►
From YouTube: Meshery Development Meeting (April 14th, 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
Three
after
so,
let's
get
started
and
let's
hope
others
will
join
if
you
haven't
already
put
your
name
in
the
meeting
minutes
and
yeah,
let's
get
started
so
welcome
everybody.
This
is
the
mastery
development
meeting
today
is
april
14th
and
without
wasting
any
time.
Let's
get
started
with
the
first
item
on
our
agenda
so
aditi,
I
guess
you're
up.
B
Hey
everyone
good
morning
or
good
evening,
so
I'll
just
talk
about
my
current
top
pr,
which
is
deleting
resources
on
system
stop
when
measuring
ctl
measure
is
running
on
platform
kubernetes,
so
I'll
I'll,
just
walk
I'll.
Just
tell
about
what
I
learned
through
this
pr.
So
when
I
first
started
with
this
pr
by
the
way,
this
was
my
most
daunting
pr
till
date.
I
wasn't
sure
how
exactly
to
proceed,
and
hence
there
was
quite
a
bit
of
delay
in
me,
starting
once
I
understood
how
the
internals
worked.
B
I
I
mistakenly
or
used
client,
go
and
deleted
all
the
custom
resources
individually
after
looking
at
the
measuring
operator,
but
thanks
to
abhishek,
he
told
me
that
he
he
told
me
that
I
I
could
just
use
the
existing
meshkit
cube
and
meshkit
utils
functions
and
that
would
delete
that
would
delete
the
resources
from
me.
If
I
just
provided
the
url
now
now
now
from
this
issue,
there
is
another
possible
opportunity.
B
One
is
one
is,
instead
of
just
hard
coding,
the
url,
the
way,
the
way
measuring
name
space
and
other
and
other
measury-wide
details
have
been
specified
in
utils
helpers.go.
B
We
could
specify
all
the
operator
file
urls
and,
as
as
you
can
see,
from
a
review
measuring
name,
though
the
name
measurey
can
also
be.
We
can
also
remove
that
hardcore.
The
way
namespace
has
been
removed.
B
So
to
sum
it
up,
I
think
I
I
think
in
this
pr
I
just
explored
meshka
tutus
more
carefully
and
next
time.
I
I
think,
instead
of
directly
using
the
client
libraries
I'll
explore
what
the
meshkit
resources
better
thanks.
A
Thanks
for
that
aditi,
I
guess
oh,
is
this
pr
ready
to
review.
C
C
In
part,
I
think
the
question
about
review
in
that
you
did
a
fantastic,
actually
yeah.
I
see
it
now
you
updated,
so
you
did
a
fantastic
job
of
identifying
work
in
progress
in
the
subject
of
the
pr,
and
then
you
also
use
the
draft
flag
to
signal
that
work
was
in
flight
so
that
that's
perfect,
that
that
didn't
stop
anyone
from
giving
you
feedback,
but
so
that
that
worked
out
that
works
out
very
well.
Thank.
B
C
Yeah
we'll
get
you
a
kubernetes
sticker
in
the
mail,
maybe
for
that
that
laptop
that
doesn't
have
kubernetes.
C
Well,
you
know
centralize
some
of
the
configuration
or
externalize
some
of
the
configuration
with
respect
to
the
namespace
in
which
measurey
is
deployed
in
kubernetes
by
default.
That's
that
namespace
is
called
mesheri
and
you
know
it
would
it'll
be
very
few
and
far
in
between
the
case
that
an
organization
or
user,
that's
deploying
mesherie,
would
excuse
me
want
to
change.
C
Maybe
I
shouldn't
say
very
few
and
far
in
between,
but
like
it's
not
the
typical
use
case,
I
would,
I
would
say
there
might
be
organizations
that,
like
to
name
space
apps
like
mesherie
into
a
management
namespace
or
something
like
that
or
maybe
they'd
like
to
they
have
a
naming
convention
where
they
prepend
someone's
name
or
some
project
name
with.
C
Then
the
name
of
the
app
in
this
case
measury
and
so
understanding
that
there,
while
it
might
not
be
the
80
like
there,
are
some
individuals
who
might
want
to
customize
that
and
I'm
probably
conflating
two
things.
So
I
guess
maybe
first
question
is
adida.
Do
you
know
if
so
like?
If,
if
through,
if
the
yep
today,
if
someone
had
deployed
measuring
into
a
different
name
space,
would
we
be
in
trouble
in
terms
of
stopping
mystery
with
this
command.
B
I
I
I
don't.
I
don't
think
so,
because
if
if,
if
at
all,
I
refer
to
it,
I
always
refer
to
it
with.
B
C
Yeah
upper
url
between
the
operator
url
and
the,
if
I
think,
is
it
a
dd
or
novendo
if
you
go
up
back
up
to
where
the
mesh
rename
space
is
right,
there
yeah
line
148.,
so
a
dd
where
we're-
and
I
don't
know
if
this
is
this-
might
be
an
older
copy
of
the
code.
I
don't
know,
but
on
line
148.
C
D
Yeah,
I
am,
I,
I
don't
see
his
instructions.
I.
D
H
C
Okay,
cool
yeah
yeah,
so
I
guess
to
help
get
us
onto
the
same
page
on
the
the
on
the
the
the
reason
that
there's
the
terminal
being
shared
is
because
there
isn't
a
terminal
being
shared
or
like
the
folks
aren't
talking
about
that.
But
there
is
a
web
browser
being
shared.
That's
open
to
a
pull
request.
The
pull
request
is
talking
about
how
to
how
to
make
sure
that
the
command
message.
A
C
System
stop
works
for
yeah.
C
C
E
C
I
C
E
H
I
hear
the.
