
that
packaged
in
have
tighter
control
over
error
handling
and
the
user
experience
for
the
user.
A
You
know,
speaking
of
scripts
hussein
since
you're,
not
you're,
not
busy
message
the
I
don't
know
if
nikhil
is
on,
but
he
was
highlighting.
A
You
know
the
he
was
highlighting
some
challenges
around
upgrading
mastery
ctl
and
you
know
I
think
it's
known
at
least
to
you
know
for
most
of
us
that
if
you
run
a
meschery
ctl
system
update
the
mesheri
containers
so
mesh
reservoir,
the
adapters,
the
database,
the
load,
generators,
etc.
They'll
they'll
be
updated
because
they're
they're
run
inside
of
containers,
so
we'll
measure
will
update
those
containers
pull
down
the
latest
images
it
today
that
command
does
not
update
mastery
ctl
itself.
It
can
it's
just.
A
D
D
I
do
not
see
any
issue
with
locally
running
mystery.
Cpl.
D
If
I
may
be,
the
issue
is
when
we
download
the
machine
script
and
try
to
start
from
there.
I
was
directly
running
mrtl
system
update
yeah.
There
could
be
differences.
A
Hussein,
if
you
don't
mind-
and
I
think
this
is-
I
don't
know
that
many
of
us
have
read
this
particular
set
of
considerations
but
to
to
fully
wrap
your
mind
around
it.
Who's
saying
it
have
a
give
this
a
read:
if
you
would
it
talks
about
individual
components,
that
measure
is
made
up
of
and
and
considerations
to
account
for
when
you
go
to
update
them.
This
is
from
a
user's
perspective,
but
it
will
spell
it
out.
A
Okay,
yeah
nice,
okay,
good
by
the
way
so
of
the
individual
who
raymond
wong
was
his
name.
That
was
having
a
challenge
with
using
gke
and
mescheri.
He
provided
a
lot
of
logs
and
details
and
seemed
to
be
very
much
so
on
hand
to
help
troubleshoot.
D
The
mastery
cdl.log
does
not
capture
any
errors
from
the
config
script,
so
I
was
not
able
to
make
much
progress
on
which
exact
step
went
wrong
in
the
gke
configuration,
maybe
I'll
sync
up
with
you
offline
or
if
you
can
provide
his
contact
details.
Maybe
I'll
talk
to
him
directly
guys
beautiful.
A
A
Yes
to
both
I'd
like
to
get
you
in
contact
with
him
directly
he's
I'll
do
just
that
you're,
not
in
the
istio
slack.
Just
yet
I
suspect,
and
so
I'll
get
you
an
invite
and
get
you
guys
introduced.
D
A
Please
yeah
I'll
even
make
an
introduction
or
abhishek,
or
it
would
cost
you
if
you
guys
want
to
make
an
introduction.
That
would
be
great
sure.
A
G
Okay,
and
so
I
think
we
have
already
mode
what
patterns
were
so
and
that,
but
that
was
actually
machine.
Video
based,
so
we've
been
working
on
a
ui
interface
for
parents,
and
now
there
are
a
few
crud
operations
that
are
available
on
patents
through
the
ui.
So
that's
what
the
demo
is
about.
G
So
this
is
the
name
section,
I'm
audible
right
now.
G
Yep
so
yeah,
this
is
a
new
section
in
machine
one
here
you
can
see
this
pattern
section.
What
you
can
do
this
is.
This
is
a
basic.
Basically,
this
table
will
allow
you
to
to
basic
thread
on
your
pattern
files.
So
here
I'm
uploading
a
pattern
file
called
this
is
the
only
demo
you
can.
G
Basically,
this
is
your
pattern
file.
You
can
change,
you
can
update
it.
Maybe
do
9.4
as
that
is
available
now
or
so
yeah
and
delete
it.
This
is
right
now
right
now.
What's
what
we
are
offering?
What
what
I
am
working
on?
