►
From YouTube: Layer5 Community Meeting (Oct 8th, 2021)
Description
Layer5 Community Meeting - October 8th 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
All
right,
let's
get
started
it's
great,
to
see.
A
So
if
you
have
cameras,
do
turn
them
on.
It
really
makes
it
more
human.
I
guess
instead
of
having
to
listen
to
just
voices
so
yep.
So
let's
get
started
with
today's
meeting.
Welcome
everyone
to
the
larify
community
meeting
today
is
the
8th
of
october,
and
the
main
agenda
for
today
is
definitely
on
kubecon,
which
starts
next
week.
A
So
measuring
and
service
mesh
performance
are
two
projects
that
came
from
this
community,
which
were
eventually
donated
to
cncf,
and
we
have
project
office
servers
for
both
projects
as
well
as
we
will
be
representing.
We
will
be
represented
in
the
cncf
tag:
networks,
service,
mesh
working
group,
so
many
of
the
initiatives
by
the
cncf
service
mesh
working
group
comes
from
the
lyrify
community
projects
like
get
nighthawk
service,
mesh
performance
and
so
on.
A
B
Sure
yeah
thank
you,
er
well,
nice
to
meet
you
all.
My
name
is
mario
arriaga,
I'm
from
mexico,
so
I
just
joined
because
there
was
this
open
source
community
event
over
here
at
mexico,
latin
america
online.
So
well,
it's
a
very
nice
thing,
part
of
this
community.
B
I
have
five
years
of
experience
with
automated
testing
in
different
tools
and
languages,
and
lately
I've
been
working
with
a
cypress
and
some
other
node.js
toolset
in
a
development
team
over
here
at
f5,
as
well
as
several
ci
cd
experience
pipelines
to
our
forum
pollumi
and
right
now,
I'm
trying
to
pick
up
kubernetes.
B
So
it's
pretty
cool
to
be
part
of
this
community
to
you
know,
learn
and
contribute
back
to
all
this
ecosystem
specifically
to
to
measure.
So
that's
all
I
have
for
now.
Thank
you.
C
Great
to
have
you
mario,
that's
good,
I
actually,
if,
if
all
that
you
do,
if
all
that
you
do
is
just
say
good
morning
occasionally
like
that's
such
a
refreshing
comment
for
me,
there
are
lots
of
folks
on
the
call
who
it's
typically
not
morning
for
them,
so
I'm
in
this
perpetual
habit
of
like
saying
good
morning
kind
of
awkwardly
to
most
people,
and
so
so
so
so
thanks
for
that
already
it's.
D
It's
it's
really.
C
Really
nice
to
have
this
really
that's
a
really
nice
introduction.
The
let
me
ask
you
this
of
the
organization
or
the
group
or
of
the
product
or
like
of
your
focus
within
f5.
Is
that
on
a
particular
product
or
is
it
a
a
team?
That's
orthogonal
to
you
know.
B
Yeah,
I
think
I
get
the
question.
Thank
you
lee
yeah.
So
right
now,
I'm
working
with
this.
It's
a
ihealth!
So
it's
this
diagnostics
and
reporting
web
application.
So
they
have
this
classic
on-prem
version
which
it's
been
in
production
for
more
than
10
years
and
that
right
now,
first
they
were
trying
to
do
a
lift
and
shift.
And
when
I,
when
I
joined,
they
were
looking
for
like
a
whole
cloud
version.
B
You
know
remake
it
so
I'd
say:
yeah,
it's
a
specific
product
where
we're
now
we're
aiming
for
2022
to
have
this
cloud
version
ready
at
least
for
internal
users,
so
f5
develops
or
offers
hardware
and
software
for
application,
delivery,
networking,
security
and
cloud
automation,
trying
to
do
a
lot
of
things
and
right
now
they're
trying
to
we're
trying
to
make
our
customers
to
have
like
a
really
seamless
experience,
because
they
have
the
opinion
or
the
perspective
that
are
the
products
are
like
very
separate:
they're,
not
really
cohesive,
so
they're
buying
a
lot
of
companies
and
products
like
volterra
nginx,
some
other.
B
Oh
okay,
yeah.
I
wasn't
aware
of
that,
but
yeah
and
basically
I
decided
to
join
because
you
know
we
have
this
diagnostics
engine
in
our
on-prem
version,
so
there
it
has
more
than
three
thousand
heuristics
and
the
cost
of
migrating
that
that
code
over
to
typescript
to
the
new
firm
engine.
We
were
we're
doing
it's
like
very
high,
so
so
we're
we're
we'll
we'll
be
lifting
and
shifting
that
part
and
they're
like
yeah
we're
moving
into
a
container.
But
I'm
like
hey
wait.
But
what
about
the
performance?
B
C
Yeah,
that's
great
yeah.
I
used
to
lead
a
well
the
engineering
for
seagate's
cloud
and
one
of
the
teams
that
I
had
we
had.
It
was
actually
it
was
kind
of
a
weird
team.
It
was
a
collection
of
like
it's
about
there's
about
40
people
in
this
one
team,
and
there
was
like
I
don't
know
like
20
phds.
Most
of
them
were
focused.
D
C
But
the
the
moral
of
the
story
is
or
the
the
interesting
part
is
it
has
to
do
with.
Maybe
the
heuristics
that
you're
talking
about
is
that
we
were
going
through
a
partnership
with
ibm
at
the
time
and
part
of
what
the
team
was
designing
was
a
well
I
won't
throw
in
more
acronyms
but
say
I
imagine
it's
somewhat
similar
to
a
sas
based
eye
health,
a
lot
of
telemetry
coming
from
the
hardware
systems
themselves.