H
E
D
C
A
C
Know
what
it
is,
it
might
be
double
check
on
your
other
browser
tabs.
If
you
would
stephen
it
might
be
that
there's
a
youtube
video,
that's
got
the
sweet.
F
C
C
Fair
enough
totally
understandable
about
the
con
confusion,
then
so
so
cool
so
so
adida
the
I
I
completed
two
things
I
think
like
the
name
of
the
app
called
meshary
and
then
the
name
of
the
namespace
and
and
like
you
were
saying
you
know,
you're,
not
you
in
so
much
as
you
can
tell
at
the
moment.
We
may
not
have
a
name,
an
issue
with
how
the
namespace
is
being
referenced,
because
what
you
have
configured.
C
What
you've
coded
up
is
something
of
a
centralized
reference
to
the
measuring
namespace
from
the
utils.
B
Actually,
in
this
pr,
this
getting
service
endpoint
and
these
two
three
lines
are
extra.
So
I'll
just
push
a
comment
now,
removing
these
I
just
realized,
I'm
not
using
them
yeah.
Even
this
thing,
even
endpoint
is
extra.
It's
it
just
starts
from
operator
url
I'll,
just
push
that.
J
C
And
yeah
minimally,
it
would
be
minimally.
It
would
be
a
step
forward
to
centralize
operator
url
into
like.
C
B
C
Abhishek
had
been
openly
pondering
use
cases
around
the
life
cycle,
management
of
measuring
operator
and
I
think
part
of
those
use
cases.
It
looks
like
you're
addressing
here
that
might
have
been
at
his
direction.
I
don't
know
yeah
but
you're
taking
it
looks
like
this
is
accounting
for
the
dismantling
of
mastery
operator.
So
it's
taking
down
the
operator
as
well.
C
C
C
Me
a
couple
of
things
that
this
this
thing
mesh
kit
was
mentioned
a
few
times
and
it's
a
library
that
adidi
is
using
and
so
there's
a
link
there
to
a
blog
post
that
one
of
the
maintainers
has
written
about
mesh
kit
and
what
its
purpose
is.
So
that's
helpful.
A
Oh
yeah,
so
I
guess
we
should
move
on
to
the
next
item.
We
have
a
couple
of
newcomers
on
call
today,
so
we
have
a
tradition
where
we
kind
of
embarrass
the
newcomers,
so
I
will
leave
up
to
khan
and
amara
to
introduce
themselves.
I
Hi
yeah
yeah,
I
guess
I
can
go.
My
name
is
kai.
A
computer
science
student
undergrad
nyu.
I
recently
heard
measuring
through
one
of
my
professor,
because
currently
working
with
him
on
distributed
tracing
so
just
take
a
peek
at
the
repository
and
yeah
just
find
some
new
issue
and
just
kind
of
try.
My.
C
C
C
I
We
were
talking,
we
talking
about
layer,
five
and
how
we
can
integrate
things.
That's.
C
Cool
yeah,
by
the
way,
to
help
folks
get
to
know
khan
a
little
bit
the
some
of
the
areas
of
focus
for
well
for
I'm
not,
I
don't
want
to
speak
in
collins
explicitly,
but
just
his
professor
and
kind
of
general
area
focus
at
least
one
of
the
things
is
a
really
intriguing
set
of
questions
being
asked
about
distributed
tracing
and
the
sampling,
algorithms
or
the
approaches
to
to
sampling
the
efficiency
of
sampling
of
traces
tracing
as
a
telemetry
signals
is
very
insightful
answers.
C
Probably
answers
more
questions
than
metrics
answer
and
depends
on
what
questions
you're
asking
maybe
tracing
answers
more
questions
than
logs
do,
but
you've
got
your
triumvent
of
those
three
signals
and
tracing's
expensive
and
that's
based
on
how
you
know
largely
like
how
many
spans
you
have
in
a
trace
and
how
many
traces
you're
generating
and
how
you're
sampling
those
and
storing
those
and
it
gets
costly
and
so
there's
a
lot
of
intriguing
thought
to
be
put
into
the
approach
by
which
you
sample
and
doing
it
intelligently.
C
And
so
so
I
can't
wait
to
nerd
out
on
some
of
that.
Some
of
those
topics,
because
that's
that's
pretty
interesting
and
I
do
think
that
there
are
there-
is
a
lot
of
I'm
pumped
that
you're
here
calling.
I
didn't
know
that
that
was
where
you're
coming
from.
But
that's
I
yeah,
there's
there's
there's
a
few
areas
to
point
out
to
you
that
may
be
of
interest,
so
yeah
yeah.
C
Now
I
understand
why
you
move
through
those
pr's
so
quickly.
Does
that
make
sense.
C
Cool
great
now
that
now
that
we've
embarrassed
on
to
embarrass
amara
welcome,
please
please
take
a
moment
to
introduce
if
you
would.
I
Cool
yeah
yeah.
I
think
the
lettuce
pr
address
the
issue
of
listing
all
contacts,
all
given
context
in
a
config
file
so
which
is
yeah.
I
I
think
it
was
pretty
fairly
straightforward
because
all
the
other
files
kind
of
constructed
similarly-
and
it's
like
the
command
line,
so
I'm
pretty
new
to
go
so
all
the
syntax
will
go
is
interesting
to
me
but
yeah.
I
should
fix
the
based
on
some
suggestion.
I
should
fix
it
and
then
it
should
be
good.
I
think.
C
Nice
navindi,
did
you
happen
to
or
khan?
Is
there
a?
We
don't
have
the
output
of
what
that
looks
like
by.