The
ui
is
the
basic
crud,
but
soon
you
would
be
able
to
do
much
more
with
patterns
right
from
the
u1.
G
If
you
want
to-
or
else
visual
details
is
obviously
available-
and
this
is
the
delete
operation
so
yeah
in
this
repository,
so
if
you're
using
a
remote
for
it,
it's
a
regular.
G
So
this
is
a
fake
demo.
Does
anyone
have
any
suggestions
that
the
ui
is
definitely
going
to
change
so
right
now,
it's
kind
of
like
we
are
not
allowing
to
mention
our
partner
file
name.
I
mean
right
now
what
what
is
happening.
Your
final
file
name
is
getting
pitched
from
the
file
name,
that
was
in
your
file
system,
but
soon
it
will
be
like
that.
G
This,
the
pattern
name
that
you
have
specified
in
here
that
would
be
actually
that
will
actually
become
the
name
of
the
pattern
and
obviously
you'll
be
able
to
provide
a
displace
like
a
different
display
name
as
you
would
like.
So
those
are
some
ui
changes
that
are
in
line
and
definitely
will
be
available.
G
A
You
know
I've
worked
with
some
bright
architects,
and
actually
this
isn't
the
comment
toward
utkarsh.
I
think
we've
already
shown
shown
him
him
favor
in
terms
of
some
compliments,
so
he
doesn't
need
any
more
compliments.
That's
not
actually
what
I
was
going
to
say.
I've
worked
with
some
bright
arcade,
so
we
don't
need
to
build
up
utkarsh
anymore.
Let's
just
be
clear
on
that.
A
I
work
with
some
bright
architects,
some
bright
engineers.
You
know
I'm
looking
at
a
few
of
them
here
on
this
call.
A
And
some
of
them
that
I've
worked
with
or
they
don't
they
don't
leave
any
room
for
anyone
else
to
talk
good
lord,
are
they
smart
people,
but
they
won't
boy?
They
just
get
such
a
tight
grip
over
the
project
over
the
thing
that,
like
nobody
else,
can
breathe
nobody
else's
ideas,
get
a
chance
to
service
and
sort
of
bubble.
You
know
percolate
sort
of
bubble
up
to
the
top.
A
A
E
You
know,
patterns
patterns
are
always
good.
Everybody
loves
patterns
and
everybody
knows
sort
of
what
or
has
a
an
expectation
or
sort
of
what
a
pattern
is
about.
A
Totally
yeah
yeah
right,
some
of
those
things
you
might
want
to
say,
are
like
hey.
You
should
use
this
when
there's
some
or
you
should
not
use
this
one.
E
Yes,
for
instance,
or
because
patterns,
can
you
know
they
can
be?
Quite
I
expect
they.
These
yaml
files
can
be
quite
big
and
then,
in
order
to
understand
what's
happening,
would
be
good
to
have
some
some
sort
of
verbose
description.
What
what's
actually
in
there.
You
know
this
is
sort
of
installing
this,
and
that
and
setting
up
your
observability
plane
and
and
yeah.
It's
not
it's
not
secured,
or
it's
secured,
or
it's
production
ready
or
it's
demonstrating
this
or
that.
So
I
think
that
will
be
good.
E
A
You
know
what
a
I've
talked
to
a
a
logo,
an
icon,
a
small
visual
diagram
of
would
help
like,
or
even
if
you
just
showed,
the
istio
logo
and
the
prometheus
logo
or
what
you
know
like.
A
H
Yes,
just
a
thought
that
popped
up,
basically,
as
as
we
define
the
ammo,
is
it
is
it
better?
Is
it
a
better
experience
that
we
show
than
the
dry
run,
not
the
dryer,
the
underlying
logic
or
the
underlying
result
that
it
would
produce?
For
example,
this
particular
pattern
has
this
yaml
file
defined
and
in
a
separate
view,
or
in
a
view
that
beside
it,
we
can
show
them
the
steps
of
what
all
are
going
to
happen.
