►
From YouTube: Meshery Development Meeting (Aug 18th 2021)
Description
Meshery Development Meeting - Aug 18th 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
Welcome
everyone
to
the
mystery
development
meeting
today
is
18th
of
august,
and
we
have
some
agenda
lined
up
today
and
if
you
have
any
other
topics
that
you
would
like
this
is
to
add
them
to
the
meeting
minutes.
A
We
have
a
couple
of
announcements
before
we
start
today's
agenda
and
if
there
is
anyone
else
who
is
new
on
the
call
who
didn't
get
the
chance
to
introduce
yourself,
please
do
that
on
the
chat.
We
would
love
to
hear
about
you
and
the
areas
that
you
are
focused
on
as
well.
A
A
So
some
of
you
folks
who
are
on
this
call,
have
been
working
on
this
action
and
we
are
in
process
of
so
so.
This
action
is
being
adopted
by
one
of
the
service
mesh
projects
currently
so
open
source
mesh
is
in
the
process
of
adopting
this
github
action
and
yeah.
We
hope
to
that
also
that
other
service
meshes
also
adopt
this
action,
so
that
is
on
the
roadmap,
and
it
will
be
done
soon.
A
You
can
check
out
this
action
here.
The
link
is
intercept
and
another
topic
is:
we
are
hosting
a
cncf
webinar
on
measuri,
so
I
believe
it
it
will
go
live
on
tomorrow,
so
you
can
rsvp
using
the
link
given
in
chat
cap,
and
we
will
also
have
some
more
webinars
coming
down
the
road
soon,
one
with
cncf
students
as
well.
So
some
of
these
are
down
the
road
moving
on
to
topics.
A
So
this
sunday
we
had
a
bug
hunting
session
for
mashery.
So
there
were
many
people
who
joined
the
call
and
tested
out
measuring
and
we
had
a
productive
session
where
we
identified
a
lot
of
bugs
so
linked
in
the
meeting
minutes.
You
will
find
a
dock
for
the
test
plan
measures
test
plan
for
the
next
upcoming
release,
which
is
v0.6.
A
A
We
are
also
proposing
to
conduct
a
a
follow-up
session
this
week
on
saturday,
so
21st
august
at
9
a.m.
Central
time
and
7
30
ist.
So
that's
the
usual
time
and
pay
time
of
this
measure
recall
so
we'll
have
a
more
we'll
have
some
more
bugs
to
hunt.
So
if
you
are
interested
I'll
send
out
an
invite
on
the
community
calendar.
So
if
you
are
interested
join
that
session
as
well
all
right
moving
on
to
dhruv,
do
you
want
to
talk
about
the
static
and
dynamic
boards.
B
Oh,
hey
hello,
guys!
So
if
you
go
through
the
attached
link
over
there,
you
will
be
directed
to
our
world
famous
meshiach
lecture
styles
and
over
here.
B
This
is
an
attempt
to
show
how
we
fetch
the
charts
in
machines
both
rafana
chats
employment,
restaurants.
I
have
demoted
this
previously
to
the
development
call.
So
if
you
want
to
go
into
detail,
then
we
can
check
out
check
out
this
particular
site.
B
I
think
what
we
wanted
to
discuss
today
was
the
readiness
of
this
slide
so
that
we
would
show
it
in
the
next
smp
meeting
or
machine
meeting.
If
I'm.
C
This
particular
piece
of
functionality
is
of
interest
to
not
only
to
measure
but
to
the
service
mesh
performance
project
as
well.
There
are
oh
there's
a
lot
of
metrics
that
that
are
needed
to
be
captured
inside
of
the
s
p,
spec
and
without
prometheus
node
exporter,
like
like
the
the
most
expeditious
way
to
re,
you
know
get
out.
Some
of
those
metrics
is,
is
through
a
prometheus
node
exporter
and
so
yeah
the
there's
a
bi-weekly
smp
call
that
that
takes
place.
The
next
one
is
tomorrow.
C
B
There
are
two
kind
of
codes
and
machines,
one
is
dynamic
mode,
one
is
static.
Now,
what's
the
difference
between
it?
So
basically
dynamic
boots
are
the
one
which
you
have
defined,
which
you
chain
in
the
final
itself,
and
we
just
we
just
switch
it
from
your
griffin
instance
and
show
it
to
you.
So
you
are
able
to
change
it
and
do
whatever
you
want
with
it.
But
then
there
are
some
statics.
B
Now
what
are
the
static
mode?
Whenever
we
do
a
s
p
test,
we
have
defined
certain
things
which
we
track.
While
your
test
is
going
on
as
part
of
the
result
of
your
of
your
performance
test,
which
you
do
so
usually
when
you
run
a
performance
test,
you
will
usually
see
a
detail
about
each
of
the
nodes
which
were
present
in
your
cluster
and
a
certain
specs
which
are
not
over
here.
I
think
they
should
be
okay.
I
will
note
that
note
to
myself
anyways
yeah.
B
So
basically,
we
captured
that
through
the
static
boards
which
we
have
so
that's
the
difference.
That's
how
we
define
static
and
dynamic
road
over
here
in
michigan,
and
usually
the
two
slides
which
you
see
in
front
of
you
are
give
you
a
gist
of
flow
on
how
we
get
the
grafana
charts
which
are
over
there
and
how
we
also
get
the
prometheus
charts
and
there's
also
some
specific
one
for
static
charts
itself,
so
usually
for
all
the
dynamics
chart
whether
it
is
from
grafana
prometheus.
B
We
use
the
the
grafana
apk
itself
to
get
your
chats,
whether
it's
a
grafana
charts
or
it
uses
prometheus
as
a
data
source
or
not.
But
in
case
of
the
static
charge,
we
directly
take
it
from
the
prometheus
sd
gate
from
your
prometheus.