C
And
I
think
yeah
okay,
this
is
interesting
like
yeah,
the
this
actually
calls
to
one
of
the
things
that
I
was
wondering
if
khan,
if
you
might
have
included-
and
it's
that
that
first
b
p
piece
of
information
about
the
user's
current
context,
do
we
denote
that
in
any
way
do
we
do
we
tell
them
that,
like
this
or
do
we
denote
that
in
any
way
with
because
in
concept,
while
this
isn't
true
all
of
the
time
in
concept,
most
of
the
time
the
user
will
be
within
a
current?
C
What
will
it
will
be
in
a
context
and
maybe
moving
between
them,
creating
them
or
deleting
them
or
switching
between
them?
So
it's
possible
that
that
they're
out
of
content
like
they
don't
they
don't
have
an
active
context,
the
current
context,
but
so
khan.
I
guess
the
question
is:
when
there's
a
list
given
do
we?
Do
you
tell
the
user
what
the
current
context
is.
I
Yeah,
my
well
I'm
just
extracting
it
based
on
the
yaml
file
that
yeah.
So
if
that
is
set,
then
it
will
tell
them
the
current
contacts
I
think,
based
on
the
latest.
I
push
the
latest
comment,
but
the
last
comment
should
list
all
available
contacts
and
tell
them
to
switch
if
they
don't
have
the
current
contact
set.
I
I
C
A
That's
good
work
on
so
yeah.
I
guess
you
made
the
changes
that
we
mentioned
yeah.
I
I
I
think
it
looks
good
yeah.
So
I
guess
we
will
move
on
to
the
next
item.
Husseina.
Are
you
on.
A
A
A
So
this
is
about
the
mystery
ctl
perf
command.
A
So,
on
top
of
what
we
have
now,
we
will
have
a
run
sub
command,
which
will
which
will
run
the
performance
test
and-
and
we
will
also
track
the
user's
performance
profile
on
the
mach3
config.
That
is
the
config.yml
file.
A
So
when
the
user
runs
a
performance
test
by
default,
it
will
check
the
current
context
and-
and
it
will
check
the
performance
profile
set
in
the
current
context
and
the
user
can
have
different
values
here,
configure
his
test
and
we
will.
A
Another
thing
is
that
we
will
add,
we
will
have
flags
where
the
user
can
override
all
these
set
values.
So
if
they
are
on
measuring
ctl
perf
run
and
add
a
url
flag,
then
they
can
override
their
their
current,
their
url
of
their
current
profile
and
similar
for
all
other
values.
A
C
Yeah,
while
other
people
are
pondering
this,
I'm
gonna
bring
up
a
terminal
just
to
help
us
explore
a
little
bit.
C
So
so
we
yeah,
so
we
should
be
seeing
a
terminal
now,
if,
if
so,
I
guess
to
help
give
divinity
some
feedback
he's
looking
at
adding
a
verb
called
run
to
the
mastery.
Ctl
perf
command
a
couple
of
things
to
note
one:
here's,
here's
what
we
have
for
mastery
ctl
perf
today,
perf
like
when
you
start
to
look
at
the
structure
of
the
commands
that
are
laid
out
from
s3ctl.
C
It's
we're
developing
a
nice
pattern
and,
as
a
matter
of
fact,
we
try
to
categorize
the.
C
Not
only
not
only
are
we
trying
to
follow
that
sort
of
hierarchy,
there's
also
something
of
a
categorization
that's
going
on.
So
if
we
just
look
at
mastery
ctl
top
level,
the
categorization
right
now,
the
top
level
items
are
but
yeah.
I
guess
a
few
nouns
system
pattern.
C
Yeah,
I
guess
yeah
now
right,
yep
anyway,
so
today
the
behavior
is
such
that
if
the
user
goes
to
wants
to
run
the
performance
test,
there's
no
verb
necessary.
They
just
do
measure
ctl
perf
and
they
supply
either
a
string
of
flags
indicating
what
type
of
performance
test
they
want
to
run
for
how
long
against
what
endpoint
or
more
simply
they
just
provide
a
service
mesh
performance
compatible
file.
So
a
file,
that's
of
the
smp
format
and
that's
a
little
bit.
You
know
easier.
C
C
Well,
one
bringing
forth
new
constructs
like
that
necessitates
changes
in
the
command
structure
to
to
help
provide
a
consistent
user
experience.
C
There's
any
time
that
you
have
a
test
result,
it
would
have
that
test
would
have
been
run
in
context
of
a
performance
profile,
and
so
that's
the
that's
where
novendo
is
coming
from
run
is
niven
sounds
like
that's.
That's
the
current
thinking
about
the
particular
verb
to
use
to
to
run
a
performance
test.
Okay,
that
you
know
for
us,
humans,
that
seems
to
you
know
be
sensical
when
we
reflect
when
we
talk
about
consistency
and
consistency,
driving
quality
of
user
experience,
so
we've
got
one
or
two
other
productive,
producing
or
productive
commands.
C
So
one
of
those
commands
being
used
elsewhere
is
apply.
Another
command
being
used
elsewhere.
Is
create
these,
are
you
know,
additive
or
positive,
or
have
positive
results.
C
C
Yeah
create
it
create,
maybe
isn't
the
nicest
of
apply
is
also
yes,
like
it
kind
of
crosses
the
boundary.
If
you
use
cube
ctl
a
lot
apply,
sort
of
crosses
the
boundary
between
potentially
creating
a
static
definition
of
something
like
a
profile
or
taking
that
definition
and
having
it
realize.
So,
if
you
do
cube,
ctl
apply
dash
f,
you
can
take
a
static
definition
and
have
cube
ctl
or
kubernetes
instantiate.
That
object.