When
you
apply
this
pattern,.
H
In
this
particular
pattern,
http
gets
installed
or
graffana
gets
installed
from
this
html.
So
so
all
of
these
details
can
be
put
up
for
one
validation
purpose
so
that
the
user
doesn't
make
any
mistake
or
the
user
is
sure
that
he's
applying
the
right
pattern
and
and
to
that,
it's
it's,
it's
always
good
to
know.
What's
going
to
happen
like
it's
always
good
to
know
the
outcome
of
your
input.
A
And
well,
both
michael
and
really
speaking
toward
the
same
thing,
both
both
having
it
from
my
perspective,
like
an
excellent
suggestion
on
on
a
way
of
I
mean
it
can
be,
you
know
really
rather
scary
if
you've
got
a
long,
a
long
manifest,
especially
if
you've
got
a
running.
Oh
thank
you.
Hussaina,
we'll
get
you
introduced.
A
If
you
have
a
long
existing
surface
matching
existing
things
going
on
you're
going
to
go,
apply
this
pattern,
you're,
trusting
that
it's
well
prescribed
and
you're
going
to
take
that
medicine
like
it'd,
be
nice
to
know
what
those
ingredients
in
the
medicine
are
you
allergic
to
any
of
those,
or
you
know,
and
the
thing
is,
is
in
a
beautiful
way
showing
someone
that
dag
showing
them
the
lends
itself
to
opens
us
up
to
the
world
of
a
workflow
engine
and
the
world
of
people
wanting
to
manipulate
those.
A
Like
it's
great
that
I
can
see
them.
Oh,
I
can
see
them
in
a
static
way.
Okay,
I
I
click
apply.
Oh,
I
can
see
it
processing
that,
oh
it's
it's
almost
there.
Okay,
great!
Oh
nope!
It
had
an
issue
on
fully
applying
the
pattern
and
it
what's
it
going
to
do.
Oh
would
I
well.
I
would
prefer
that
it
rolls
back.
I
would
prefer
that
it
restarts.
A
I
would
prefer
to
just
stop
and
just
leave
what
you
did
just
stop
right
there
like.
Okay,
oh
actually,
no,
you
know
what
I
would
prefer
to
change
this
around
and
rejigger
the
dependencies
like
all
of
a
sudden
you're
facilitating
workflow
design.
You've
got
a.
What
do
you
have?
We
have
a
service
mesh
designer
a
configurator,
we'll
shoot.
We've
got
that
in
the
ammo
right
here.
We
just
don't
have
that
visually.
Okay,.
G
A
Yep
totally
so,
from
my
perspective,
like
the
a
quick
step,
is
to
yes
to
potentially
go
ahead
and
expose
a
column
that
has
any
number
of
logos,
of
the
components
that
would
be
that
are
being
touched
or
being
installed
or
being
manipulated
in
the
pattern
and
that
I
don't
know
that
it's
obvious
to
everyone.
I
think
this
has
been
shown
in
the
somewhere.
A
We
should
probably
show
it
again,
but
these
patterns,
you
know
another
way
of
ref.
So
another
way
of
referring
to
these
would
be.
Is
you
know
more
or
less
your?
It's
an
environment,
config,
it's
a
service,
mesh
config
it
the
thing
about
the
patterns.
Is
they
take
it
further?
They
they
can
also
describe
something
that
you
would
like
to
happen
over
time
and
maybe
repetitively
over
time.
In
this
particular
pattern,
it's
a
roll
out
pattern.
So
it's
it's
not
just
it's
not
static.
A
You
know,
anyway,
but
to
the
point
about
showing
a
visual,
dag
or
a
visualization
of
what
all
is
in
here
is
that
mesh
map
is
on
the
cusp
of
being
capable
of
interpreting
this
same
yaml.
As
a
matter
of
fact,
it
is
capable
of
interpreting
this
yaml
there's
a
measuring
measure
will
have
the
ability
to
take
and
convert.