Instead,
instead
of
going
through
the
grafana
instance
to
the
prometheus
acid
data
source.
So
that's
the
only
difference
between
that's
the
major
difference
between
the
static
graphs
and
the
dynamic
graphs
that
make
sense.
C
This
is
great,
and
so
for
is
neja
on.
Oh,
my
goodness,
she
has
been
hot
to
trot
on
this,
and
so
by
the
way,
that's
a
term.
That's
a
term
that
my
wife
really
hates
when
I
use
so
it's
good
that
she's,
not
here
just
fyi
the
sunnae
has
been
raring
to
go
on
this
particular
area,
and
so
we'll
have
to
play
back
the
recording,
because
this
would
be
helpful
to
her
understanding
and
it's
not
just
her
that
we
need
to
be
involved
in
order
to
deliver
this
one
further
step.
C
One
piece
of
analysis
that
hopefully
neha
can
start
on
is
well
is
I'll
both
state
it
and
ask
a
question
at
the
same
time.
So
there's
a
collection
of
there's,
there's
some
metrics
that
are
defined
with
that
are
of
interest
and
they're
defined
inside
of
the
service
mesh
performance
specification.
C
The
spec
is,
there's
always
more
metrics
to
collect,
potentially
like
there's
feedback
on
the
spec
that
that
additional
metrics
would
be
gathered,
and
so
the
this
initial
task.
This
initial
analysis
that
neha
and
maybe
others
can
do,
is
to
look
at
and
it's
very
it's
fairly
readable
and
navindu
or
drew
or
whomever
is
sharing
it.
C
You
might
be
able
to
navigate
to
like
the
s
p
site
to
the
s
p
protos
and
just
give
people
a
pointer
as
to
where
they
can
digest
what
metrics
are
specified
as
mandatory
or
specified
as
being
captured
by
the
spec
today
versus
what
metrics
are
being
collected
in
the
static
board.
There's
a
delta,
I
suspect-
and
that's
that's
part
of
the
analysis
that
needs
to
be
done
so
I'll.
I
don't
know
who's
who's
sharing,
but
I
can
send
you
a
link
if,
if
you
can't
find
it,
you
can
go
to
the
there.
C
C
And
so
it's
you
know,
so
this
is
why
we're
why
it's
good
that
you
are
bringing
it
up,
because
it's
not
unless
you
have
someone
pointed
out-
it's
probably
not
the
most
obvious
or
the
most
legible,
but
within
here
at
some
point,
we'll
start
to
see
things
about
cpus
and
memory
and.
C
C
So
there's
like
the
amount
of
traffic,
the
number
of
millicourse,
the
number,
the
amount
of
memory,
good,
the
amount
of
memory
for
a
proxy,
the
amount
of
memory
for
an
ingress
gateway,
an
amount
of
memory
for
an
egress
gateway
like
if
you're
gonna,
analyze.
How
well
your
mesh
is
running.
Those
are
you
want
to
know
that
and-
and
it's
not
just
that
it's
it's
additional
metrics,
so
the
yep
so
I'll
say
it
for
the
third
time.
C
C
C
Well,
you
know,
most
of
the
service
meshes
come
with
a
small
instance
of
prometheus
embedded
as
like
an
optional
deployment,
because
they
too
want
to
show
metrics
about
how
well
they're
performing,
and
so,
but
they
also
recognize
that
for
the
most
part,
anyone
who
has
a
production
deployment
of
kubernetes
like
that
they
probably
have
already
deployed
prometheus
elsewhere.
C
It's
status,
part
of
the
status
quo
to
have
a
prometheus
around,
and
so,
instead
of
deploying
a
new
one,
that's
specific
to
just
the
service
mesh
control
plane
generally
measuring
would
look
to
connect
to
an
existing
prometheus
here's.
C
The
challenge
is
that,
well,
when
people
stand
up
like
some
of
you
who
are
contributing
to
the
project,
you
might
stand
up
just
a
small
cluster
just
to
do
development
and
check
things
out,
and
so
you
can
separately
deploy
prometheus
or
perhaps
you
might
want
from
measuring
to
do
that
for
you,
especially
with
the
configuration
that
mastery
would
need-
and
that's
part
of
that
analysis
is
like
hey
what
the
node
exporter
there's
different
metrics
exporters
that
you
can
deploy
with
prometheus
the
one
that's
of
particular
interest
is
the
prometheus
node
exporter.
C
A
Thanks
drew
thanks
lee
any
other
comments
here,
so
those
who
would
like
to
help
out
here
and
reach
out
to
richard
in
the
in
the
slack
also
we
will
be
looking
at
some
of
this
in
tomorrow's
snp
call,
so
you
can
join
that
as
well.
D
So
before
I
share
my
screen
so
a
little
bit
about
the
little
bit
about
the
unit
test,
so
we
have
covered
almost
all
of
the
commands
except
two,
so
we
have
mesh
command
remaining
and
the
version
command,
so
anyone
who
wants
to
start
contributing
to
golang.
So
this
is
a
very
beginner
friendly.
The
you
can
write
tests
for
the
version
come
on,
it's
very
beginner
friendly.
If
you
run
into
any
problem,
you
can
hit
push
or
me
or
other
than
that.
D
Yes,
it
is
okay,
so
actually
I
was
having
some
connectivity
issues
today.
So
that's
where
that's
what
I
was
asking.
So
there
is
this
function
validate
version
and
it
is
being
called
in
multiple
functions,
so
it
is
being
called
in
when
we
validate
any
context.
This
function
is
repeatedly
called.
So
what
happens
due
to
this
is
this
makes
these
api
calls
due
to
which
we
are
exceeding
the
github
rate,
limiting
whatever
the
threshold
has
been
set
back
it
up.
D
So
to
avoid
this,
what
I
was
proposing
was
to
actually
cache
these
to
cache
these
versions
in
the
documentary
folder.