C
C
C
She
said
what's
going
on
what
what
are
these
anyway
and
they
go
to
list
that
they
they
go
to
just
find
a
common
user
experience,
so
they
it
becomes
intuitive
after
a
time
and
so
that
intuitiveness
reinforces
consistency
of,
and
so
my
part
of
my
point
here
is
and
just
an
open
thinking
about
the
use
of
the
word
run
and
while
it
like
well,
I
think,
when
you
consider
it
just
myopically
unto
mastery
ctl
perf
alone
as
an
individual
command
run,
sounds
pretty
great,
like
that's
very
descriptive
of
what's
gonna
happen.
C
C
Anybody
what's
that,
so
what
would
I
propose
as
an
alternative?
That's
a
great
question.
I
don't
know:
let's
explore
that,
so
if
we
did
perf
apply
dash
f
that
might
get
us
there
does.
Does
that
feel
entirely
awkward
to
anyone?
If
you
invoked
this
command
and
a
performance
test
was.
C
Run
you
might
not
be
applying
a
file
in
this
case,
you
might
be
doing
something
much
more
convenient
around
the
performance
profile,
so
navendu
right
now.
Do
you
recall
what
you
have
proposed
for
the
interfacing
with
the
profiles?
A
Oh,
no,
we
actually
use
only
the
listened
view
command.
A
Okay
and
yeah,
and
to
like
the
list
and
view
commands,
can
be
used
for
like
viewing
the
whole
whole
whole
performance
process,
all
the
performance
profiles
or
the
or
like
the
individual
results
or
yeah,
but
it
may
need
some
reconsideration,
as
as
we
are
moving
into
like
we
are
adding
the
performance
profile
consideration
in
the
config.yml
file,
so
some
changes
might
be
needed.
A
Like
for
right
right
now,
we
don't
need
to
like
we
specify
the
since
the
performance
profiles
are
specified
in
the
config
file.
We
would
just
need
to
call
the
run
command
or
run
command
and
we
can
change
the
the
profile
with
the
flag.
So
that
is
what
we
have
currently.
C
C
Nice,
well
maybe
we'll
do
a
so
so
this
brainstorming
has
been
helpful.
Is
what
I'm
hearing
good
or
it's
been
or
it's
been
disruptive
I
don't
know
it
depends.
F
F
C
Yeah,
I
don't
know
if,
yes,
I
think,
sometimes
the
those
two
words
end
up
being
interchangeable
between
like
argument
like
an
argument
with
a
a
global
flag.
Rather,
I
guess
is
that
the
parlance
it,
although,
having
said
that
there
is,
there
are
exceptions
to
the
rule
at
the
moment.
So
if,
if
we
look
at
a
different
area
of
mastery
ctl
system,
this
is
a
fairly
well
like.
This
is
a
fairly
decently
done.
Category
of
of
measuring
ctl
capabilities.
C
Commands
so
system
channel
config,
and
so
if
we
look
at
one
of
these
like
config
or
rather
context,
I
don't
know
them
so.
There's
there's
measure
ctl,
there's
the
noun
kind
of
the
area,
the
category
in
this
case
this
ends
up
really
being
the
category
of
things
that
we
want
to
get
into.
We
want
to
deal
with
measuring
as
a
system
and
talk
to
and
manipulate,
measuring
and
its
deployment,
and
that's
what's
being
in
some
cases
that
works
very
well
like
there's
the
noun
there's
the
verb,
we're
good
to
go
in
other
cases.
C
It's
like
well
yeah.
We
could
start
the
entire
system,
but
in
other
cases
like
well,
I
just
want
to
tweak
an
aspect
of
what
we've
got
going
on
here,
like
I'd
like
to
switch
from
the
stable
channel
over
to
the
edge
channel,
and
so
this
is
a
valid
command
and
you
end
up
with
a
couple
of
nouns,
then
a
verb
and
an
argument.
C
So
it
isn't.
You
know
strict
adherence
and.
C
Not
out
of
a
desire
not
to
be
consistent,
but
just
so
dhruv
and
vijay,
and
in
part,
why
I'm
going
to
give
you
this
next
suggestion
and
I'll
say
this
is
in
part,
why
I
was
mostly
trying
to
brainstorm
is
because
a
lot
like
when
feedback
is
given
on
a
user
interface
in
the
in
a
web-based
user
interface
mesh
for
your
ui.
It's
easy
enough
to
look
at
one
area
and
say
that
font
looks
too
small
like
or
wouldn't
wouldn't
black
look
nicer
or
whatever
it
is.
C
But
when
you
step
back-
and
you
look
at
not
just
that
page
but
this
screen
and
this
screen
it's
like
oh
well,
that
would
be
totally
inconsistent
that
wouldn't
we
shouldn't
do
that
and
and
so
they're
in
live
the
challenge,
for
you
know
great
design
on,
because
this
is
the
exact
same
thing:
it's
a
user
experience
and
it
needs
to
be
consistent
and-
and
so
part
of
my
hesitancy
anymore
is
giving
a
strong
opinion
about.
C
C
This
is
a
verb
that
actively
switches
the
user
real
time
from
that
to
the
restart
measure
and
move
you
over
like
it
does
a
specific
thing
and
it's
it
was
necessary,
like
if
you
were,
to
apply
edge
or
create
edge
like
those
things
didn't,
so
we
needed
to
bring
forth
this
new
verb
and
while
the
verb
isn't
used,
ubiquitously,
it's
still
appropriate
and
that's
the
kind
of
scrutiny.
I
think
that
we,
you
know,
I
encourage
all
of
us
to
give
to
something
like
run.
It's
like
hey.
C
It
may
be
the
moment
in
time
where
we
do
need
to
create
a
new
verb,
because
it
is
a
different.
