►
From YouTube: Meshery Development Meeting (Sept 23rd, 2020)
Description
Welcome @Techwinx, @oluwabamikemi!
A
B
D
A
A
E
E
E
F
F
E
E
E
No,
actually,
no,
that's
fine!
Actually
I
was
I
had
my
whole
zoom
off.
I
was
still
still
looking
at
bomi.
Anyway.
Sorry
there's
been
a
rash
of
of
a
healthy
interest
in
internships
within
the
community,
which
is
awesome.
We
do
a
number
of
internships
and
have
been
for
a
few
well
two
years
now.
I
guess
we
end
up
doing
internships
in
a
few
ways.
E
So
some
of
those
are
through
programs
like
google
season
of
docs
google's
summer
of
code,
we've
just
we've
chatted
with
the
major
league,
hacking
folks
in
the
past,
and
are
doing
a
little
bit
of
chatting
with
them
right
now
we
haven't
gotten
engaged
in
outreachy,
but
we
have
been.
We
have
been
engaged
in
community
bridge
and
this
post
from
yesterday.
E
What
is
today,
it
just
says
this
winter
yeah,
this
post
from
this
week
says
that
we're
participating
in
some
service
mesh,
centric
community
bridge
projects,
and
they
are
they're
kind
of
described
toward
the
bottom
of
the
the
post.
I
think,
and
so
there
have
been
there's
been
a
lot
of
interest.
It's
been
great.
There
are
only
so
many
internships
to
go
around.
We
do
layer,
five
as
a
as
a
company
does
offer
direct
internships.
E
E
I
also
will
recognize
that.
Well,
maybe
I'll
make
a
quick
analogy
and
say
those
of
you
who
have
attended
high
school
or
attended
university.
You
go
you
go
there,
there's
a
structure
to
the
learning
path.
There
are
learning
paths
to
find.
There
are
books
that
you
can
read
and
lectures
to
listen
to
teachers
to
engage
with.
E
I
think
you
could
say
much
of
the
knowledge,
maybe
not
the
wisdom,
but
maybe
much
of
the
knowledge
that
you
could
learn
about
a
particular
subject.
You
could
probably
go
over
pick
up.
The
book
read
read
through
the
same
text:
do
the
labs
or
do
the
things
you
probably
walk
away
with
about
the
same
knowledge,
which
is
to
say
that
you
don't
necessarily
have
to
be
in
a
structured
environment
to
get
gain
that
same
experience
or
gain
that
same
knowledge.
E
E
Hopefully
the
zoom
is
still
coming
through.
Okay,
but
it
is
about
accountability.
People
saying
here's
the
thing
to
do:
here's
a
chapter
to
read:
there's
an
exam
coming
and
there's
a
deadline
here.
You've
got
to
do
these
things
and
it's
helpful.
You
know
the
reason
I'm
drawing
this
analogy
is
to
say
I
consider
you
could
still
if
you
didn't
have
to
have
the
deadline.
If
you
were
stronger
than
me
and
could
force
yourself
to
read
the
book,
you'd
probably
gain
the
same
knowledge
and
suffice.
So
what
I'm?
E
What
I'm
saying
here
is
there
are
only
so
many
internships
for
those
that
inevitably
don't
make
it
into
an
internship.
That
doesn't
mean
that
you
can't
gain
the
same
knowledge
or
even
more
by
just
engaging
meaningfully
on
an
aspect
of
a
project
here,
if
you,
if
you've
done
that
by
the
time,
you're
done
at
least
I'll
speak
for
my
part,
we
will
have
had,
we
will
have.
We
will
have
had
quite
some
relationship.
E
We
will
have
talked
a
lot,
which
means
that,
for
my
part,
anyone
who's
giving
of
their
time-
hopefully
learning
things
here,
but
but
giving
and
making
the
community
better
and
the
projects
better.
I'm
gonna
do
everything
within
my
vast
power
to
to
help
you,
which
means
which
could
mean
any
kind,
all
kinds
of
things
introductions
to
people.
E
So,
okay,
that's
part
of
it,
the
other,
so
people
have
been
asking
hey.