A
What's
an
an
open
application
model
compatible
specification,
this
one
and
convert
it
to
a
few
different
things,
translate
it
over
to
kubernetes
manifests,
of
which
this
isn't
too
far
removed
from
it.
But
it
is
this
won't
work.
If
you
go
apply
it
to
kubernetes,
it
needs
to
be
changed,
so
mesher
is
capable
of
well
measuring.
Adapters
are
capable
of
doing
that.
A
A
Have
it
to
do
that
programmatically,
maybe
in
your
pipeline
or
or
in
response
to
something
else,
going
on
like
in
response
to
something
else
going
on,
like
maybe
it's
time
to
roll
out
the
next
version
of
of
a
workload
yeah,
the
the
one
of
the
sweet
things
about
this
particular
format,
this
pattern
file
format-
is
it
couldn't
be
more
succinct?
A
I
don't
think
like
it's
about
as
high
level
as
you
can
get
it's
about
as
short
as
you
can
get,
which
I
think
lends
itself
to
popularity
potentially
remember.
A
I
think
that
there's
some
merit
here
to
michael's
point
is
we
go
to
put
an
effort
to
service
mesh
patterns
at
large
and
and
by
the
way,
I
think
I
shared
with
all
of
you
that
the
the
book
called
service
mesh
patterns
it's
going
into
early
release
now,
so
I
in
about
a
week
it
should
be
out
and
available
in
early
release
and
we'll
just
have
a
couple
of
chapters
and
one
of
the
chapters
just
is
just
all
about
measuring
just
talks
about
getting
mastery
up
and
installed
and
it'll
be
the
tool
that
you
use
to.
A
You
know
interface
with
all
these
patterns,
and
so
so
to
michael's
point
like
there's
a
lot
else
that
is
written
about
the
metadata
around
around
the
patterns
like,
and
so
one
of
the
things
that
I
would
suggest
that
that
we
all
think
on
is
well
get
like
giving
a
unique
identity
to
the
pattern
so
that
you
know
it's,
it
can
be
kind
of
referenced.
Like
hey,
I'm
going
to
use
pattern
5009
as
my
starting
point
for
a
canary
rollout,
but
you
know
my
environment's
a
little
different,
so
I'm
gonna,
I'm
gonna
modify
it.
A
I'm
gonna
or
like
there's
these
other
variants,
but
maybe
I'd
like
to
know
hey
what
else
is
what
else
should
I
be
considering
about
this
thing
when
when
is
it
really
good
to
use
or
not?
Well,
is
anybody
else
running
this
and
if
you
search
for
that
code,
pattern
service,
mesh
pattern,
5009,
hopefully
you'll-
find
you'll
find
a
book
on
it
or
you,
but
you'll
also
find
other
practitioners
out
there
saying.
A
Well,
you
know
we
came
out
with
5010
because
such
and
such
you
know,
nine
didn't
cover
it
for
us
and
so
there's
a
new
one,
but-
and
so,
if
you
were
sharing
that
here,
you
go.
If
you
want
to,
you
know,
use
the
new
and
improved
pattern.
A
There's
a
repo
there's,
a
github
org
called
surface
mesh
patterns
and
a
repo
service
mesh
patterns.
I
think
it's
closed
at
the
moment,
mostly
because
it's
doesn't
have
much,
but
the
intention
is
for
it
to
have
these
and
for
it
to
be
open
and
maintained
by
others
are
contributed
to
by
others,
and
all
of
you
are
sitting
in
the
most
prominent
spot
to
to
to
make
a
mark
on
that
to
have
your
name
on
it
have
your
name
against
patterns,
so
so
we
should
get
that
we
should
I'll
drop
a
link
into
the
meeting.
A
That's
and
just
I
don't
know
that
that
there's
anything
there
to
look
at,
but
but
if
you're
interested
then
we'll
I'll
we'll
get
you
access
thanks.
I
thought
I
thought
we
were
gonna.