You
know
it
is
distinct
from
how
these
other
verbs
are
being
used
today
and
it's
not
like
yeah,
it's
not
like.
We
need
to
necessarily
rule
this
with
an
iron
fist
or
like
that.
We
should
we
should
have
as
few
verbs
as
possible
and
it
should
all
be
like.
C
I
don't
know
that
that's
going
to
be
really
helpful
either
like
because
if
you
have
a
fluctuation
of
behavior
when
you
use
the
same
verb
but
with
different
commands-
and
you
get
a
little
bit
of
fluctuation
in
behavior
like
well,
that's
bad
because
that's
inconsistent
as
well,
and
so
so
my
encouragement
is
to
go
read
through
navendi's
spec,
because
in
order
to
get
it
right,
it
needs
to
be
a
holistic
consideration.
C
I
think
across
that
and
it's
gonna
by
the
way
your
your
noodle
will
get
tickled,
because
there
are
multiple
concepts
being
included
in
perf
profiles
and
results
and
the
the
your
the
ability
to
create
some
static
profiles,
a
definition
of
something
to
look
at
something.
Historically,
that
has
transpired
look
at
results
from
different
runs
and
then
also
to
invoke
those
runs,
and
so
it's
a
totally
surmountable
task.
It's
just
why
we
have
to
write
it
down
into
a
spec
to
reflect
on
it
so
vijay
or
dhruv.
J
A
Oh
sorry,
I
was
on
mute.
I
guess
we
will
move
to
abhishek,
but
I
don't
find
him
on
the
car.
L
So
a
while
I'm
trying
to
share
the
screen.
Basically,
what
I'm
trying
to
show
in
here
and
that
basically
I'll
be
showing
is
the
basically
I'm
working
on
the
ability
of
measuring
so
that
it
can
start
importing
mystery
patterns
from
github.
That's
the
that's
something
that
I'm
targeting
right
now.
So
what's
exactly
important
from
bihar
means.
Basically,
you
can't
create
mystery
patterns
for
yourself
that.
L
L
You
may
find
something
like
this
if
you
log
in
with
with
the
right
provider,
so
you
can't
see
some
patterns
in
here.
These
patterns
have
been
imported
from
the
file
system.
This
is
functional
feature
right
now.
You
can
use
it,
but
what's
been
worked
on
right
now
is
the
ability
to
add
basically
import
these
patterns
from.
L
As
well
as
the
provider
which
basically
enhances
nursery
with
the
ability
to
import
gardens
from
from
github,
this
is
the
over
here.
If
someone
is
interested,
they
can
take
a
look
in
here.
L
This
is
the
endpoint
that
so
what
you
can
do
right
now
is
if
you
would
go
and
sign
in
with,
if
you
would
go
and
sign
in
with
it
will
ask
for
permission,
because
we
have,
we
have
to
basically
increase
the
scope
so
that
we
can
actually
gain
access
to
repositories
so
right
now,
basically,
if
you
would
go
and
try
to
log
in
through
github,
once
upgrades
will
merge,
you
will
get
this
screen.
L
You
can
authorize
industry
so
that
it
can
gain
access
to
your
repositories
and
can
start
importing
buttons
from
there
now,
if
you
would
so
because
the
ui
hasn't
been
worked
on,
so
what
I'd
be
sharing
is
an
endpoint.
L
L
Yep,
it
has
only
three
pattern
files
which
I
just
dropped
in
here,
so
that
I
can
test
a
few
things.
So
what
I'm
trying
to
do
in
here
is
I'm
specifying
that
hey
the
provider
I
want
to
import
from
is
github.com.
This
is
my
this
is
the
owner.
You
can
specify
the
organization
also,
and
then
this
is
the
repository
name,
and
this
is
the
part
this
part
can
be
like
could
be
of
any
dip.
L
That
is,
you
can
specify
that
and
go
to
this
representative
in
this
directory,
and
what
mission
will
do
is
basically
import
the
pattern
from
there.
It
will
also
ensure
that
the
pattern
is
a
pattern.
I
mean
it
complies
with
with
the
structure
that
we
have
defined
for
patterns
and
it
will
import
it
for
you,
so
this
is
a
basic
response
that
we
got.
This
is
the
pattern
that
I
tried
to
improve.
L
It
was
not
much
of
a
visual
tree.
That
is
because
it
was
just
an
api
endpoint
still,
if
someone
has
this,
this
is
actually
still
an
open
pl.
If
someone
okay.
L
This
is
an
open
er,
so
someone
can
take
and
look
in
here
because
this
is
this
is
a
very,
very
early
feature.
So
it's
still
in
development
still
in
consideration.
That
is
so
a
few
thing
that
I'm
considering
right
now
is
ability
like
introducing
the
generic
http
endpoints.
This
is
not
only
from
github,
you
can't
you
should
be
able
to
import
from
somewhere
else
also
and
the
ability
to
basically
go
and
go
and
look
at
previous
versions
of
the
file
so
yeah.
L
L
L
F
F
C
J
C
C
Make
bad
jokes,
okay,
good
the
crush
the
mechanism
or
the
criteria
by
which
the
the
in
the
to
be
imported
file
is
verified
as
being
compatible.
L
Yeah,
so
basically
still
so
what
happens?
Is
we
have?
So
it's
it's
support.
Basically,
that
is
responsible
for
verifying
this,
so
we
have
already
defined
these
trucks,
which
are
which
will
represent
the
pattern
file.