So
we
can
have
something
like
this.
This
is
a
test
json,
so
we
can
have
a
versions
field
that
specify
all
of
the
versions
so
like
we
have.
I
have
right
now
I
have
the
latest
version
and
the
metadata
can
be
created.
This
is
a
uni,
basically
epoch
timestamp.
D
We
can
have
this
book
timestamp
that
when
was
this
last
cached
at
and
the
validity
whether
this
particular
version
is
valid
or
not,
and
then
we
have,
we
can
have
a
custom
threshold
right
now
it
has
been
set
to
1r.
So
what
will
happen
is
before
making
the
api
call.
D
It
will
come
here
and
check
if
the
version
is
actually
here
and
if
the
version
is
here,
whether
it
is
valid
or
not,
and
if
it
is
valid
whether
it
has
crossed
the
threshold
or
not,
and
if
all
of
the
things
are
okay,
then
it
returns
true.
Otherwise
it
returns
false.
So
this
is
not
messy
ctrl.
This
is
just
a
demo
file,
so
I
have
we
have
different
functions.
Like
has
valid
cached
version.
This
basically
returns
true.
D
If
we
can
actually
avoid
the
api
call
and
if
we
actually
make
the
api
call
after
the
api
call
is
being
made,
we
can
actually
add
that
version
with
the
this
function.
You
can
call
add
version
and
add
that
version
here
so
that
for
the
next
one
r,
it
wouldn't
hit
the
github
api
and
the
same
can
be
done
for
the
latest
stable
release
tag
so
yeah.
That
was
that
was
it.
D
So
if
anyone
has
any,
if
anyone
has
any
issues
with
it
or
any
recommendation
of
how
this
json
file
should
be
made-
or
you
can
give
your
criticism
or
anything.
D
Yeah,
so
I
have
basically,
I
find
the
time
dot
now
dot,
basically
the
current
unix
time
stamp
and
then
subtract
it
from
the
created
at
if
it
exceeds
the
threshold.
It
means
that
here
you
can
see
this.
Basically
it's
this
line,
number
120.
If
I
can
just
just.
C
C
Oh,
it's,
the
cash,
it's
the
cash,
it's
the
gap
very
necessary
or
very
depending
upon
the
situation,
and
I
guess
the
question
is-
and
I
haven't
I'll
think
aloud
with
all
of
you
for
a
moment,
because
because
getting
this
wrong
was
you
know
not
good
and
the
cli
isn't
the
only
place
that
calls
out
that
mesh
recalls
out
to
github
anytime
that
you
refresh
this.
I
think
I
guess
just
the
dashboard,
maybe
in
measure
ui.
It
also
checks
for
the
latest
version,
and
so
it
doesn't
take.
C
You
don't
have
to
be
too
much
of
a
power
user
to
start
to
broach
the
limit
in
part,
because,
anytime,
that
you
do
a
measuring
ctl
start
command
anyway,
there
may
be
any
number
of
of
calls
out
now
there
a
deity
is
contemplating
or
has
contemplated,
and
so
is
husseina.
C
These
two,
I
think
in
particular,
given
some
thought
toward
like
you
know
the
term
caching
wasn't
used
before
well
or
yes,
it
was,
but
just
the
concerns
around
I'm
an
air
gap
deployment
so
outside
of
a
rate
limit.
What
happens
if
you
don't
have
network
connectivity
to
retrieve
manifest
or
to
pull
down
to
check
on
the
latest
version
and
well
I'll
tell
you
what
happens
is
measure
ctl
doesn't
work,
you
can't
use
it,
so
you
gotta
go
off
and
find
helm
or
something
else.
C
C
We
are
there
other
than
I
mean
so
there's
particulars
about
that
that
need
to
be
considered
for
how
the
cash
would
function
and
it
doesn't
have
to
be
super
complex
like,
but
what
could
be
done
in
lieu
of
a
cache
or
like
there's
some
simple
steps,
and
I
think
push
may
have
done
this
already
or
someone
might
have
already
taken
an
initial
step
and
instead
of
querying
github's
api
for
the
latest
mesh
reversion
with
every
single
mastery
cli
command,
which
is
you
know
in
general,
that
would
have
been
fine,
but
with
the
rate
limit
now
it
is
so
that's
been
reduced.
C
I
think
to
just
like
a
couple
of
commands
so
that
that
reduction
doesn't
eliminate
the
problem,
but
it
reduces
it
to
to
a
point
where
maybe
it's
not
a
problem?
Maybe
that's
not
the
case.
For
me.
I
hit
the
rate
limit
all
day
yesterday,
all
the
day
the
day
before
so
do
think
about
this.
If
you
would,
what
else
could
be
done
if
not
for
cash
and
if
not
for
just
reducing
the
number
of
times
that
it's
called?
C
C
C
C
I
know
that
some
of
you
have
thought
about
this
and
you're,
probably
politely
being
quiet.
Anyone
think
that
iterating
on
design
for
cash
is
like
that,
we're
starting
off
on
the
wrong
foot
that
we
should
be
doing
things
before.
Looking
at
a
cache.
A
No,
like
considering
the
use
case,
I
believe
we
are
going
in
the
right
direction.
I
believe
about
some
of
it
today,
like
we
are
caching,
the
manifest
files
and
using
that,
so
this
one
first
step
we
are
taking
towards
air
gap
deployment
and
also
like,
like
your
work,
seems
like
something
that
we
should
consider
doing,
because
the
version
check
api
call
is
might
be
the
most
might
be
the
leading
cause
for
the
reaching
the
rate
limit.
A
D
Yes,
so
utkarsh
asked:
why
do
we
need
the
valid
field?
Actually,
we
need,
because,
if
user,
basically,
if
we
get
some
invalid
version
from
the
from
the
config.tml
user,
is
going
to
again
and
again
hit
the
github
api
for
that
invalid
thing.