I'd
like
to
the
internship,
looks
pretty
neat
I'd
like
to
learn
more
I'd
like
to
get
the
internship.
How
do
I
best
to
go
about
doing
that?
The
answer
is
pretty
simple
and
if
you
think
about
it
for
a
moment,
it's
kind
of
intuitive
the
way
to
best
position
for
an
internship
is
engage.
E
Those
are
the
individuals
that
get
an
internship
like
actually
in
many
respects,
like
that's
very
fair,
so
the
answer
is:
if
you
want
an
internship,
I'm
kind
of
spending
a
long
time
on
this,
because
there's
been
like
50
something
people
that
have
asked
in
the
last
week.
You
know
how
to
how
to
land
one
of
these
and
it's
it's
just
that
it's
identify
an
area
of
need,
pick
it
up
go
forward.
E
If
this
round
goes
by
and
you
don't
you-
either
don't
land
one
directly
or
through
one
of
these
other
programs
by
the
way
shredi.
If
you
could
visit
the
layer,
five
dot
io,
maybe
just
the
the
main
site
itself
or
the
slash
pro
the
slash
programs
link.
It
just
lists
out
the
fact
that
we're
layer
5
offers
direct
internships,
but
also
through
these
other
programs.
What
I
was
going
to
say
is:
if
you
don't
land
one
right
now,
we've
been.
E
E
E
E
Cool
all
right,
so
the
first
topic
up
so
into
mesh
redevelopment
stuff
first
topic
was
well.
I
was
hopeful
that
clement
might
who's
just
joined.
The
community
he's
been
getting
familiar
with
the
code
base
and
things
we
were
gonna.
I
was
just
gonna
suggest
he
look
at
this
area,
but
I
don't
think
he's
on
so
we
will
catch
him
next
time.
It
actually
means
that
he.
E
E
I
was
just
going
to
suggest
it
to
him,
so
the
the
topic
is
that
we've
recently
created
two
release
channels
for
meshary,
a
stable
release
and
an
edge
release
right
now,
when
anyone
starts
when
anyone
downloads
mesh
free
and
begins
to
just
use
it
unbeknownst
to
them
they've,
they
have
implicitly
chosen
to
use
this
stable
release
or
to
be
on
the
stable
channel
as
it
stands
in
the
project.
Currently
there
aren't
there
isn't
control
or
facilities.
E
E
You
know
in
the
settings
area
you
should
be
able
to
like
switch
between
stable
channel
and
an
edge
channel
by
the
way,
these
two
release
channels
stable
means
that
we
release
less
frequently
things
are
better
tested.
You
should
expect
to
be
able
to
point.
You
know
that
you
would
have
less
issues
there.
E
I
think
this
is
fairly
intuitive
for
all
of
us,
but
people
might
need
to
switch
back
and
forth
between
the
channels
because
they
might
be
looking
for
this
new
feature
that's
out
on
edge,
and
we
should
make
it
easy
to
switch
back
and
forth
one
of
the
ways
there's
two
ways
in
which
we
would
like
to
be
able
to
support
people.
Switching
one
is
inside
measures,
ui
just
go
to
settings
and
kind
of
you
know
toggle
the
switch,
the
ui
component.
That
is
not
there
today.
E
D
E
D
E
Measuring
itself
so
actually
put
it
down
as
a
topic
as
you.
If
you
would
and
we'll
talk
about
that
as
well,
we'll
talk
about
all
the
components
in
there:
okay,
all
right
nice,
the
next
topic
up,
if
no
one
has
any.
You
know
any
anyway-
that
that
invitation
for
for
takers
on
the
release
channel
stuff
is
open
through
half
an
hour
after
the
end
of
this
call,
after
which
we'll
go
out
talk
to
clement
abhashek
is,
is
he
with
us
yep?
He
is
abhishek.
Do
you
want
to
talk
about
the
open
pr?
E
G
Yes
sure
so
so
we
have
like
a
couple
of
years
in
different
places,
basically
on
osm
and
puma
adapters
and
and
the
major
change
we
are
on
measuring
repo
and
also
a
bit
of
them
on
learn
layer
5
as
well.
So
all
of
these
are
a
transition
state.