L
So
when
I
import
when
I
basically
send
update
requests
to
this
particular
part,
it
went
to
the
provider
and
it
fetched
the
file
from
from
a
github,
and
then
it
kind
of
it
kind
of
verified
the
this
these
bytes,
because
this
data
was
invited
and
then
it
verified
that
against
a
struct
and
because,
because
it
didn't
create,
I
mean
had
it
been
failed,
it
would
have
contained
that
this
thing
that
you
tried
encoding
is
not
a
pardon
file,
but
because
it
was
a
valid
pattern,
but
it
just
would.
C
Is
there
so?
This
is
the
contents
of
the
pattern
of
the
pattern
file
makes
tons
of
sense
with
respect
to
the
import
process.
What
additional
metadata
is
being
tracked
about
the
the
file,
the
pattern.
C
In
specific
to
I
get
well
geez,
I
don't
know,
I
mean
maybe
it's
kind
of
both,
but
it's
the
here's.
A
specific
example
in
question
is,
is
the
location
from
which
someone
has
imported
the
file.
Is
that
retained
as
metadata
about
the
pattern.
C
To
ponder
that
out
loud
with
you
for
a
moment
if,
if
if
there
was
a
location,
value
or
piece
of
metadata,
that
was
that
that
was
tracked
for
the
to
identify
the
genesis
of
any
of
the
patterns
that
mushri
has
actively
loaded
at
any
given
time.
C
C
So
if
there
was
a
location
field
that
was
an
attribute
of
any
pattern,
sources
of
input
or
sources
of
where
patterns
come
from
are
at
least
like
for
the
at
the
moment
seems
like
our
from
local
file
system.
So
if
there
was
a
location
attribute,
that
would
be
like
the
local,
the
path
to
that
on
the
file
system
at
the
time
if
it
was
imported
from
github
or
another
generic
http
endpoint
like
that
url
would
be
the
look.
You
know
the
value
of
the
location,
I
think
there's
a
just.
C
C
You
have
to
correct
me:
udkarsh
people
who
are
using
meshri
now,
should
everybody
be
seeing
patterns
in
their
ui
or
is
this
a
feature
flagged
thing?
Is
this
a
an
experimental
thing
that's
yet
to
be
exposed.
C
Okay,
so
okay,
good
anyway,
so
some
of
you
have
seen
this,
but
in
the
mesher
ui
you
might
see
a
menu
item
called
patterns
and
it's
what
crush
has
demoed
a
few
times.
You'll
see
a
table
listing
of
those
patterns
and
it's
just
yaml
files
that
you
can
upload
and
edit,
and
so
you
can
so.
My
point
is
the
other
location
from
which
a
pattern
can
be
loaded
is
visually
here,
so
that
a
user
clicks.
C
C
So
that
would
be
so
in
terms
of
like
the
location
that
location
attribute
one
might
that
just
might
be
another
http
endpoint
to
meshmap,
or
that
might
be
a
value
called
mesh
map.
I
don't
know
anyway,
okay,
so
whatever
those
values
are
why?
Why
do
we?
What's
the
point
of
tracking
that
attribute?
Why
do
we
need
to
have
the
location
of
the
pattern?
C
C
File
from
github
is
because
well,
if
you're,
you
know,
if
you're
serious
about
managing
your
infrastructure
and
particularly
doing
it
as
code,
then,
if
you're
going
to
invoke,
you
know,
apply
a
pattern
and
over
time
you
tweak
it
well,
you
might
want
to
keep
a
revision
history
of
that
or
you
might.
You
might
keep
your
manifest
files
and
your
pattern
files
in
a
source
control
system,
and
so
it
makes
sense
that
you,
you
know
that
that
makes
intuitive
sense
as
we
go
to
do
a
one-time
import.
I
would
posit
that.
C
That's
not
a
one-time
thing
like
the
initial
import
is
sort
of
you
might
use
the
term
one
time,
but
but
it's
really
an
ongoing
communication.
That
would
we
would
ultimately,
you
probably
want
to
be
aware
of
or
facilitate,
or
at
least
let
people
whether
or
not
mesherie
does
anything
about
it.
Just
at
least
give
them
that
additional
metadata
because
it
becomes
useful
to
distinguish
because
right
now,
it's
you
know.
Unless
I
had
foresight
to
giving
the
the
best
file
names
to
my
patterns,
it
might
not
be
enough
distinguishing
information.
C
K
I
just
wanted
to
say
something:
it
looks
like
when
you
say
that
you're
in
some
way
referring
to
requirements
and
so
at
a
particular
time
when
you
have
a
particular
requirement,
you
you're
happy
you're,
you're,
generating
or
coming
up
with
this
new
pattern
file,
and
so
I
was
thinking.
Is
it
possible
to
have
like,
like
a
notes,
field
where
you,
where
you
can
specify
the
reason
why
you
have
this
new,
this
new
version,
which
may
be
called
that
or
maybe.
K
L
K
I
I
I
mean
what
was
the
reason
for
having
a
new
word
for
having
a
source
control
because
you're
having
you're
bringing
up
the
source
control,
because
there
is
a
change,
there
is
a
new
version
and
what
is
the
reason
for
the
new
version
is?
Is
it
like
you're
saying
you
know
that
there
is
a
change
to
the
infrastructure
right
and
so
there's
a
change
to
the
infrastructure,
because
there
has
been
a
change
to
the
requirements.
C
C
It's
funny
actually
because
the
those
that
coined
the
term
git
ops-
I
don't
see
this
from
them
and
and
I
described
it
to
them-
they
got
excited
about
it
three
years
ago
and
they
still
haven't
done
it
so,
but
we're
going
to
do
it
and
it's
in
part
based
on
the
fact
that
we're
doing
this
and
what
we're
going
to
do
is
provide
context
to
what
you
just
said
and
there's
a
couple
of
ways
of
interpreting.