Oh.
B
Yeah
we
process
a
quick
view,
which
is
just
a
tower
and
it's
sometimes
a
tar
with
tars
horror
balls.
So
it's
like
a
lot
of
xml
processing,
really
heavy,
because
you
know
software
and
hardware
for
the
for
the
that
the
customers
use.
You
know
it
generates
a
lot
of
information,
so.
C
Yeah
is
the
heuristics
that
you're
referring
to
are
those
are
those
those
are
for
determining
health
in
an
intelligent
way?
Is
that
what
those
are
about.
B
Well,
it's
mostly
there's
different
articles
at
f5's
knowledge
base.
So
there's
specifics
of
the
this
is
for
big-ip
systems.
B
Honestly,
wasn't
aware
they
existed
until
I
joined
f5
to
be
honest,
but
basically
these
big
ips
they
have
their
own
kernel.
You
know
it's
kind
of
like
running
besides,
I
don't.
I
don't
know
which
flavor
of
linux
or
or
unix,
but
basically
it
generates
different
information
based
on
you
know
their
own
modules.
B
So
if
there's
a
specific
error
that
an
article
would
let
our
end
users
know
it's
either
some
let's
say:
what's
the
name
when
it's
vulnerability,
it
could
be
a
vulnerability,
something
failed.
You
know,
there's
something
configuration-wise
it
it
wasn't
done
properly
or
there.
You
know
something
has
to
be
done
so
there's
errors
being
thrown,
so
it
would
look
through
different
xml
files.
So
that's
the
ones.
B
I
know,
for
example,
if
if
a
cpu
fan
or
or
a
chassis
fan,
isn't
working
correctly
in
the
hardware
there's
specific
errors,
it
throws
and
are
heuristics.
Sometimes
you
know
they
look
for
those
errors
and
they
have
different
rules
that
are
defined
right
now
there
it's
it's
an
old
language.
I
just
forgot
the
name:
it's
not
pearl,
it's
not
it
rhymes
with
lint.
No,
but
it's
not.
B
A
E
B
That's
kind
of
it
there's
other
kind
of
heuristics
and
I'm
not
familiar
with
the
3000
of
them
yeah,
but
yeah.
C
Cool
okay,
yeah,
so
not
not
here
is
six
in
the
sense
of
machine
learning,
but
here
is
six
in
the
sense
not
yet
like
yeah.
I
have
a
rule-based
system
that
where
humans
have
captured
like
what
you
know,
look
for
this
thing
and
then
if
you
see
that
then
look
for
this
yeah,
that's
nice!
That's
great!
There's
a
lot
of
there's
a
similar
concept.
That's
needed
in
measuring!
C
There's
an
open
issue
describing
a
feature
to
be
added.
Well,
it
describes
it
for
measuring
cli
measuring
ctl,
it's
well,
it's
a
rip.
What
would
it
would?
What
do
we
call
it?
The
is
a
report
measuring
ctl
system
report
or
system
diagnostics,
but
it's
the
concept
that
it's
not
a
dissimilar
concept
in
so
much
as
when
someone
bumps
into
a
bug,
it's
like
okay,
great,
hopefully,
they'll,
come
and
open
it
up,
and
hopefully.
C
Yeah
yeah
but
yep.
So
so,
like
a
diagnostics
package,
yeah
yeah,
I
mean
a
very
similar
concept
where
hey
there's,
there's
meshri
has
logs
it's
helpful.
When
people
drop
those
logs
into
the
issue,
it's
painful
for
people
to
have
to
go,
collect
the
logs
and
then
describe
the
environment
and
then
maybe.
C
Yeah,
actually
the
tar
ball
that
actually
it
says,
zip
here,
but
like
the
tar
ball
that
you
just
mentioned,
it's
a
very
similar
concept.
Hey
just
you
know,
grab
the
local
logs
grab,
some
diagnostics
info,
some
environment,
yeah
cool
and
then
that
way
when
they
come
it's
just
it's
just
a
convenience
operation
like
that,
such
that
when
they
come
to
open
an
issue
that
they
can
just
drop
it
in
and
then
it
becomes
consistent
for
the
rest
of
the
community
who's
trying
to
go
through
and
figure
out.
C
C
Yeah
yeah
people
might
be
a
little
more
encouraged
to
like
when
they
come
to
some
some
projects.
You
go
to
you're,
going
to
fill
in
a
form
for
a
defect,
and
the
maintainers
have
done
a
great
job
of
of
saying,
okay,
here's
the
template
and
then
you
just
sort
of
scroll
down
the
screen.
It's
like
you
know,
fill
in
all
this
info,
like.
C
Yeah,
it's
very
it's
very
much
needed
it's
painful
to
the
end
user
who's
like
well.
You
know
what
I
you
know:
it's
not
even
worth
it.
You
know
forget
it,
but
but
if
they
made
that
a
click
button
thing
like,
if
so,
it
can
be
done
either
or
both
from
mesri
ctl,
which
is
what's
described
here
or
andor
from
ashrae's
ui,
either
of
its
two
clients
to
facilitate
that
so
anyway,
yeah
yeah.
It's
it's
things
like
that.
Like
eye
health,
that
aren't
aren't
why
people
come
looking
for
big
ip.
C
C
So,
speaking
of
so
since,
since
we're
totally
putting
mario
on
the
spot,
this
whole
time
there's
this
coming
week,
there's
the
mesh
the
the
recently
renamed
bi-weekly
call
meshery.
I
think
it's
the
build
and
release
meeting
it's
on
thursday
at
one
o'clock,
central.