So
we
also
need
to
keep
track
whether
that
thing
is
valid
or
invalid.
So
if
the
valid
field
is
set
to
true,
it
means
that
yes,
it's
valid.
If
it
is
not
set
to
true
what
will
happen
is
we
can
also
cache
the
invalid
things?
D
So
maybe
the
user
is
going
to,
he
has
made
some
typo
or
latest
one
or
latest
two
something
that
that
is
some
kind
of
invalid
version.
So
I
also
need
to
cache
the
invalid
version
so
that
I
can,
I
don't
have
to
go
and
hit
the
github
api
to
know
whether
that
thing
is
valid
or
invalid.
So
I
have
two
states
in
the
cache,
whether
it's
valid
or
invalid
invalid.
B
B
You
can
cache
the
version
that
is
written
in
front
of
the
hammer
and
store
it
there
and
if
it
is
updated,
then
we
actually
hit
the
api
endpoint
and
update
the
json
okay,
something
like
that
not
making
it
this
complex,
keeping
it
really
simple.
B
Okay,
I'll
try
to
explain
what
happens
at
the
end
is
whenever
we
are
actually
trying
to
get
any
measuring
ctr
like
something
from
config,
we
hit
a
hand
like
a
function
that
is
called
validate
version,
okay,
and
that
is
what
the
reason
behind
exceeding
the
github
limit.
Okay,
so
my
point
is:
let's
just
cache
the
stuff
that
we
want
to
kind
of
make
sure
of
it
only
regarding
the
version
that
has
been
written
in
the
config
camera.
B
Okay,
so
we
are
only
caching
it
for
one
hour
and
if
it
is
like
it
was
created
one
hour
before,
and
the
version
is
still
the
same:
okay
4.5.0,
we
will
kind
of
return.
The
response
from
the
cache
part
confirm
it
that
this
is
a
valid
version
and
if
it
is
updated-
and
it
is
not
matching
with
the
cache
thing,
we
will
create
a
new
request
to
github
and
update
this
json
file
and
make
it
like
it
was
created
just
now,
and
it
is
very
for
the
next
one
getting
it.
D
About
thinking
about
what
you
are
saying
but
see,
this
thing
is
dynamic,
I'm
only
caching
what
the
version
is
specified
in
the
config,
I'm
not
the
extra
things
that
I
don't
know
why
it
looks
complicated
to
you.
I
mean
there
are
just
two
extra
fields
just
to
get
just
to
I.
I
also
want
to
store
the
metadata
about
when
this
was
created,
so
that
I
can
dynamically
delete
them
every
time.
The
time
stamp
every
time.
D
D
F
A
Because
a
lot
of
people
has
a
lot
of
suggestions
here,
so,
let's
put
it
on
a
design
spec,
I
can
send
you
the
link
to
create
a
design
spec
ourselves.
C
You
know
yeah
it's
good
to
holistically,
give
consideration
as
to
how
the
system
overall
functions.
It
could
also
be
that
there's
another
an
alternative
way
of
considering.
C
That
is
that,
if
measuring
cli
as
a
client
is
concerned
with
it-
and
you
know
to
take
steps
to
address
either
a
rate
limited
api
or
a
disconnected
network,
because
these
concerns
kind
of
go
hand
in
hand,
so
too
maybe
should
the
other
client
the
ui
have
similar
concerns,
and
maybe
it's
lesser
of
a
concern
of
measuring
server
than
it
is
the
two
clients
I
don't
know,
but
but
yes,
it
doesn't
mean
that
they
both
have
to
be
addressed
at
the
same
time.
B
So
I
added
a
doubt
about
the
rate
limiting
thing
like
so
it
is.
I
guess
good
question
was
done,
so
I
wanted
to
ask
that
slash
api
slash
providers.
Kpa
lbd
is
right.
It
pings
it
pings
meshi
cloud,
but
machi.
Does
that
also
count
as
rate
limiting
when
you
use
that
particular
endpoint
to
get
the
capabilities.
G
A
Like,
since
you
are
going
to
talk
about
caching
of
manifest,
would
you
like
to
show
your
recent
pr.
A
E
So,
as
you
can
see
on
earlier,
kubernetes
did
not
have
a
skip
update
flag,
so
the
flag.
I
have
added
attempts
to
skip
update
in
the
sense
that
it
uses
your
cash
manifests
if
possible.
Now
the
conditions
for
using
your
cash
manifest
include
in
the
measuring
deployment
file.
The
version
given
should
match
with
the
config
file,
the
operator
manifest
should
be
present,
and
so
should
the
adapter
manifests.
E
If
that
is
possible,
it
instead
of
read
or
re-downloading
and
then
starting
you
can
see
it
just
starts.
E
If-
and
this
applies
to
the
operator
manifests
as
well
so
earlier
when
this
flag
in
there
every
time
there
would
be
a
fresh
download-
and
this
is
like
one
of
the
steps
towards
reducing
our
dependence
on
you-
know
a
net
connection
like
an
air
gap
deployment
so
partly
for
docker
deployment,
since
they
use
operator
manifest
too.
This
extends
to
that
as
well,
and
the
next
thing
I'm
going
to
do
is
extend
this
to
the
docker
compose
file
as
well.
E
If
you
guys
have
any
way,
I
could
improve
this
or
anything.
Just
let
me
know.
A
C
They
are
trying
to
they're
being
enhanced
to
provide
better
support
for
individual
versions
of
each
service
mesh
that
they
support
as
they
do
that
they
will
in
the
in
their
repository
of
each
of
the
mesh
readapters
they'll
have
a
pre-parsed
collection
of
information
about
that
specific
version
of
that
service
mesh,
and
given
that
each
service
mesh
has
many,
you
know
publicly
released
versions.
C
The
file
system
of
the
source
code
for
those
adapters
it'll
have
one
folder
per
service
mesh
version,
and
so
the
the
question
is
hey.