I
think
what
you
just
said,
but
but
one
of
the
ways
in
which
I
do
it
is
that.
C
Your
the
static
nature
of
your
patterns
that
this
is
a
is
a
descriptor
of
how
you
want
your
infrastructure
running
your
infrastructure
may
or
may
not
be
running
like
this.
You
may,
and
hence
you
know
in
large
respect
like
the
core
concept
to
part
of
what
kubernetes
does,
which
is
just
a
big
fat
reconciliation
loop
of,
like
you,
know,
here's
the
here's,
the
description
please
make
that
reality.
Please
keep
reality
and
sync
with
the
description
just
sit
here
and
and
make
it
so
well.
There
are
various
forces
that
affect
changes
to
your
infrastructure.
C
Some
of
those
are
bugs
and
apps.
Some
of
those
are
hackers
hitting
on
your
apps.
Some
of
that
is
rolling
out
a
new
version
of
your
app
those
things
come
from
different
vectors,
and
some
of
that
is,
and
so
that
that
one
of
like
rolling
out
a
new
version
is
like
you
would
think
that
would
be
an
event
that
everyone
everyone
on
your
team
would
be
very
well
aware
of.
Like
hey
yeah
yesterday
we
released
this
next
version.
C
I
was
like
well
yeah,
maybe
they'll
be
aware
of
it
when
you
have
like
you
know:
50
micro
services,
100
micro
services
running
around,
and
you
know
different
teams
of
different.
You
know
intentionally
and
that's.
The
whole
purpose
is
that
they're
releasing
as
and
when
ever
they're
releasing
and
if
they're
a
high
performing
team
like
they're,
releasing
like
six
times
a
day.
I
mean
they're
like
releasing
all
the
time
and
they
should
be
and
and
it
actually
it
makes
me
nervous.
I
get
really
nervous
when
we're
like
six
weeks
without
a
measuring
release.
C
I
want
for
us
to
be
releasing,
even
though
this
isn't
a
continuously
delivered
thing.
It
just
reduces
the
risk
as
you
release
every
time,
and
my
point
is
tracking
that
event.
If
you're
going
to
do
git
ops,
if
you're
tracking
that
event
and
understanding
that
your
versioning
system
is
supposed
to
be
representative
of
your
the
reality
of
your,
what
you're
actually
running,
there's
a
lot
of
opportunity
in
there
to
provide
users
with
insights
about
not
just
current
cpu
and
memory
consumption.
C
V2
v3
here.
Was
these
events
showing
them
and
who
to
blame?
Because
who
made
that
code
change?
Clearly,
that's
usually
the
event
music
and
then
like
they're
there
and
lie
I'm
a
significant
opportunity.
Hence
in
part,
why
you're
going
to
want
to
have
ongoing
interaction
with
you
want
to
get
really
friendly
with
your
git
based
system
or
your
source
code
system.
J
J
So
when
the
one
that,
when
the
when
the
guys
you
guys
before
presented
on
the
perfect
on
the
on
the
patterns,
it
was
great
and
the
advancement
is
great
and
so
on,
but
there
are
two
different
perspective
where
you
can
see
the
things
from
the
development
and
feature
implementation.
J
This
is
how
we
went
and
also
the
thinking
how
it's
going
to
be
used.
I
don't
want
the
user,
how
is
the
user?
It's
it's
the
sort
of
the
mindset?
Okay,
I
have
this
deployment,
so
it's
like
the
management
of
deployments,
or
things
like
that
that
you
have
in
your
in
your
mind.
So
for
me
it
will.
I
have
like
one
question
or
two
questions
it
could
be.
One
would
be
if
it's
already
available
for
sure
it
is.
J
I
just
didn't
look
enough
just
to
take
out
the
source
code
and
run
and
do
exactly
what
is
how
or
or
at
least
one
feature
how
it
was
presented
today.
This
is
advancement
or
how
we
advance
and
so
on
and
the
other
one
will
be
installing
it.
You
know
in
a
setup,
so
what
was
presented
before
you
don't
necessarily
need
to
have
a
what
the
cluster
the
applications
and
things
like
that.
J
It
will
be
installing
it
because,
let's
say
I'm
doing
my
my
normal
deployment
or
stuff
like
that,
and
now
I
run
the
the
perf
to
see
how
it
is
behaving
or
something
like
that
when
I
so
what
I'm
saying
the
second
perspective,
it's
the
perspective
where
I'm
taking
granted
what
the
guys
build
and
I'm
just
using
it
for
my
for
for
the
purpose
of
the
deployment
or
for
these
things,
it's
invisible.
J
C
I'm
sorry
I
was
just
playing
around
with
the
cat.
I
didn't
yeah
yeah.
I
totally
get
what
yes
you're
hitting
on
you're.
If
I'm
getting
you
correctly,
you're
striking
on
a
lot
of
the
value
that
the
oh
yeah.
Thank
you
by
the
way
everybody
we're
like
10
minutes
after
so,
if
you've
got
to
go,
no
offense,
you
know!
Please,
please,
go
we'll
see
you
next
time.
C
The
the
quick
value
is
that
yeah
you
you
want
to
trust
but
verify
with
your
developers
that
you
assume
that
they
haven't
regressed
in
terms
of
performance,
for
what
they're
doing
or
or
that
when
you're
turning
on
a
big
service
mesh
feature
that
it's
not
going
to
slow
the
whole
system
down.
So
you
might
want
some
tooling
like
this
to
check
on
that
for
you
as
you
go
to
do
that
deployment.