At
the
moment,
if
I
say
so,
there
are
some
dependencies
with
the
implementation,
wise
or
the
feature
wise.
G
B
A
G
Which,
which
is
purely
a
dependency
update,
that
we
are
stuck
with
and
they've
tried
sort
of
rendering
the
packages?
This
is
purely
and
specific.
This
is
the
kuma
adapter
that
we're
talking
about
here
and
we've
tried
several
ways
to
fix
the
the
stat,
the
checks
before
making
videos,
and
we
we
sort
of
made
it
for
two
or
three
checks,
but
there
are
still
three
more
checks
that
is
yet
to
be
passed,
and
this
basically
depends
on
a
third
party
package
that
that
the
ci
is
not
able
to
fetch
and
yeah.
G
So
this
is
the
same
case
for
the
osm
adapter
so
because
they
both
run
on
run
using
the
same
package
version.
So
they
are
the
same
dependency
check
errors.
Apart
from
that,
the
machine
repo
has
the
changes
ready
for
the
the
recent
feature
that
we
worked
on,
basically
to
bring
up
the
table
of
semi
conformance
in
the
ui,
and
this
is
still
pending
because
of
the
hanging
checks
that
design
our
adapters
right
now.
G
G
E
That
so,
just
to
add
some
context.
The
two
community
bridge
internships
that
opened
up
one
of
the
one
or
both
of
them
have
very
much
to
do
with
what
abhishek
was
just
showing,
and
that
is
that
mescheri
is
the
official
tool
for
validating
whether
or
not
a
given
service
mesh
is
service,
mesh
interface
or
smi
compliant
and
so
yep.
That's
what
the
first
community
bridge
internship
is
about.
It's
half
about
smi
and
it's
half
about
smp
service
mesh
performance.
E
A
C
C
Instead
of
continuously
adding
the
details
same
details
again
and
again
suppose
he
wants
to
do
a
test
for
google.com,
for
example,
but
just
want
to
edit
some
values
and
perform
the
same
test.
So,
instead
of
creating
entering
the
data
again
and
again,
the
person
user
will
have
a
test
profile
saved
and
he
can
just
click
on
a
button
and
run
the
test.
C
So
we
have
this
feature
being
implemented
in
mastery
ctrl
current,
but
the
ui
part
which
lacks
so.
I
was
working
on
it.
It's
been
a
while
now
not
able
to
give
much
time
to
it
so
yeah
I
have
some
progress
over
here.
I
will
be
showing
that
in
just
a
second,
so
this
is
the
basically
the
performance
profile,
the
page
will
which
will
get
open
up
when
the
user
first
logs
it
and
doesn't
have
any
profile
created
so
on
creating
on
clicking
this
button
create
a
test
profile.
E
Nikhil
I'm
going
to
be
in,
or
I'm
going
to
risk
being
rude
to
interrupt
just
for
a
moment
just
because,
because
I
can
help
help
myself
but
give
context
for
everyone
else.
I
think
you
might
have
said
this,
but
I
just
want
to.
I
want
to
be
sure
that
everyone
understands
that
what
what
nikhil
is
doing
here
is
is
well,
it's
actually
quite
powerful.
It's
pretty
convenient
today.
E
If
you've
used
measuring
you
can
you
can
see
that
you
can
go
over
to
a
performance
area
fill
in
of
a
form
to
run
a
performance
test
measure
will
generate
load,
calculate
statistics
on
how
well
the
mesh
is
doing,
etc.
Graphic
chart
all
these
things
you'll
do
some
mathematics
that
I'm
you
know
that
are
over
my
head
or
but
the
the
pain
in
the
rump
with
that
is
that.
E
Well,
if
you
want
to
rerun
the
same
test
or
run
a
test
of
a
different
configuration,
you
have
to
go
back
and
fill
back
in
the
form
and
when
you
want
to
run
it
again,
you
have
to
go
back
and
fill
in
the
form
and
what
nikhil
is
trying
to
achieve
here
and
well?
Well,
we'll
achieve
here,
he's
got
you
know.