Are
we
as
we
as
measuring
ctl
deals
with
air
gap
considerations
and
then,
potentially,
the
user
has
resident
on
their
local
file
system?
You
know
any
number
of
versions
of
manifest.
C
You
know.
How
are
we
going
to
structure
that
is
that
just
all
the
same
manifest
folder
are
those
files,
your
name
differently?
Are
they
structured
into
a
there's,
a
design
spec
that
needs
to
happen
on
this
particular
set
of
concerns?
If
you
guys
would
and
ashish
as
you
go
to
create
this
rate,
limiting
design
spec
that
you
know
more
or
less
has
very
similar
concerns
and
almost
goes
hand
in
hand
with
just
these
air
gap
to
concern.
C
E
C
Okay,
why
do
we
need
that
pr
merge?
First
to
me
that
sounds
like
coding
before
doing
design.
E
No
we've,
I
I
had
a
design
spec
for
this.
It
didn't
have
this
folded
version
thing,
but
whatever
I've
done
in
the
pr
is
in
part
of
the
design,
spec.
C
Got
you
okay,
nice
yeah,
the
forcing
the
rest
of
us
to
review
your
design
species
would
be
good.
Oh
nice.
C
Yeah
yeah,
no,
no,
no,
I
just
mean
like
forcing
yeah
right,
you're
doing
your
job
for
sure,
like
I
just
mean,
and
actually
you're
doing
it
right
here,
you're
like
hey,
forcing
everyone
to
look
at
it,
which
is
good,
which
is
like
it's
a
liability
if
we
don't
collectively-
and
so
this
is
great,
you
know,
if
you
all
would
like
the
overarching
concern.
I
think,
for
this
area
is
kind
of
like
it's
kind
of
like
air
gap
deployments
sort
of
the.
C
I
don't
know,
if
that's
the
best
term,
but
there's
a
lot
of
considerations
within
there
and
as
a
an
area
of
concern
it
drives.
C
Caching
manifests
it
drives
rate,
limiting
concerns,
excuse
me,
and
so,
if
you
would
try
to
tie
those
together,
creating
hyperlinks
between
the
docs
and
things,
so
we
can
get
a
full,
a
holistic
picture
and
the
reason
being
is
like
is
that
there
are
any
number
of
snafus
kind
of
gotchu's
in
something
as
complex.
Now
as
it
is
mastery
ctl,
where,
if
you
like,
it's
easy
to
get
yourself
into
a
catch-22
and
you
just
become
deadlocked,
you.
C
Deploy
with
one
version
on
one
channel,
you
switch
the
channels
change
the
version.
Try
to
then
stop
the
other
like
like
it's
easy
to
get
your
you
to
end
up
in
a
place
where
you
don't
have
the
correct
manifests
available
to
you,
and
so
we
have
to
write
down
the
use
cases
like
the
starting
points.
So
user
is
running
the
latest
version.
Then
they
start
it.
Then
they
restart
it.
Then
they
stop
it.
Then
they
delete
it.
C
E
C
Oh,
it's
good,
so,
for
my
part,
it
puts
me
slightly
on
edge
to
merge
the
pr
without
like
it
is
probably
just
churn.
It's
actually.
One
thing
I
wanted
to
convey
to.
I
think
all
of
us
is
there's
a
number
of
users
who
are
getting
busy
with
mesheri
and
they
are
running
into
any
number
of
bugs,
and
that's
doesn't
mean
that
any
of
you
are
including
myself
are
doing.
C
Images
are
doing
anything
wrong.
It's
part
and
parcel
to
writing
one
line
of
of
code
like
there's,
inevitably
a
you
know
a
bug
in
in
that
line.
C
Folks
are
becoming
hip
to
the
fact
that
there's
kind
of
there's
this
pretty
cool
project
happening
over
here
and
I
gotta
tell
you.
I
don't
know
that
all
of
you
heard
the
story,
but
it
was
samir
and
I
were
gonna
present
to
the
cncf
service
mesh
working
group
a
week
ago,
and
I
woke
up
that
morning,
groggy
completely
having
forgot
that
that
was
actually
today
or
that
day.
C
So,
there's
like
you,
know,
20
to
30,
end
user
large
or
you
know,
enterprise
organizations
that
are
part
of
the
cncf
that
focus
on
service
measures
specifically
because
they're
running
service
meshes
wanted
we're
asking
for
a
presentation
of
measuring
end
of
s
p,
and
so
we
gave
it
to
them,
and
then
they
held
me
on
the
line
for
an
additional
30
minutes
after
that
hour-long
call
because
they
had
so
many
darn
questions,
especially
when
they
saw
a
mesh
map.
C
They
just
well
soiled
themselves,
I
think,
was
sort
of
what
happened
and-
and
I
wish
that
and
and
that's
the
kind
of
feedback
that
I
wish
to
for
all
of
you
to
receive
directly.
It's
no
joke
what
you're
working
on,
but
I
also
got
feedback
from
others
that
well-
and
these
are
you-
know,
principal
architects.
You
know
these
are
these.
Are
these
are
folks
that
know
what
they're
doing
I
just
couldn't
get
the
thing
working
and
in
the
you
know,
they're
referring
to
mastery
itself
or
like
they
ran
into
so
many.
C
You
know
they
ping-ponged
around
different,
bugs
it's
like
so
so
we'll
have
to
reel
in
the
quality.
So
I
gave
that
as
a
backdrop
to
saying
I
don't
know
that
we
should
merge
that
idd,
not
that
it
isn't.
Necessarily.
I
don't
even
know
I
don't
barely
even
know
what
it
is,
I'm
just
using
it
as
an
example
of
like
if
we
know
that
we
need
a
design
spec
for
something.
It's
like.