C
It
is
a
great
place
to
like
peel
back
the
layers
on
integration
testing
on
site.
Like
there's.
It's
not
it's
not
ju.
It's
on
cypress,
which
is
great.
You
know
really
helpful
for
our
integration
testing
on
what
we're
doing
there
have
an
understanding.
It
was
like
what
we're
missing
there.
There's
a
test
plan.
That's
currently
in
a
spreadsheet
today,
there's
like
there's
a
document
called
test
strategy
which
has
really
like
no
no
strategy
defined
in
the
doc
today,
and
it's
not
like
it
needs
to
drone
on
for
a
long
time.
C
It
doesn't
really
need
to,
but
it
would
just
identify
some
things
like
hey.
We
do
a
lot
inside
of
github
actions,
there's
a
lot
of
workflows
that
that
execute
those
are
really
helpful,
particularly
for
ongoing
unit
and
integration
testing.
We
can
spin
up
a
kubernetes
cluster
within
as
a
pr
is
sent
in.
We
can
spin
up
a
cluster
exit,
build
that
code,
execute
that
code
all
within
the
same
all
before
we
merge,
but
a
lot
of
that
infrastructure
is
there.
C
There
are
too
few
people
involved,
curating
it
and
adding
additional
approaches
to
the
additional
tests
or
approaches
to
the
test.
That
is
like
cyprus
in
particular
like
we
just
we
just
dipped
a
toe
into
the
water
like
six
months
ago
or
something
and
then
we
no
one,
has
come
over
to
grab
it
by
the
horns
and
wrangle.
It.
A
All
right,
we
have
one
more
newcomer
on
this
call
nico.
Would
you
like
to
introduce
yourself.
A
From
india
and
currently
I'm
learning
one
stack
and
I
have
to
contribute
to
some
project
way
to
do
the
front-end
part,
and
recently
I
was
just
hovering
over
some
guitars
and
open
source
looking
for
some
open
source
communities.
So
I
just
came
across
this
I
before
joining
this,
I
even
looked
up
for
the
youtube
video
and
it
was
quite
good
and
that
engagement
was
quite
fascinating.
Engaging
to
me,
so
I
thought
give
it
a
try.
Okay,
so
that's
it.
A
That
sounds
really
good
glad
you
had
a
good
time
good
on
boarding
process
here.
Welcome
to
the
community
nicola
really
great
to
have
you
here.
Did
I
miss
anyone
else?
Is
anyone
else
new
on
this
call.
A
If
so,
please
introduce
yourself
or
chat
or
turn
on
your
mic
and
introduce
yourself
all
right.
So
we
have
a
couple
of
announcements
today,
so
nitish
karthik,
who
has
been
contributing
to
mashri
for
a
while.
Now
us
was
nominated
as
a
maintainer
and
it's
I
think,
it's
time
that
we
announced
all
the
positive
responses
from
everyone
and
that
he
is
now
a
mashery
ui
maintainer.
So
congratulations.
E
This
wasn't
what
I
was
expecting,
but
yeah,
I'm
really
glad
that
this
happened
like
I
always
try
to
be
like.
I
always
try
to
own
mystery.
I
mean
like
not
in
not
in
a
good
way.
I
mean
like.
I
feel
that
mysteries
I
would
like.
I
treat
myself
my
own
project
and
I
try
to
do
everything
that
I
would
do
to
keep
it
up
and
running.
So
I
really
don't
have
anything
to
say
right
now,
but
yeah.
I'm
really
glad.
C
Is
an
important
concept,
and
hopefully
those
you
know
all
of
you
that
are
on
the
call
end
up
with
this
feeling,
either
immediately
or
or
very
soon
that
there's
a
bunch
of
stuff
to
do.
Mastery
has
a
messagery
as
one
of
the
projects
that
are
here
actually
is
one
of
two
cncf
projects
that
we've
donated
recently
we've
got
get
nighthawk,
that's
underway
as
well,
that
those
won't
be
the
last
projects.
Mesher
itself
has
quite
the
roadmap,
it's
intentionally
at
a
dot.
Six
release.
C
Suffice
to
say,
there's
a
ton
of
things
to
do
here
and
it'll
take
more
than
the
number
of
people
that
are
on
this
call
right
now.
It's
important
that
that
you
do
get
a
sense
of
ownership.
C
You're
going
to
screw
things
up,
which
you
see
you'll
follow
my
lead
in
this
regard.
You'll
screw
things
up
like
and
that's
okay,
but
but
that's
part
of
what
it
takes
to
become
an
owner
of
the
thing
or
form
an
opinion
about
the
thing
and
and
if
people
aren't
doing
that,
if
people
don't
feel
like
they're
welcomed
to
do
that,
we're
going
to
fail.
C
So
little
did
nitish
know
like
that's
in
fact,
at
the
core
of
why
it
is
that
he
was
nominated
for
a
maintainer
role
and
why
it
is
that
he
was
overwhelmingly
you
know
ushered
in
or
that
nomination
was-
and
that's
like
it's
for
any
number
of
other
reasons,
but
like
one
of
the
central
themes
to
what
nitish
had
done
was
to
say
wait.
A
second
like
this
is
this
is
amazing.
There's
a
project
here.
C
That's
got
300
something
contributors
to
it
already
and
just
went
to
the
cncf,
and
it's
just
you
know
off
to
do
even
bigger
and
better
things.
But
it
looks
like
any
number
of
these
contributors
have
been
kind
of
focused
on
here
and
now
on,
like
the
thing
that
they're
trying
to
accomplish,
and
not
taking
ownership,
not
stepping
back
and
saying
wait
that
may
not
be
good.