Half
of
it's
done
is:
is
the
convenience
of
filling
in
the
form
identifying
the
fact
that
you're,
you
know
understanding
that
you're
probably
going
to
run
this
test.
E
Maybe
every
time
you
do
a
software
release
every
time
you
do
a
configuration
change
in
your
service
mesh.
Maybe
you
want
to
see
what
how
that
has
affected
the
performance
of
the
mesh
that
you're
going
to
want
to
run
the
same
test
a
bunch
so
so
he's
bringing
forth
the
ability
to
save
the
cat,
the
parameters
of
the
performance
test
as
a
profile
to
give
it
a
name
and
be
able
to
recall
that
profile
and
run
that
same
performance
test
profile
over
and
over.
C
Oh
yeah
delicious,
so
this
is
the
page
which
I
guess
most
of
you
or
all
of
you
in
the
call
are
familiar
with.
So
this
is
a
page
from
where
we
then
run
the
performance
test
and
currently,
after
entering
all
the
fields
where
the
test
name
is
not
mandatory,
but
we
can
enter
and
we
can
run
the
test
directly
from
here
or
we
can
create
a
test
profile
which
will
ask
for
a
profile
name
which
we
can
enter
and
then,
after
confirming
it,
a
new
profile
will
be
created.
C
C
C
E
C
Yeah
we
can
do
that.
Actually,
I
had
a
question
regarding
that.
That
is
it
possible
that
we
can
enter
manually
because
last,
which
I
know
from
the
update
from
kanishka,
the
test
id
gets
generated
automatically
which
is
stored
so
which
we
are
actually
the
profile
id
so
which
we
can
get
the
profile
later
on.
E
Oh
for
sure,
having
like
we
want
to
proliferate,
ids
every
which
way
ev
it
should
be
like
the
the
oprah
show.
Like
you
get
an
id
you
get,
everybody
gets
an
id
like
they're
super
cheap.
We
can
spin
up
a
good
all
the
time
and
we
should
definitely
every
object.
Everything
in
here
should
have
its
own
identifier,
every
instance
of
a
profile
of
a
performance
test
run
and
every
performance
test
profile.
E
E
You
know
that
our
environment
is
doing
well
now,
there's
not
really
much
that
changes
for
for
them
in
that
profile,
so
they
want
to
go
back
and
play
it.
You
know
on
a
weekly
basis
at
some
point,
though
they
decided
to
add
two
more
nodes
to
their
cluster.
So
at
some
point
they
changed
their
environment
and
that's
significant,
like
that's,
going
to
affect
the
behavioral
characteristics
of.
E
What's
going
on,
they're
we're
going
to
want
to
empower
them
with
a
faucet
with
a
either
the
ability
to
version
their
performance
test
profiles,
so
they
could
say
look.
This
is
still
the
same
profile.
It's
still
our
normal
weekly
soak
test,
but
when
we
compare
historically
over
time
something
changed
and
it
could
be
something
that
they
changed
in
the
profile
itself.
Maybe
they
decided
not
to
run
it
for
an
hour
every
week,
but
like
90
minutes
every
week,
or
they
added
the
two
plus
two
new
nodes
to
their
clusters.
E
It
may
not
be
a
version-
that's
the
most
appropriate
faucet
to
to
track
this
change.
It
might
be
more
of
like
well
an
annotation,
a
label
if
you
will
just
so
that
as
they're
looking
back
historically
historically
so
anyway,
I'm
getting
into
a
deep
example.
I
guess
my
point
is
yeah.
We
want
to
give
an
id
to
every
object,
yeah
having
a
name
having,
I
think,
like
with
just
from
a
ux
perspective.
E
If
you
go
over
to
the
performance
test,
page
people
are,
if
again
like,
I'm
gonna,
I'm
gonna
be
self-centered
here
and
and
make
a
guess
that
there
are
some
other
people
in
the
world.
Like
me,
that
will
go
to
a
page
and
when
the
first
field
says
test
name
and
they
all
of
the
sudden
have
to
draw
upon
their
creative
juices
to
like
come
up
with
a
name
for
this
stupid
thing
like
I
can
barely
do
it
every
time
I
face
that
text
field.