C
Let's
go,
let's
go
chew
through
that
and
hopefully
have
less
churn
and
there's
time
and
place
by
the
way
for
rip
roaring
through
code,
which
is
what
we've
mostly
been
doing,
especially
if
you
look
at
the
number
of
commits
that
adidi
has.
In
particular
I
was
just
looking
at
that.
That's
fantastic!
We
need
to
get
some
stuff
in
there
and
things
tangible
and
things
to
use.
C
C
Great
I'm
looking
forward
to
the
design
spec
we'll
we'll.
Thank
you
for
writing
one.
Let's
get
some
eyeballs
on
that.
It
looks
like
a
lot
of
people
have
a
feedback
in
this
area.
C
Is
fine
young
well
there's
a
fine
gentleman,
his
name's
phil
olberg.
I
thought
I
saw
a
message.
This
is
really
coincidental
because
I
thought
I
saw
a
message
from
him
in
slack
yesterday
I
thought
and
he's
the
one
he
works
for
a
bank
so
happens
to
be
here
in
austin
where
I
am,
and
air
gapped
is
what
they
need
like
they
can't
and
he
doesn't
really
doesn't
particularly
care
how
many
hoops
he
has
to
jump
through
to
deploy
it
air
gap.
C
He
just
needs
to
know
what
are
those
instructions
and
that
document
that
we
have
hopefully
call.
I
know
it
calls
out
some
already.
So
that's.
Actually
the
starting
point
is
like
what
are
the
concerns?
What
can
a
user
even
do
if
they
are
air
gapped?
In
some
cases,
there's
manual
steps
for
them
to
do?
Let's
start
there,
here's
a
good
point.
E
I
was
also
thinking.
Could
we
like
have
selective
downloads
for
specific
adapter
files
because,
like
that's,
that's
not
a
complete
air
gap
deployment
but
that's
reducing
our
dependency
on.
You
know
the
internet
like
if
the
user
is
okay
using
an
older
file
for
let's
say
sto,
but
wants
a
newer
one
for
console.
They
should
be
allowed
to
specify
that
with
the
adapter's
flag.
C
I
I
think
it
to
be
to
give
you
quick
feedback
is
like
there
are
more
egregiously
outstanding
things
to
go
chase
after
before
that,
like
even
supporting
multiple
versions
like
today,
if
you
go
use,
try
to
deploy
service
mesh
embarrassingly
it
doesn't
measure,
it
doesn't
even
give
you
the
choice
of
what
version
you
want
to
deploy.
It's
like.
Oh,
my
gosh.
We
regressed
and
functionality
there,
but
we're
well
positioned
to
blow
past
what
we
were
offering
previously
previously.
C
It
was
a
bunch
of
sort
of
hard-coded
support
for
hard-coded
versions,
anymore
rudraksh
and
shrayas,
and
shayatan
and
ashish,
and
there's
a
list
of
folks
that
have
been
working
on
making
that
very
dynamic
making
it.
So
it's
a
bit
self-sustaining
and
that
there'll
be
a
long
history.
A
long
list
of
individual
version
numbers
that
every
service
mesh
would
support.
Adapter
would
support.
E
C
We
want
to
get
to
that
first
and
then
start
to
iterate.
On
top
of
that,
where
people
can
then
selectively,
you
know,
choose
what
to
the
adapters
will
come
pre-bundled
with
support
for
as
many
as
many
as
we
can
stuff
in
there,
because
the
support
itself
is
more
about
is
extremely
small
in
terms
of
footprint,
the
the
actual
binaries
and
the
images
that
will
be
installed
to
run
the
mesh
those
are
still
stored,
separately
and
produced
by
that
service
mesh
project.
C
E
C
C
A
Thank
you.
Thank
you
aditi
thanks,
so
we
will
iterate
upon
spec.
A
A
B
Okay,
many
commands
have
been
added
to
mr
ctr
from
last
two
months.
One
of
them
is
messenger
check.
These
three
flags
are
added
to
it
and
two
are
still
pending.
So
if
anyone
wants
to
contribute
here,
most
welcome.
B
Okay,
let's
move
to
second
command.
That
is
added
to
this
video
that
is
perf
and
almost
everything
is
working
and
has
been
added
to
perf.
One
thing
that
is
not
working
correctly
is
the
file
thing.
It
is
being
fixed
by
multiple
people,
so
I
think
it
will
be
eventually
fixed
very
soon
in
the
mesh
command.
I
can
see
validate
valid
views,
finding
deploy
and
apply.
So
these
two
are
still
areas
to
be
worked
upon
in
patterns.
Everything
is
working
seamlessly.
F
In
app
suhani
added
all
the
features.
B
B
B
Most
of
them
have
been
changed.
If,
like
many
of
you
must
be
aware
of
it,
because
I
am
using
a
new
pr
each
week,
I
guess,
and
only
two
things
are
remained
a
filter
and
the
mesh
manager.
The
pr's
have
been
already
raised,
so
that's
it.
When
will
this
whole
column
will
be
green.
A
Thank
you
for
the
update.
You
sure
looks
good,
it's
nice,
that
we
are
almost
done
with
reworking
the
api
endpoints
as
well,
and
as
push
mentioned,
the
the
areas
that
are
not
green
in
this
spreadsheet
are
potential
areas
to
work
on
in
mastery
ctl,
so
reach
out
to
push
and
jump
on
jump
on.
A
If
you
could
share
the
link
to
spreadsheet
in
the
chat,
it
would
be
great
so
moving
on
to
the
next
topic
sameera,
would
you
like
to
talk
about
some
concepts
in
measuring
ui.
D
A
We
can
maybe
stay
back
after
the
call
and
discuss
that
or
catch
up
on
slack.
I
guess
yes,
but.
D
D
Yeah
something
like
this
so
now,
if
I
just
go
to
prometheus,
we
saw
it
was
connected.