While
that
accomplishes
this
issue
and
we
can
get
this
pr
merged.
C
Is
that
sustainable?
Is
that
consistent
with
the
rest
of
the
code?
You
know?
Would
I
want
to
come
back
to
this
a
few
months
later
and
be
responsible
for
that
like
and
that's
the
attack
that
nation
was
taking?
It's
not
only
that
he
was
taking
the
attack,
but
also
performing
that
work
like
you
putting
his
fingers
where
his
mouth
is,
that's
a
thing
and
yeah.
So
people
generally
don't
like
suggestions
generally
don't
get
rejected
around
these
parts.
C
They
usually
get
embraced,
and
I
think
sometimes
maybe
that
surprises
people,
because
things
are
suggested
but
not
not
followed
up
on
necessarily,
but
that's
okay
as
well,
but
just
because
you're
suggesting
something
doesn't
mean
that
you
need
to
go.
You
know,
do
it
and
I
mean
like
there's
some
ground.
I
don't
know
like
a
half
of
most
of
you
on
the
call,
already
kind
of
know
these
things
or
have
felt
them,
but
there's
there's
other
there's
other
norms
about
the
community
that
are
healthy
to
understand.
One.
C
Is
it's
okay
to
call
our
babies
these
projects
ugly,
like
if
that's
the
case,
they
need
to
be
called
other.
We
need
to
poke
holes
to
them
so
that
they
aren't
like
that.
One
of
the
one
of
any
number
of
potential
downfalls
for
the
the
projects
one
is
bugs
like
features,
attract
people
bring
them
in
mesry.
Does
what
manages
how
many
service
measures,
like
that's
ridiculous,
there's
no
other
project
in
the
world
that
does
that
none.
Nor
will
there
likely
ever
be
not
to
the
degree
that
measuring
will
it's
pretty
unique.
C
That's
a
pretty
cool
statement
to
say
and
to
say
it.
Factually
okay,
we
should
go
contact
the
guinness
world
book
of
records
or
something
if
they
really
care
boy,
if
guinness,
if
they
actually
tracked
software
stuff
that'd
be
a
thick
thick
book.
But
point
is
there's
some
awesome
things
that
we're
doing
grab
grab
a
piece
of
the
projects.
There's
there's
a
bunch
of
edges
to
latch
on
to
call
it
ugly
beat
it
up,
poke
the
holes
in
it.
C
C
Someone
go
over
and
try
to
use
the
thing
and
stumble
sub
your
toe
on
an
issue
and
then
complain
about
it.
Like
that's,
really
helpful,
it
really
is
helpful
and
then
mario
will
go
write.
An
automated
integration
test
to
you
know
cover
that
next
time,
but
but
it
is
really
helpful
because
because
if
you
don't
stub
your
toes,
someone
else
will
and
they
may
not
take
the
time
to
complain
about
it
and
and
that
issue
will
just
perpetuate
people,
have
a
bad
experience.
C
They'll
tell
their
friends,
their
friends
will
tell
their
mom,
their
mom
will
tell
their
and
then
a
word
will
get
around
and
instead
the
hopefully
the
word
that
will
get
around
is
like.
Oh,
you
know
it
was
didn't,
do
everything
I
needed
for
it
to
do,
but
it
was
delightful
or
like
you
know,
it
worked
boy.
I
really
wish
they
would
do
this
and
it's
okay
to
say
that
that's
a
those
are
road
map
things.
C
A
This
congratulations
again,
so
I
think,
let's
get
on
to
some
topics.
We
are
already
almost
half
an
hour
in
so
the
main
topic
for
today
is
measuring
and
service
smash
performance
at
kubecon.
A
If
you
joined
the
call
little
late,
I
was
talking
about
measuring
and
snp
being
two
projects
from
the
layer
5
community,
which
were
donated
to
cncf
a
couple
of
months
ago.
So
as
cncf
projects
we
will
be,
we
will
be
represented
at
kubecon,
which
is
which
happens
at
next
week.
So
we
have
two
project
office
hours,
one
for
measuring
and
and
another
for
service
mesh
performance.
A
We
also
have.
We
will
also
be
represented
at
the
cncf
tag-
networks,
service,
mesh
working
group.
So
some
of
the
projects
that
are
of
interest
of
the
service
measure
working
group
come
from
the
larify
community,
so
get
nighthawk
is
one
such
project
and
service.
Mesh
performance
is
also
another
project
that
we
discuss
at
the
service
mesh
working
group,
so
yeah.
A
The
links
to
the
events
is
shared
in
chat
and
other
than
the
office
servers.
So
we
also
have
a
bug
bash
session
at
kubecon.
So
when
kubecon
kicks
off
on
on
11th,
there
will
be
a
bug
bash
event
where
contributors
can
pick
up
issues
and
they
can
get
points
for
fixing
the
issues
and
the
people
with
most
points
get
some,
I
think,
get
some
prices
and
also
get
some
bragging
rights
so
yeah.
The
current
contributors
can
also
participate
in
the
webpage.
A
And
lee,
would
you
like
to
add
anything
else.
C
Very
briefly,
the
come
join
the
office
hours
for
either
the
two
projects.
It's
a
good
way
to
learn
about
them
if
you're
still
new
and
if
you've
been
here
for
a
while
like
come,
come,
join,
come
represent.
We'll
call
you
out
if
you,
if
we
see
you
there
there'll
be
a
few
of
you
that
will
be
panelists
there
and
be
helping
tell
people
about
the
projects.
C
C
We'll
also
have
a
talk
at
servicemeshcon
for
those
that
might
that's
kind
of
a
day,
one
or
a
day,
zero.