Do
like
a
six
minute
thing,
so
this
is
good
any
other
back
for
crash
comments
on
this.
A
All
right,
all
right
so
anish,
I
think
you're
up
mastery,
ctl
tokens
and
kubernetes
support.
I
Hello
regarding
token,
like
we
have
actually
updated
the
conflict
to
store
the
token
the
way
it
was
meant
to
be,
as
per
in
the
machine
cli
reference
documentation.
I
I
can
not
actually
show
you
the
current
platform
support
for
google.
It
is
because
I
have
not
been
able
to
you
know
completely
run
it
properly
on
my
laptop,
so
yeah
sort
of
like,
and
regarding
that
token
too,
so
I
have
to
outside
that
too.
So,
first
time
working
on
it
so
like
I
wouldn't
like
what
is
that
token.
A
So
to
answer
your
question,
I
need
the
mastery
ctl
as
a
client
of
mesherie
server,
some
of
the
rest
apis
that
mastery
ctl
invokes
on
meshfree
server.
A
Some
a
number
of
those
apis
are
protected
and
are
meant
to
be
invoked
in
context
of
a
user
calling
them
because,
if
you
imagine
your
messaging
as
a
system
has
a
lot
of
access
to
kubernetes
and
to
workloads
and
service
meshes
and
and
so
ensuring
that
as
meshri
server
sits
there
as
a
long-lived
process
and
is
watching
and
helping
manage
the
infrastructure
that
people
can't
really
nearly
take
measuring,
ctl
and
start
to
tell
measure
server
to
delete
a
bunch
of
stuff.
A
And
so,
when
you
sign
in
to
measurey
as
a
user
in
the
upper
right
hand,
corner
you'll
find
the
ability
to
right.
Now.
It's
just.
It's
called
get
token.
So
you'll
find
the
ability
to
download
your
sessions
token,
and
then
you
can
use
that
token.
In
combination
with
measuring
ctl
to
authenticate
to
measures
api.
A
Everybody
everybody
used
cooper,
you
know,
touched
cube,
ctl
and
if
you
you've
noticed
cube
ctl
has
you
know
you
can
you
can
use
the
same
cube,
ctl
client
to
talk
to
different
kubernetes
deployments,
different
kubernetes
servers?
A
Well,
how
do
you
know
when
you're
using
cubectl
and
you're
going
to
get
pods
or
whatever
you're
going
to
do,
which
kubernetes
and
which
of
your
kubernetes
clusters?
Is
it
going
to
talk
to?
Well,
it's
going
to
talk
to
your
current
context.
It
has
so
what
do
you
mean?
Let's,
if
we
look
at
that,
if
you
look
at
your
in
your
local
file
system,
where
your
your
home
directory
is
you'll,
find
a
dot
cube,
a
hidden
folder
that
has
a
file
called
config
in
that
config?
A
A
Set
context,
kind
or
yeah
whatever
anyway
so
for
measuring
ctl,
there's
now
a
context
command
and
if
you
execute
just
this
will
say
well,
hey
your
current
context
is
called
local.
Maybe
you
wanted
a
new
one
called
well
beloved
bob
navendu
hates
bob,
because
bob
is
a
name.
That
is
the
test
case
that
that
we'll
use
and
bob
causes
him
a
lot
of
trouble.
So
so
you
can
go
anyway.
What
am
I
so?
So
we
I'm
totally
digressing,
because
I
need
to
sort
of
ask
like
hey.
A
What's
this
token
for
and
so
yeah
it's
to
authenticate,
and
then
what
I
was
trying
to
say
is
that
when
you
look
at
the
given
that
you
might
be
using
measuring
ctl
to
talk
to
different
instances
of
measuring
well
those
different
instances,
you're
going
to
have
a
session
over
here
and
a
session
over
there,
so
you're
going
to
have
two
different
tokens
to
talk
to
two
different
measuring
servers.