But
when
I
go
here,
it
still
says
error
communicating
from
this
and
this
same
thing
there
are
few
things
with
filters
and
patterns
which
actually
I
wanted
to
discuss
so
just
wanted
to
know
from
where
I
get
all
this
like.
Is
it
from
my
series
server
or
is
it
from
my
cluster?
A
C
Yeah,
that's
it!
Those
are
those
last
so
those
last
two
questions
are
good
questions
of
like
measuring
is
it
highlights
the
fact
that
measuring
is
a
platform
that
is
extensible
that
you
can
bring
your
own
content
to
that
in
some
areas
we
would
do
well
to
seed
content
and
provide
sort
of
out
of
the
box
readily
available
filters,
a
couple
of
readily
available
patterns,
those
that
we
know
that
many
would
might
like
to
use.
D
Independence,
like
you
know
here,
if
I
see
go
in
patterns
now,
if
I
upload
it
really
doesn't
tell
me
which
kind
of
files
I
can
upload.
So
is
there
a
good
way
to
like
you
know,
give
indication
that
or
can
do
a
usa
validation?
If
I,
whatever
I'm
uploading,
is
a
proper
one
or
give
some
indication,
maybe
just
camera
or
json.
C
Or
yeah,
or
a
bit
of
yeah
yep,
that's
good
feedback.
I
mean
your
patterns
are
an
emerging
concept
in
measuring
and
the
there's
a
lack
of
context
for
what
the
heck
the
user
is
supposed
to
do
and,
like
you
know,
a
set
of
examples,
a
bit
of
like
a
bit
of
an
in-product
in
project
tour.
C
F
C
This
you
weren't,
being
you
weren't
being
rhetorical.
You
were
actually
asking
like
I
great,
I
see
when
it
was
uploaded
and
when
it
was
last
updated,
okay
and
then
some
other
name,
that
of
what
I'm
looking
at.
I
have
no
idea
what
those
are
it's
like
yeah
in
general.
The
concept
is
that
users
themselves
will
be
creating
these,
whether
they're
visually
designing
them
using
mesh
map
or
whether
they
are
grabbing
a
set
of
yaml
like
pre-existing
patterns,
maybe
as
starting
points
that
they
can
augment.
C
C
Where
two
things
are
not
happening,
one
is
that
if
you
have
prometheus
and
graffana
deployed,
meshsync
will
identify
those
and
the
prometheus.
The
metrics
drop
downs
in
the
settings
page
should
be
pre-populated
with
links
to
them.
They
should
actually
auto
connect.
If
those
are,
if
you
only
have
one
reference
to
prometheus
one
reference
to
grafana
meshri
should
auto-connect
auto-register.
C
Right,
yeah
yeah,
that
makes
sense,
that's
a
good
ask,
so
they
ask
is
like
hey.
Can
we
have
you
know
it's
nice
that
you
can
do
the
ad
hoc
ad
hoc
connectivity
test
good
measuring,
can
talk
to
prometheus
okay.
Now,
what
now
I'd
like
to
look
at
prometheus?
Is
there
a
link?
Can
you
provide
a
convenient
link
and
you
can
see
the
url
it's
like
so
yeah,
an
external
link
to
open
up
prometheus
sounds
about
right.
C
Right,
if
you
get
once
per
me
once
grafana
is
connected,
there
is
a
set
of
in
settings
under
metrics
you,
you
can
add
any
number
of
boards
any
number
of
panels,
but
it'll
it's
only
there
after
grafana
is
connected.
C
Well,
actually,
that's
not
true,
actually,
I
guess
yeah
if
you
take
definition
of
yeah.
Actually,
this
is
funny.
I've
not
seen
a
user
get
this
far
in
a
long
time.
So
this
is
great.
If
you
add
that
do
you
down
below.
Do
you
see
it's
not
yeah,
but
anyway,
you
would
see
charts
here.
C
Okay,
if,
after
you
get
grafana
connected,
I
think
you'll
have
a
bit
more
success.
D
It's
really
one
thing
is:
is
this
map
which
is
see
here?
Is
it
the
mess
map.
D
C
Nice
and
samir,
if
you
can,
you
do
try
to
connect
grafana
and
I
think
you'll
have
you'll,
have
more
success.
B
Also
to
add
a
few
things,
the
the
different
patterns
and
footage.
Obviously
you
are
seeing
it
because
now,
since
your
account
is
as
part
of
the
program
itself,
you
have
access
to
all
the
public
public
filters
patterns
and
performances,
that's
why
you
were
seeing
them
for
the
first
season
and
sorry
about
mesh
map.
Actually,
yesterday
we
were
trying
to
debug
the
rate
limiting
thing,
and
in
that
case
there
were
a
lot
of
unwanted
packages
which
we
released.
So
we
will
fix
that
soon.
Sorry,
for
that.
A
Thanks
amir
thanks
drew
so
we'll
be
covering
a
couple
of
more
topics
today,
so
please
stick
around
if
you'd
like
to
listen
to
it
and
see
some
demos
so
otherwise
thank
you
for
joining
and
we'll
see
you
tomorrow
in
the
smp
meeting
or
friday
on
the
community
meeting,
and
if
you
are
new
here,
you
can
jump
in
the
new
commerce
meeting
that
happens
tomorrow
as
well.
A
Moving
on
to
the
next
topic
nikhil,
would
you
like
to
talk
about
the
snack
bar
update
and
possibly
demo?
It.
D
So,
as
we
discussed
in
the
last
meeting
like
the
plan,
was
to
have
have
a
way
to
deploy,
machining
and
connect
to
nats
manually
if
the
user
wants
to
in
any
in
some
cases.
If,
if
the
deployment
or
the
connection
is
not.
C
D
Know
automatically
done
so
that
was
what
we
are
planning
to
do
and
to
take
a
step
towards
that.