You
know
add-on.
We
are
going
to
debut
what
will
be
a
really
popular
plugin
for
meschery
will
debut
mesh
map
the
so
service
mesh
con
all
about
service
meshes.
It's
a
couple
of
tracks
a
two-track
day.
I
think.
C
We'll
talk
about
service
mesh
patterns
in
that
talk,
service,
mesh
patterns
are
becoming
fundamental
to
what
measuring
is.
There's
a
book
being
authored
now
on
service
mesh
patterns
through
o'reilly,
there's
some
details
about
it
on
the
layer,
5io
website,
but
all
of
the
patterns
that
will
go
into
that
book
and
potentially
a
subsequent
book
they'll
be
implementable
by
measuring.
As
a
matter
of
fact,
measuring
is
called
out
prominently
in
that
book
is
like
in
the
intro
chapter.
C
The
second
chapter
is
nothing
but
about
measuring,
like
here's,
how
you
get
going
with
measuring
here's,
how
you
can
take
and
import
patterns
into
measuring,
and
so
so
that's
what
that
service
mesh
con
talk
is
about
service
mesh
patterns
about
the
book.
Therefore,
it
ends
up
being
about
measuring
measure.
It
gets
demoed
on
on
stage
as
well
as
mesh
map
gets
demoed
a
little
bit.
So
so
that's
exciting.
So
that's
another
talk
there
in
general.
C
I
would
say,
like
hey:
go
whether
you've
done
a
ton
of
work
here
or
not
go
you
know,
go
represent,
go
tell
people
about
well
like
a
some
of
the
resources
that
are
here
you
all
of
you
like
suhani
is
on
the
call
have
assisted
with
or
led
the
way
on,
the
interactive
labs
that
are
free
on
this
site.
There's
a
book.
One
of
the
books
here
is
free.
Actually,
two
of
them
are
free.
C
C
C
So
there's
a
lot
anyway
go
tell
folks
about
the
resources,
like
it's
a
great
way.
Most
people
in
the
world
that
you
know,
95
percent
of
the
people
or
of
those
that
are
in
tech
still
don't
know
what
the
heck
a
service
mesh
is
and
how
it
works
and
what
you
can
do,
and
all
that
and
so
acknowledging
that
the
communities
puts
effort
into
into
all
this.
So
it's
good
it's
good
to
go
helpful
to
people.
Hopefully
that
will
get
them.
Some
exposure
to
meshi
meshri
was
originally
created
in
part
as
a
learning
tool.
C
A
hosted
mesh
where
people
can
just
come
sign
in
dink
around
with
an
actively
running
well
measuring
the
provision
service
meshes
deploy
sample,
apps
change
rules
like
see
how
all
this
stuff
works,
that'll,
hopefully
be
a
great
service
for
again
for
like
learning
getting
familiar
good
but
anyway,
having
said
all
that,
you
know,
go
be
active
at
cubecon.
C
Learn
some
stuff,
maybe
come
back
and
share.
If
somebody
did
this
community
call,
we
talk
a
lot
about
measuring,
but
in
the
past
sometimes
all
we've
done
is
talk
about
how
the
community
is
doing
like
how
how
many
visitors
are
there
to
the
different
sites
that
we
have,
how
many
people
are
signing
in
signing
into
meshary
how
many
people
are
in
slack
how
they
joined?
What's
that
experience
when
they
do
join,
we
really
try
to
take
a
lot
of
time
to
curate
that
experience.
What
does
the
contributor
ladder
look
like
here?
C
C
So
one
thing
that
we
should
do-
I
think
we've
almost
done
this
in
the
past,
but
you
know
like
this:
this
zoom
we
should
get
a
some
some
measuring
and
some
some
project
ones
or
some
layer,
five
ones
for
as
resources
for
people
there's
a
lot
of
resources
in
the
community
drive.
If
you
don't
know,
as
a
matter
of
fact
recently,
I
think
there
was
a
copy
of
squid
game
sitting
in
there.
C
So
just
fyi
we
had
our
first
formal
incident
reported
even
though
yeah
hey,
don't
the
community
drive
is
not
for
not
for
copyrighted
copyrighted
material.
So.
C
You
know
some
of
the
meetings
will
be
canceled,
specifically
the
surface
mesh
performance
meeting
the
next
service
mesh
performance
meeting
we'll
have
a
googler
jacob,
sobun,
he'll
join
and
he'll
talk
to
us
all
about
c,
plus
plus,
with
respect
to
nighthawk
that'll
be
that'll,
be
exciting,
not
because
he's
a
googler,
but
because
he's
going
to
go
because
we've
been
waiting
to
like
really
dig
into
adaptive
load
controllers,
which
are
going
to
be
pretty
interesting,
he's
going
to
cover
that
so
next
week
do
we
have
anyone
on
the
call
that
thinks
that
they
might
like
to
host
the
website's
call
someone
who's
a
little
bit
has
been
a
little
bit
familiar
with
that
one.
C
Think
about
it
and
if
you
you
think
you
might
want
to
you're,
not
quite
sure
what
that
entails,
and
then
ping
me
or
I
would
say
another
community
manager,
but
a
couple
of
them
are
out
and
ping
sayatin
goes,
but
yeah,
otherwise,
we'll
just
we'll
just
cancel
for
next
week.
That's
not
a
big
deal.
C
Cool,
I
think
those
are
the
these
are
the
kubecon
updates.
We
normally
have
more
talks
than
we
can
handle,
and
this
this
cubecon
is
is
the
same.