A
This
config
file-
I
I'll
refer
to
it
as
like
mesh
config
sort
of
just
like
cube
config,
there's
a
mesh
config.
So
it's
this
file
here
when
you
look
at
it,
here's
my
context.
I
just
created
bob
as
a
context,
it's
pointing
to
my
local
mastery
server.
It's
going
to
use
this
token.
If
I
want
to
invoke
certain
api
calls,
the
thing
is
is
like:
maybe
I'm
not
using,
maybe
the
token
that
I'm
using
is
valid
on
both
of
the
meschery
servers
or
maybe
not.
A
A
It
takes
it
takes
a
smart
engineer
to
crack
through
the
anisha.
Well,
fortunately,
no,
what
will
happen
is
if
you,
if
you
choose
the
none
provider
you
know
measuring,
gives
you
it
doesn't
give
you
a
multi-user
experience.
It
gives
you
this
implicit
single
user
experience,
even
though
you
don't
that
that
single
user
doesn't
really
have
a
name,
that
single
user,
that
unnamed
anonymous
single
user
does
have
a
session
and
and
therefore
does
also
have
its
own
token,
and
so
it
still
bears
the
same
requirement
to
auth.
C
G
C
C
Yeah,
I'm
so
sorry
yeah
right
so
does
this
user
have
like
a
constant
token
or
is
it
generated
nearly
each
time,
but.
A
A
A
A
Okay,
so
so
to
anish
to
your
point
like
hey,
there's,
there's
really
not
much
in
here
other
than
just
identifying
the
fact
that,
like
it's
using
none
and
so
now,
you're
gonna
jog,
my
memory
okay.
So
if
we
come
over
to
so,
we've
got
an
active
session
and
we
want
to
do.
We
want
to
invoke
something
that
takes
a
that
requires
a
token,
maybe
like
a
performance
test.
So
in
the
performance
test.
A
A
It
is
responding
or
like
the
get
requests
are
going
so
it
should.
It
didn't
really
matter
like
it
should.
A
Okay,
it
should.
A
Here's
error
failed
to
add
auth
details
to
the
request,
so
the
by
the
way,
the
token
was
working
a
little
bit
differently
in
the
past,
where,
if
you
had
like
you
know
previously
loaded
it
up,
then
I
forget
the
details
like
it
would
store
it
in
a
cache
file
so
that
it
would
make
it
convenient
for
the
user
not
to
have
to
always
add
the
token
on,
but
in
this
case,
so
so
for
right
now
it's
saying
hey:
this
is
a
protected
interface.
Sorry,
you
you've
got
to
give
me
a
token
okay.
A
Well,
I
just
downloaded
my
more
or
less
empty
token.
This
is
it
and
it's
saying
hey.
I
don't
know
what
that
shorthand
is
so
we've
got
some
discrepancies
in
our
cli,
like
because
in
the
other
area
in
the
I
don't
know
in
in
the
first
like
the
patterns
for
some
of
the
other
commands,
they'll
take
a
dash
t,
so
a
niche
actually
a
good
note
to
take.
A
We
should
go
fix
that,
so
it
looks
like
it
offed,
it's
probably
taking
30
seconds
to
go,
run
a
test
whether
or
not
the
test
is
successful
or
not.
Is
a
second
question,
so
the
question
is
right
now
is
like
well
hey,
then,
just
can't
it
to
a
to
a
hacker
who
becomes
familiar
with
meshri.
They
understand
that,
like
there's
this
empty
file
or
basically
that
they
don't
need
a
if
there
is
a
an
issue
here.
A
I
think
that
if
someone
is
running
mystery
in
a
production
environment
and
they
allow
people
to
off
anonymously
if
they
like,
if
they
allow
the
none
provider,
then
right
now,
what
we're
seeing
is
that
people
can
invoke
a
it
that
they
can
invoke
commands.
I
think
our
recommendation
would
be
if
we
didn't
change
any
other
code.
Our
recommendation
would
be
hey.
Don't
do
that.