E
So
my
point
is:
if
they've
done
it
once
there,
and
then
they
go
to
save
it
as
a
profile
minim,
if
we're
going
to
ask
them
for
yet
another
name
and
ask
them
to
muster
all
of
their
creativity
for
a
second
time,
then
we
minimally
should
suggest
to
them
that
it
is
test
one
that
it
is
the
same
name
as
the
test
that
they've
given
it,
but
I
actually
would
argue
that
we
just
shouldn't
ask
them
again.
It
should
just
be.
E
E
E
H
C
C
Yeah,
so
I
think
adding
a
profile
name
would
be
a
better,
but
if
that's
repetitive,
which
I
agree,
the
point
you
mentioned
that
if
they
enter,
they
think
why
is
it
asking
for
it
again
for
the
name
again,
then
I
guess
we
can
put
the
same
test
name
as
the
profile
name
and
instead
of
popping
up
showing
up
the
model
for
entering
the
name.
We
can
show
up
a
small
snack
bar
stating
your
profile
has
been
added.
E
It
just
assumes
that
they're
going
to
keep
the
same
name
between
the
test
and
the
profile
it
creates
the
profile
when
they
go
over
to
the
profiles.
Excuse
me
over
the
profiles
page
that
it
would
be
at
that
point
that
they
would
have
the
ability
to
change
the
name.
Potentially
we
don't
have
to.
We
don't
have
to
do
that.
I
think
for
sure.
I
think
we're
agreeing
that
if
there
is
a
separate
profile
name
that
it
would
be,
that
field
would
be
auto
populated
with
the
test
name
as
well.
E
C
A
A
A
A
E
Teach
folks,
speaking
of
internships
kush
is
just
wrapping
up
the
google
summer
of
code
internship
and
he
was
working
on
enhancing
the
performance
tests.
That
nikhil
was
just
showing
he's
building
in
support
for
a
third
load
generator.
This
load
generator
is
called
nighthawk
it's
from
the
envoy
project.
E
It
does
things
a
little
bit
differently
than
the
other
generators.
It's
not
to
say
that
one
is
necessarily
better
than
the
next,
but
they
have
different
algorithms
that
they
use
to
perform
statistical
analysis
of
the
response
time
and
the
latency
and
the
throughput
and
things
so
he's
building
in
support.
For
that
that's
interesting,
but
what
makes
it
really
much
more
interesting.
Is
that
part
of
what
is
being
worked
on?
There
is
the
ability
not
to
just
yeah.
E
G
So
I
was
just
away
from
keyboard
for
around
a
minute.
I
guess
so.
The
presentation
which
I
wanted
to
give
was
that
my
talk
being
bundled
with
mystery
ctl,
so
that
if
you
just
download
mystery
ctl,
you
will
be
able
to
use
nighthawk
as
it
is
like
which
you
do
with
photo
right
now.
The
following
feature
is
not
available
with
wrk2.
G
E
A
I
A
I
I
I
I
I
use
it
in
quite
a
few
use
it
in
quite
a
few
projects,
they're
quite
easy
to
set
up,
but
they're
quite
easy
to
set
up
quite
easy
to
set
the
tests
typical
test
tests
looks
like
this.
For
instance,
this
test
should
return
expected
supported
operations
like,
for
instance,
an
operation
called
console,
install
an
operation
called
install,
install
book
info
and
so
on
and
so
forth.
I
So
this
helps
me
to
feel
safe.
When
I
refactor
the
code,
there
were
absolutely
no
tests
in
the
measuring
console
adapter
and
one
can
sort
of
like
think
about
where
which
kind
of
tests
do
you
want
to
start
writing
when
there
are
no
tests
at
all,
and
of
course
that
can
be
discussed.
But
an
easy
approach
is
because
one
doesn't
necessarily
know
how
the
internals
of
the
system
work
but
know
how
the
system
as
a
whole
is
expected
to
behave
that
one
starts
with
them.
I
I
E
I
E
I
H
A
I
I
G
E
E
But
that
means
that
every
time
we
go
to
do
a
release,
it's
a
scary
moment
because
we're
not
quite
sure
that
everything
is
working.
It
really
slows
us
down.