J
Yep-
and
I
it
does-
I,
you
know
how
to
do
it
to
combine
and
so
on,
but
for
instance,
for
me,
it's
very
hard
to
put
the
questions
in
a
context
for
the
ones
that
implemented
the
measuring
feature
to
understand
where
I
want
to
go
with.
J
Yeah
yeah,
so
I
don't
know
how
you
do
it,
but
yeah
it's
been
a
year
since
I'm
mostly
on
the
meetups
on
the
on
the
meetings
and
you
you
go,
you
go.
You
go
ahead,
so
also
the
developers
and
so
on.
If
we
isolate
this
too
much,
then
the
it's
the
gap,
it's
it's
bigger,
but
if
it's
isolated
it's
faster
on
on
how
they
they
are
working.
So
for
me
it
would
be
so
what
I
wanted
to
install
and
run
like
development
mode
and
then
also
to
see.
C
J
C
I
see
yeah,
I
think
the
second
part
of
our,
if
I
interpreted
it
right,
like
the
first
part
of
what
you
were
discussing,
was
kind
of
the
merits
of
the
function
or
the
merit
of
the
functionality.
That
measure
can
offer
to
users,
and
yes,
so
tying
the
the
context
together
and
between
static
stuff
what's
running
and
then
the
performance
of
it
as
well
and
kind
of
the
whole
like
the
utility
that
mastery
can
offer.
C
And
then
the
second
bit
of
what
you're
saying
is,
I
think,
you're,
reflecting
upon
the
manner
in
which
this
community,
like
when
we
get
together
the
sequence
by
which
we
demonstrate
things
to
show
progress
and
get
feedback
that,
like
there's
a
little
bit
of
tooling
today,
that
helps
facilitate
making
that
a
little
easier
with
the
edge
channel,
which
is
what
you
were
calling
out.
And
it
was
were
you
and
then
take
me
from
there?
Were
you
suggesting
a
change,
or
were
you
just
sort
of
asking
for
clarification
on
what
that
procedure
is.
J
I
I
have
a
gap
at
this
moment.
I
understand
what
is
presenting
and
so
on,
but
I
I
cannot
go
in
practice
or
where
I
can.
I
can
use
it.
I'm.
J
C
Yeah
yeah
totally,
I
mean
that
that
highlights,
I
think,
two
things,
or
at
least
two
things.
Three
things
like
one
is
in
that
when
you
go
to
the
docs,
is
there
a?
Is
there
a
service
mesh
patterns,
a
guide
to
using
service
mesh
patterns,
and
that
explains
the
concept
and
kind
of
walks
you,
through
an
example
of
like
why?
What
a
pattern
is
why
you
would
use
it?
No
there's
nothing
written
in
the
docs
about
this
stuff.
C
So,
like
that's
a
gap,
one
and
then
two,
if
you
as
a
as
a
integrated
community
member,
want,
you
know,
you're
being
asked
to
provide
feedback
and
to
facilitate
that,
if
you,
if
you
to
make
that
easier
for
you
to
actually
go,
do
the
thing.
Well,
there
are
certain
aspects
of
what's
developed
here,
like
a
like
when
someone's
updating
the
docs
that
great
there
will
be
a
preview
of
what
that
change
is,
it
will
be
built
and
then
a
url
will
be
given,
so
that
all
of
us
can
go
by
real
quick.
C
We
don't
have
to
go
check
it
out,
build
it,
run
it
open
it.
You
can
just
click
the
link
and
see
what
the
change
was
give
some
feedback.
The
same
type
of
a
thing
can
be
facilitated
with,
like
measuring
ctl
changes,
every
time
that
someone
checks
in
a
change
about
the
command
line,
interface,
a
build
happens
and
a
binary
is
produced,
where's
that
binary
like
why
don't
and
like
we,
we
can
totally
take
that
binary
and
make
it
available
right
there
on
the
pr
that
says,
click
to
download
the
binary
that
was
built.
C
If
you
want
to
test
out
this
new
change
that
this
person
did
so
that
you
can
get
offer
up
feedback
so
that
that
feedback
is
facilitated,
that's
possible,
and
it's
just
missing
today
like
we
should
we
should
do
that.
That's
a
good
ci
centric
thing
for
us
to
that,
would
just
help
with
reviews
and
like
like
that's
part
of
it,
and
then
I
think
the
third
component.
To
that
I
would,
I
don't
know
if
this
is
explicit,
what
you're
saying,
but
I
think
I
would
give
us
an
f
on.
C
Is
the
marketing
centric
like
user
facing
like
hey,
you,
hey
you're,
looking
at
measuring
and
your
question
is:
oh
thank
you
to
venue.
Your
question
is
what
do
you?
What
should
I
do
with
this
thing
now
I
see
you're
saying
it
has
a
lot
of
capabilities,
but
can't
do.
Is
there
like
a
guide?
Is
there
a
couple
of
example
things
I'm
supposed
to
go?
Do.
J
Yeah,
I
have
a
real
example
from
today
from
my
work
today.
That's
why
I
exact
but
yeah,
that's
the
question.
Actually,
yes,
I
have
a
like
a
situation
where
it
seems
that
it
might
be
used
and
the
solution.
I
went
with
the
infrastructures,
the
c
support
and
so
on.
They
have
no
clue
and
the
developers
only
understood
the
part
of
it.
So
I
put
the
yaml
file
in
git
and
okay.
We
create
when
you
do
on
app
number
one
on
the
stack
number
two.
You
do
a
release.
J
J
J
J
C
Good
good,
okay,
fair
enough
all
right!
Thank
you
very
much.
Thank
you
very
good
talk
to
everyone.
Soon,
we're
20
minutes
over
we'll
catch.
You
catch
you
guys
shortly!
Bye
y'all,
see
you
guys.