Don't
allow
people
in
your
environment
to
run
to
use
the
none
provider.
A
A
I
have
to
go
look
at
some
of
the
other,
so
I
mean
that's
a
that's
something
to
think
about
and
maybe
do
something
about,
because
while
you
can
explain
it
away-
and
you
can
say
hey
that
you
don't
it's
against
best
practices
to
deploy
in
that
model,
you
still
want
to
take
care
of
it
and
really
not
let
people
shoot
themselves
in
the
foot.
A
I
have
to
go.
Look
at
the
in
mystery
itself
when
it
is
built
at
build
time.
It
has
a
salt.
It
has
a
secret
string
that
it
will
use
to
to
communicate
to
to
facilitate
secure
communications
that
secret
is
injected
at,
like.
I
was
saying
it's
injected
at
build
time
of
not
build
time
of
measuring
through
its
ci
system.
So
the
only
place
that
that
secret
is
stored
is
in
the
ci
system
injected
in
at
build
time.
A
But
I
don't
know
that
that
really
bears
weight
in
this
interaction
right
here
that
does
bear
weight
if
people
are
if
people
are
submitting
these
performance
tests,
these
results
anonymously
people
have
the
choice
of
sending
in
anonymous
performance
results
or
not,
so
they
can.
They
can
disable.
A
These
functions
that
salt,
that
secret
string
that's
used
to
facilitate
anonymous
or
secure
anonymous
communication
is
really
oriented
for
this
because
part
of
mesherie's
goal,
if
you
well,
I
don't
know
I
was
going
to
say
if
you
recall,
I
don't
know
that
everyone
recalls,
because
I'm
not
everyone
was
here,
but
but
part
of
the
genesis
of
the
project
was
to
help
answer
this
question
of
like
what's
the
overhead
of
running
a
mesh.
Well,
all
the
environments
are
different.
Let's
start
to
collect.
A
Let's
start,
let's
give
people
a
tool
to
run
it
and
test
it
out
and
let's
start
to
collect
those
so
that
we
can
analyze
them,
let's
partner
with
universities,
to
analyze
those
and
do
research
on
them.
A
So
if
I
have
so,
if
we
go
do
since
we're
talking
about
this,
let's
do
this
real,
quick.
Let's
go
do
another
test.
I
haven't
done
this
test
right
here,
I'm
assuming
that
this
is
the
case
that
I
disabled,
the
sending
of
anonymous
performance
test
results,
so
I
should
see
in
here.
A
A
If
I
go
and
look
at
the
ui
here,
I
refresh
we're
at
1953
1953.
Okay,
if
I
come
over
here
and
we
say
well,
you
know.
Actually
it
is
our
preference
to
persist
these
performance
results
and
we
go
run
a
test
for
five
seconds.
A
1954
good
so
anish,
you
know
like
there's
something
for
us
to
to
consider
like
it's
a
reasonable
response
to
give
to
people
to
say
to
to
empower
them
to
to
disable
the
none
provider
like
if
they,
if
they
only
want
to
have
certain
providers
to
turn
them
off
that
we
should.
You
know,
try
to
make
it
easy
to.
You
know,
facilitate
that
type
of
a
thing
and
as
people
can
lock
down
their
environments,
prevent
access.
A
A
And
so
yeah,
hopefully
this
helps
explain.
I
don't
know
some
of
those
idiosyncrasies
a
couple
things.
I
think
we
should
do
one.
We
should
create
an
issue
to
have
a
dock
that
says
what
does
a
production
deploy?
What
are
the
considerations
for
a
measuring
production
deployment.
A
A
Do
we
disable
the
ability
to
interact
with
this
thing
through
anything
other
than
the
ui?
It's
like?
Well,
because
even
the
ui,
it's
a
client,
it's
it's!
It's
just
like
mastery
ctl.
It
itself
needs
to
auth
to
the
server.
It's
able
to
do
it
a
little
bit
special
at
the
moment,
because
it's
built
into
the
same.