So
this
this
is
nice.
I
have.
I
have
a
couple
of
technology
questions,
so
yeah
part
of
I'll
make
some
statements,
maybe
who
you
can
tell
me
if
these
are
accurate.
E
E
I
So
so,
but
I
think
when
we
get
there,
we
we
we
just
have
to
discuss
it
and
have
to
look
at
the
consequences.
Well,
my
sort
of
you
know
my
my
dream
would
be
that
either
it's
in
in
these
pipelines
or
that
we
have
some
books
somewhere.
That
is
running
all
these
tests
like
every
night
or
every
day
or
yes,
automatically
so
yeah
yeah.
E
And
to
your
point,
like
the
it's,
not
just,
there's
kind
of
two
two
issues
with
having
two
separate
test
frameworks,
the
one
that
you
mentioned,
which
is
g,
is
just
the
cost
of
creating
the
both
and
sustaining
them
both,
but
then
also
to
the
extent
that
you're
seeing
failures
in
one,
but
not
in
the
other,
or
that's
just.
You
know
that
and
so
yeah.
If,
as
we
look
to
push
forth
with
bats
and
potentially
run
that
you
know
continues,
you
know
all
the
time
as
part
of
the
process.
E
It
might
be
that
the
creation
of
an
action
there
is
I'm
assuming
is
probably
relatively
straightforward
to
the
extent
that
you
just
you
know
it's
a
linux
environment,
you're
grabbing
to
run
the
the
item.
That
might
be
a
little
more
that
might
make
putting
this
together
a
little
more
interest.
Interesting
is
the
spin-up
of
a
small
kubernetes
cluster.
I
Exactly
I
was
actually
more
worried
about
that
than
running
getting
bats
to
run
so
here
I
think
shipyard
could
really
be
great
shepard
io,
but
I'm
not
quite
sure
how
this
works
on
you
know
in
in
in
the
context
of
a
github
pipeline,
whether
that's
easy
to
actually
achieve
or
not
getting
up,
kubernetes
cluster.
There.
E
E
I
I
E
It's
very
much
neat.
I
mean
this
whole
like,
irrespective
of
the
technology
like
this,
is
very
much
needed
in
the
linker
d.
Adapter
is
where,
if
memory
serves
that's
where
suco
was
building
in
the
deployment
of
a
of
a
kind
cluster,
I'm
just
looking
through
it
now,
but
actually,
if
you
don't
mind
sort
of
tagging,
asuko
and
maybe
nick
jackson
as
well,
to
discuss
a
little
bit
about
shipyard.
E
The
suku
and
I
were
having
a
conversation
recently
in
which
you
know
part
of
you
know
as
we
look
to
as
we
look
to
refactor
the
mesher's
adapters
and
provide
support
for
provisioning
service
meshes
in
a
couple
of
different
ways,
because
because
some
of
those
ways
are
universal-
and
some
of
them
aren't
like
a
universal
way
might
be
like
the
most
universal
in
a
kubernetes
environment
might
be
here's
a
set
of
manifests
for
a
mesh.
You
just
push
to
manifest
like
that,
should
be
the
case
that
you,
you
should
be
able
to
deploy.
E
All
meshes
like
there
are
some
service
mesh
projects
that
really
don't.
How
do
you?
How
do
I
phrase
it?
It
seems
like
they
don't
want
you
to
do
that,
because
they
and
you
know
for
good
reason.
They
want
to
have
a
good
experience
and
they
want
you
to
customize
the
deployment.
So
they
want
you
to
use
either
helm
or
maybe
they
don't
want
you
to
use
anyway.
E
My
point
is,
it
looks
like
inevitably
we're
gonna,
probably
you
have
to
support
one
or
two
or
maybe
three
mechanisms
of
provisioning
service
meshes
and
my
I'm
intrigued
at
the
pos,
by
the
possibility
of
maybe
one
of
those
mechanisms
being
shipyard
and
I've.
It's
worth
a
conversation
with
nick.
He
very
much
so
is
a
friend
of
the
community
and
wants
wealth
for
what
we're
doing,
and
I
think
it
would
maybe
help
his
project
as.