It's
part
of
my
goal
for
most
of
you,
whether
you
desire
for
it
to
be
or
not
utkarsh
is
still
trying
to
hold
off
on
me,
but
it's
part
of
my
goal
for
for
me
not
to
be
on
stage
anymore,
but
for
others
for
all
the
rest
of
you
is
part
of
why
anita
is
presenting
at
kubecon
china.
C
C
Cool
well
also,
I
don't
know
how
many
of
you
know:
ritac
martika
she's,
a
director
of
cloud
native
engineering
at
intel,
will
probably
nominate
her
as
a
maintainer
on
smp.
So
there's
a
couple
of
s,
p
maintainers
that
are
on
the
call
now.
So
just
heads
up
to
start
thinking
about
that,
and
that's
it
for
me.
A
Yep
thanks
lee,
so
I
hope
to
see
you,
everyone
at
the
project
office
servers
at
kubecon,
so
we
also
have
an
update
from
darren
on
helm,
charts
it's
installing
measuring
with
helm,
charts
so
darren.
Would
you
like
to
talk
about
it
and
hopefully
share
your
progress
with
the
measuring
cd
update.
D
Yeah
sure
first,
can
I
shoot
my
screen
so
that
I
can
explain
things
a
bit
better.
D
D
So
it
fails
to
install
the
nginx
service
mesh,
because
this
guy
the
label
namespace,
doesn't
have
the
permission
to
do
the
pre-installation.
I
believe
I
don't
really
know
like
what
this
means
at
the
moment,
but
I
believe
this
is
not
going
to
be
a
big
issue
and
luckily
the
other,
so
I
also
tested
like
even
before
dpr
got
merged.
D
I
also
tested
you
know
most
of
the
installation
for
the
other
adapters
and
then
they're
all
working
the
nsm
and
then
the
console
the
istio
was
also
working,
but
then
the
osm
one
is
also
a
bit
tricky
because
it's
failing
in
the
init
container,
but
I
don't
think
the
issue
comes
from
the
hem
shot
it.
The
error
message
is
saying
that
it's
not
able
to
create
default
mesh
config
because
neither
the
config
nor
the
master
was
specified.
D
I'm
not
really
sure
if
this
issue
existed,
you
know
have
been
existing
or
not
because
oh
it
has
been
okay
yeah.
So
it's
not
the
yeah.
That's
what
I
thought
because
it
doesn't
look
like
it
is
a
hamstring
issue,
so
yeah.
So
for
my
own
testing,
you
know:
that's
the
only
hem
chart
issue
that
I've
found,
but
since
there
are
like
other,
you
know
like
much
more
other
functionalities
within
the
system.
D
I
really
do.
Do
you
need
some
help
on
testing
other
functionalities
because
so
far
I've
been.
You
know,
testing
the
installation,
for
you
know
for
each
adapters.
That's
what
I
have
been
testing
since
I'm
still,
you
know
new
to
community
and
I'm
still
not
aware
of
other
functionalities
and
yeah.
So
that's
the
only
small
issue
that
I've
found
so
far
and
in
terms
of
the
progress
of
using
those
hamstrings
in
the
mastery
ctl,
I
did
have
something
to
share
so
so
for
the
current
implementation.
D
D
But
you
know
the
mastery
part
is
only
used
latest
when
the
context
version
is
latest
and
the
context
channel
is
stable.
That's
the
point.
I'm
a
little
bit
confused
like
why
it
is
this
case.
I
can
show
you
the
source
code
that
I'm
referring
to
so
yeah.
Here
is
the
pause.
So
basically
we
we
change
the
so
so
actually
we
got
this.
D
D
I'm
not
I'm
going
to
be
confused
about
the
implementation,
to
be
honest
and
then,
and
then
you
know
after
we
change
this
image
version.
We
basically
you
know,
call
this
function
to
actually
go
into
the
mmo
file
in
the
file
system
and
then
modify
the
image
version,
and
then
we
apply
the
manifest
using
helper
functions
and
then
for
the
adapters.
We
always
install
the
you
know
the
latest
version
about
the
image
version
with
the
latest
tag
image
tag
so
yeah,
I'm
not
really
sure
about
that
part.
D
But
besides
that
the
migration
is
pretty
simple
for
the
system
start
and
stop,
but
for
the
update
command,
it's
a
little
bit
tricky
because
for
this
for
the
start
and
stop
we
have,
you
know
the
apply
hand,
chart
helper
function
from
the
mesh
mesh
kit
library,
so
under
so
under
the
hood.
It's
basically
just
you
know
equals
to
this
ham.
Install
you
know,
release
and
then
charge
is
basically
same
as
this
command
that
we
do
manually.
D
D
I
think
it's
a
bit
cleaner
because
you
know
we
don't
need
to
mess
with
the
the
file
system
for
the
user
so,
but
for
the
system
update
traditionally
with
ham,
we
can
do
the
ham
upgrade
command.
D
D
But
the
thing
is
that
for
the
current
implementation,
we
don't
really
have
a
helper
functions
to
help
us
do
the
hem
upgrade.
So
I
was
thinking
to
add
the
helper
function
to
the
match.
Kit,
library,
you
know-
let's
say
you-
know:
hem
well-
upgrade
hem
charts
to
basically
support
these
commands
so
that
we
can
keep
supporting
this
system
update
command
that
we
have
been
having.
D
So
so
in
a
world
that
I
think
the
I
do
have
a
clear
idea.
You
know,
but
I
have
like.
I
have
the
implementation
for
the
start
and
stop,
but
you
know
I
still
need
a
little
bit
more
time
for
the
system
update,
because
it
requires
some
implementation
for
the
magic
kit,
library
and
then
also
need
to
brief,
but
a
bunch
of
exciting
existing
implementation
using
all
those
manifest
and
then
that's
the
well.