E
A
It
makes
sense
yeah
or
that
there's
a
a
default,
a
lot
alarm
or
alert
that
sits.
D
A
A
That
makes
a
lot
of
sense
makes
a
lot
of
sense
to
me
there.
It
michael
in
a
very
timely
way,
there's
an
issue
out
here,
calling
for
a
configuration,
wizard,
a
multi-step
form
that
will
help
people
where
I
think
that
you
know
more
or
less
what
you're
saying
can
be
integrated
into
it's
to
it's
yeah,
it's
to
help
facilitate
people
getting
their
message,
deployment
correct,
connecting
mystery
to
kubernetes.
A
So
all
right
good,
so
we've
got
you
know.
We've
got
one
minute
left,
which
means
we
get
just
enough
time
for
road
map.
So
pranav,
let's,
let's
plan
out
to
talk
about,
maybe
we
can
talk
about
image
hub
on
friday,
potentially
might
give
us
a
give
you
no
issues.
Yeah,
it's
your
fault!
It's
your
fault!
You
guys
yeah!
A
K
Topic:
okay,
I
was
intended
for
tomorrow
for
the
ci
group
and
I
wanted
to
send
the
mail
today,
but
that
I
went
out
and
yeah
cool
thing
is
that
it's
it
was
working
with
the
determinable.
K
I
was
able
to
to
use
like
digital
ocean
as
a
as
a
an
adapter,
the
the
way
you
have
mini,
cube
or
kind
to
have
it
with
with
so
that
was
cool.
I
don't
know
I
yeah
what
you
said
friday.
K
Okay,
I
have
a
question
which
kubernetes
distribution
did
you
say
you
said
in
previous
meetings
to
use
we're
gonna
migrate
from
mini
cube
or
from
I
don't
know,
but
I
don't
remember
which
one
you
said:
you're
we're
gonna
use.
A
Well,
let
me
rephrase
the
quote,
which
kubernetes.
K
A
No
problem-
hey,
hopefully
yeah
we
can,
you
know
I
don't
know.
Hopefully
this
is
just
a
high
level
stab
at
what
some
of
those
milestones
might
be
working
our
way
toward
1.0.
A
A
A
Is
that
before
some
of
these,
like
are
all
of
these
necessary?
I
don't
know
that
they-
I
don't
know
that
they
are.
I
think
I
think
that
they
are
my
my
perspective.
Right
now
is
that
they
are
but
something
to
think
on.
A
And
some
of
these,
I
expect
aren't
obvious
to
to
you,
know
what
what
they're
all
about.
So
you
know
we
we've
held
off
on
holding
separate
steering
committee
meetings.
Like
you
know,
road
map,
you
know
things
and
would
not
of
that
size
necessarily
just
yet,
but
we
could
dedicate
a
measuring
dev
meeting
or
to
these
topics
and
kind
of
highlight
them.
It
might
be
more
productive
to
just
expand
on
them
and
articulate
them
in
written
form,
send
it
around,
and
maybe
let
people
digest
and
then
set
up
a
time
for
a
discussion.
K
A
Yeah
yep
it
it
would
be
this.
Would
that
see
that
that,
because
the
answer
to
the
question
is
like
that,
it's
unrealistic
to
expect
that
we
would
go
faster
than
a
quarter
per
minor
release
like
that
or
or
considering
what
is
stuffed
into
each
of
those
milestones.
A
A
A
I
don't
know
we
wrote
this
almost
two
years
ago
now,
yeah
but
yeah
about
about
once.
I
know
it
says
minor
release
here.
Minor
releases
like
that
would
be
great.
We
have
too
much
stuffed
into
those
the
milestones
currently
to
make
that
happen.
We
don't
necessarily
have
to
hit
a
1.0
in
order
for
people
to
start
using
it
in
production.
A
J
A
Nice,
okay,
anything
else.
A
See
you
guys
all
tomorrow
and
friday.