The
the
only
part
that
I'm
not
clear
about
is
the
image
version.
C
A
C
A
If
that
is
a
use
case,
but
it
wasn't
done
because,
like
we
ideally
want
users
to
always
have
the
latest
adapters
or
latest
machines,.
D
Oh,
so
so
the
mastery.
So
by
default
we
do
the
so
it
basically
follows
the
mastery
cto's
version
for
the
mastery
part.
Is
it
correct.
A
Yep
so
yeah,
so
basically
yeah
you
can
change
the
version
in
the
measuring
config
file
and
yeah
that
the
docker
image
that
was
released
at
that
particular
version
will
will
be
pulled.
So.
A
That
was
the
logic
behind
it
and
since
the
adapters
are
released
separately,
they
don't
have
the
same
version
as
they
don't
have.
The
yeah.
A
D
C
It
was
there
like,
didn't,
wasn't
able
to
complete
the
remainder
of
the
the
task
so
to
speak
to
this
there's
so
there's
two
two
release
channels
stable
edge.
C
They
each
have
this
concept
of
latest
stable
latest
edge
latest
the
the
late
the
quote-unquote
latest
is
becomes
this
keyword
for
a
subscription
and
the
the
notion
being.
C
What
nubindi
was
saying
is:
is
that
ideally
user,
like
actually
is
kind
of
interesting
in
the
time
that
we've
done
this
docker
desktop
or
docker,
rather
has
changed
their
free
version
of
docker
desktop
like
if
you're
gonna
use
the
free
version
you'll
be
forced
to
upgrade,
and
it
makes
sense
like
they're,
sustaining
costs
and
things,
and
so
for
measuring
it's
kind
of
the
same
thing.
We
wanna
get
people
on
the
latest
stuff,
one
because
bugs
are
fixed
two,
because
there's
new
functionality
and
so
yeah
by
default.
C
When
you
create
a
context
in
your
mesh
config
well,
by
default,
it
would
be
on
the
you'd,
be
subscribed
to
the
stable
you'd,
be
subscribed
to
the
stable
channel
and
subscribe
to
being
stable,
hyphen
latest,
and
so
good
people
can
switch
between
channels
if
they
want
to.
D
C
D
C
D
C
That's
one
thing
by
the
way
to
your
other
question
about
the
notion
that
hey,
why
are
we
writing
to
file
system
when
we
could
avoid
complexities
that
are
involved
with
like
keeping
deployment
files?
Local
well
and
there's?
That's
a
great
point.
The
the
logic
behind
that
is
the
user.
There's
any
number
of
users
who
want
air
support
for
air
gap
deployments,
and
so
they'll
want
to
so
they'll
need
deployment
manifests.
I
use
the
word
manifest
in
that
broader
sense
of
like
whether
it
was
docker
composed
that
they'll
need
those.
You
know,
local.
C
That
brings
in
the
unwanted
complexity
of
in
part
what
you
were
alluding
to,
which
is
okay.
Well,
so
if
they
deploy
one
version,
great
they've
got
a
copy
of
that
that
one
version's
manifest.
C
D
C
C
Well,
it
isn't
so
much
the
download
part
as
it
is
the
seek
source
them
locally
apart
and
so
in
order
to
facilitate
that
it's
like
well,
if
someone
you
know,
if
they're
on
a
machine
and
they
switch
networks
to
the
air
gap
network
that
they
would
yeah,
I
mean
we
have
to
we're
still
not
quite
it's,
okay,
that
people
might
have
to
manually
download
the
helm,
charts
but
place
them
in
that
folder,
but
at
least
the
infrastructure
within
measure
ctl
is
there
to
source
locally.
First.
C
F
D
Yeah
yeah
yeah
yeah.
I
was
saying
that
you
know
it
also
supports
getting
the
hem
shot.
You
know
from
online
urls
like
without
like,
like
avoiding
download
the
hamstring,
because
I
didn't
really
know
the
idea
behind
donald
d
manifest.
But
now
I
do
understand
you
know
after
lee
explained
that
to
me.
D
D
Yeah,
I
think
I
get
it
now
I'll
try
to
finish
up
the
pr
either.
Well,
probably,
I
need
to
wait
until
weekend.
Well,
yes,
sunday,
I
can
submit
appear
by
the
end
of
sunday.
I
believe
I'd
like
to
cover
to
fix
the
entire
thing,
but
the
system
starts
stop
and
then
the
update
command.
F
Photo
for
a
week
actually,
instead
of
probably
instead
of
creating
a
new
function,
we
can
have
a
new
attribute
in
this
drop
that
apply
himself
takes
a
like
is
upgrade
the
reason
for
doing
that
could
be
that
it
would
be
that
a
lot
of
things
plan
chart
does
right
now,
which
we
even
like
we
will
have
to
replicate
in
a
upgrade
hem
chart
like
referencing
local.
Hence,
so
there
are
tons
of
things
that
apply
himself
is
capable
of
doing
right
now.
D
Yeah,
but
there
are
two
different
functionalities
right,
so
I
thought
you
know
isolated
would
give
a
little
bit
cleaner
implementation
like
in
terms
of
the
implementation
itself,
yeah
yeah.
We
can
certainly
you
know,
extend
what
we
have
in
the
apply
hem
chart
function
or
the
api,
where
we
call
it
right,
but
so
yeah.
So
this
is
the
application,
and
then
I
traced
it
down
to
the
to
the
end
of
the
code.
D
It
basically
used
the
hem
api
about
the
hem
library
from
hem
itself,
so
they
have
a
golden
library
and
then,
at
the
end
of
it,
just
basically
use
the
you
know,
create
this
hamstring
and
then
it
yeah
I
shouldn't
configure.
Indeed
that's
how
it
do
the
ham
install
you
know
from
the
code,
wise
yeah,
I
could
yeah.
I
can
do
it
that
way,
really
up
to
you
like
what
we
really
want.
C
Yeah,
I
think
that
yeah,
it's
probably
a
little
easier
to
be
dry,
to
not
repeat
ourselves
if
we're
augmenting
with
an
is
update,
parameter,
which
it
kind
of
that's,
it's
actually
it's
kind
of
interesting,
because
it's
sort
of
the
same
line
of
questioning
that
we're
asking
ourselves
here
in
like
a
separate
function
versus
augmenting
the
you
know
augmenting.
What's
there
it's
like
kind
of
the
same
thing
with
the
the
charts
versus
kubernetes
manifests
like
the
more
the
left
is
less
surface
area
to
get
things
wrong
more
like.
D
Well,
if
you
look
at
the
you
know,
the
ham
is
the
implementation
for
the
command,
so
they
have
the
install.go
and
the
upgrade.go
which
matches
to
hem
install
and
have
in
upgrade.
There
are
two
different
objects,
so
you
know
these
run
well.
They
both
implement
the
same
interface
right
golden
interface,
but
then
it
comes
it's
coming
from
different
objects,
so
this
guy's
certainly
coming
from
the
install
objects
that
owns
this
run
method
and
then
the
upgrade
we
have
the
same.
You
know
same
function
signatures,
but
then
it's
coming
from
the
upgrade
objects.
D
D
F
Actually,
when
I
was
working
on
the
flying
chart,
I
did
notice
that
the
actually.
That
is
the
reason
that
this
generate
action.
You
would
say
upper
order
function.
High
order
function
exists
because
it's
instructing.
F
What
action
it
was!
It
tries
to
abstract
that,
so
I
did
notice
that
yeah.
They
have
in
fact,
at
some
places,
upgraded
and
installed.
They
differ
also,
but
is
to
to
the
extent
that
our
use
cases
are
or
were
we.
We
were
able
to
successfully
struck
that
out
and
put
this
put,
that
behind
a
generic
interface,
so
probably
the
reason
as
far
as
I
could
think
after
looking,
I
also
actually
referred
to
help
him
score.
F
So
as
far
as
I
could
think,
the
reason
they
had
different
structs
is
because
there
is
an
overlapping,
but
for
some
functions
they
the
interfaces
actually
differ
so
for
them,
probably
it
made
sense
to
have
different
structs
so
that
they
can
keep
the
implementation
separate,
but
for
our
use,
cases
like
which
are
pretty
minimal,
it
hasn't.
F
Yet
it
hasn't
yet
sort
of
came
back
to
bite
us
when
when
we
are
actually
putting
it
behind
the
generic
interface,
that
is,
we
are
just
creating
our
upper
order
function.
We
have
tried
to
abstract
this
out
of
in
this
particular
function
reduction.
I
hope
that
makes
sense.
D
Okay,
okay,
yeah!
So
if
that's
actually
this,
it
could
be
used
for
upgrade
as
well
right
because
it
only
takes
the
name
from
the
contest
this
okay
yeah.
I
can't
pick
from
the
charts
the
choice.
Okay,
yeah.
I
can
use
yeah.
I
can
try
to
update
the
system,
update
command,
to
use
the
same
functions
to
see
if
it
works.
D
That's
really
helpful,
yeah!
That's,
I
think.
That's
all
I
have
you
know
for
the
hem
shots.
So
please
you
know,
let
me
know
if
you
know
I
know
there
are
some
other
people
also
testing
the
hamstring
insulate.
Well,
not
not
the
installation
part,
but
you
know,
after
using
hamstring,
to
install
the
system.
I
know
the
you
know
we,
as
you
know
some
other
people
that
are
more
familiar
familiar
with
the
system
to
help
testing
that.
So
please,
let
me
know
if
there
are
any
issues,
but
that's
the
only
one
that
I
found
so
far.
A
Thanks
darren
still
now
I
didn't
face
any
issues.
Apart
from
the
osm
install
that
you
mentioned,
I
will
keep
on.
D
A
All
right,
so
we
are
almost
at
time
so
before
we
end
the
call,
I
just
wanted
to
bring
everyone's
attention
to
some
outstanding
prs.
So
some
of
these
prs
are
for
the
upcoming
release
and
there
are
40
outstanding
prs
in
the
mastery
repo.
So.
D
C
There's
a
couple
of
quick
items
on
some
megan
is
on
the
call,
so
is
zane,
so
megan,
megan
and
zane,
at
least
for
my
part,
I'll,
be
spending
a
little
bit
of
time
with
you,
either
asynchronously
or
synchronously
over
the
next
couple
of
days
about
the
the
blogs
for
the
dot
six
release.
So
thank
you.
Thank
you
for
stewarding
them.
For
my
part,
I
will
finally
turn
my
attention.
There.
That's
good.
C
Oh
and
yeah
we'll
try
to
so.
I
don't
think
tunisia
is
on
the
call
right
now,
but
he
had
been
working
on
a
short
what
we
call
a
short
marketing
video
for
to
to
improve
what
we
have
on
measuring.io.
C
B
C
C
A
Yep
I'll
try
to
go,
go
through
some
of
this
and
do
that
as
well.
All
right
everyone,
that's
it!
Thank
you
for
joining
this
week's
call
see
you
next
week
at
kubecon,
bye.