We
have
implemented
this
thing
that
we
right
now
have
two
chips,
but
the
functionality
functionality
of
this
chips
is
that
if
it
is
already
connected
it
would
on
clicking
on
mesh
sync.
It
would
actually
try
to
redeploy
the
missing
deployment
and
in
case
of
nas
it
will
try
to
reconnect
to
the
latch
server
okay.
So
I
I
have
some
problem
with
my
cluster.
That
is
why
it's
not
working
just
before
this.
D
It
was
working
fine,
so
but
the
functionality
is
basically
working,
but
I'm
not
able
to
demo
it
right
now,
that's
the
issue
and
yeah,
so
this
machine
will
actually
redeploy
the
whole
machine
deployment.
If
you
press
on
it
and
then
an
app
will
actually
connect
try
to
connect
to
the
nats
broker
again,
if
it
is,
if
there's
a
problem
with
that,
basically
it
doesn't
actually
do
it
all
the
time
it
will
first
check
if
it
is
already
connected
or
not,
and
if
it
is
not,
then
then
it
will
try
to
reconnect.
B
Yeah
andrew
you
have
something
to
add
to
this
yeah
so
also
the
state
which
snaps
up
we
have
added
one
more
state,
basically
disabled,
active
and
connected,
because
those
three
are
different
and
usually
some
people
are
stuck
where
the
net
is
not
connected
and
they
want
to
connect
it
by
themselves.
So
they
can
see
over
here
that
it's
not
connected
and
they
can
connect
it.
B
D
F
D
Yes,
so
it
happened
twice
yesterday,
so
what
I
did
so
the
third
time
when
I
downloaded
this
thing,
so
I
saved
those
manifest
and
just
applied
with
cube.
Ctl
apply.
D
In
in
in
dot
messaging
folder
in
the
many
in
manifest
folder,
it
just
saw
only
two
files,
most
of
the
time
so
which
are
just
like
you
know,
just
v1
and
messages
and
operator
just
two
files.
I
can
see
sometimes
and
when
I
try
for
the
third
time,
then
I
got
all
the
files
like
service
deployment
default.
Everything.
D
D
D
C
Yes,
and
maybe
that
was
a
rate
limiting
issue
at
the
time
it
kind
of
goes
un
hidden,
sometimes
but
they're,
okay.
So
that's
part
of
potentially
part
of
the
issue.
The
other
part
of
the
issue
and
part
of
the
reason
that
we're
would
like
to
shift
to
only
using
helm
charts
is
that
is
that
is
sustaining
two
different
things.
So
the
kubernetes
manifests
the
direct
manifest
that
we
have
to
install
mastery
today,
articulate
and
characterize
a
mesh
redeployment
and
then
so
too
does
a
helmet
a
set
of
charts
that
we
have
well.
C
That's
just
two
separate
things
to
sustain
and
make
sure
that
they're
right
doing
it
in
just
one
spot,
so
the
the
manifests,
don't
don't
have
different
configuration
settings
for
mesh
redeployment
than
the
charts.
Do
that's
another
issue
that
could
have
affected
what
was
going
on.
D
Yeah
and
I
think
as
a
part
of
release
artifacts
right,
so
we
are
releasing
dot
like
you
know,
all
the
mercedes
ctl
executables
in
that
case,
can
we
also
export
the
mystery
camel
file?
So
so
it
does
go
and
download
it
from
there.
Maybe
just
just
a
thought.
C
C
You
really
just
need
to
grab
mastery
ctl
and
it
will
take
you
from
there,
although,
except
in
the
case
where
you're
in
air
gap
deployment
or
where
you're
dealing
with
github
rate
limiting
things,
maybe
you
should
be.
Maybe
you
know
it's
not
just
measuring
ctl
that
should
be
downloaded,
but
a
bundle
of
templates
and
manifests.
C
So
this
is
a
good
example
as
to
why
it
is
that
a
design
spec
needs
to
be
written,
that
one
off
myopic
consideration
to
how
functionality
is
implemented.
Does
us
a
disservice,
so
holistic
consideration
put
into
that
design
spec
with
a
bunch
of
comments
like
from
samir
as
well.
This
is
gonna.
Inspire
confidence,
gonna
be
good
yeah.
I
think
we
should
do
what
you're
suggesting
from
my
part.
A
With
this,
would
you
like
to
call
out
the
other
topic
as
well?
Yeah.
D
Yeah
sure
regarding
this
will
connect
and
try
to
get
this
thing
working
okay.
So,
regarding
the
next
thing
like,
we
actually
have
an
open
proposal
for
some
changes
for
state
management
and
measure
ui,
and
if
you
can
look
in
the
meeting
minutes,
you'll
have
the
link
for
the
proposal
you
can.
Actually
you
will.
D
You
can
feel
free
to
go
and
take
a
look
at
it
and
suggest
something
like
if
you
have
any
reviews
to
it
or
if
you
want
to
add
something
to
it
or
you
know,
whatever
the
opinions
that
you
have
you're
welcome
to
put
it
over
there,
we
can
have
a
discussion
as
well
as
to
so
basically
that
it
is
about
the
whole
state
management.
The
state
of
state
management
and
measuring
currently
is
not
it's
not
very
efficient.
D
There
are
some
problems
that
I
addressed
in
the
proposal,
so
it
would,
it
will
clearly
explain
things.
What
are
what
are
the
problems
that
you're
trying
to
solve
so
yeah?
That's
what
I
wanted
to
call
out.
A
So,
thank
you.
I
have
shared
the
link
to
the
spec
in
the
comments
moving
on
to
the
next
topic.
Gabra,
would
you
like
to.
A
B
It's
running
okay,
so
we
can
just
select
any
of
the
I'm
doing
it
for
the
application.
The
same
will
work
for
the
patterns
as
well,
and
we
will
click
on
this
and
this
thing
will
get
delete.
B
B
B
Just
yeah,
so
in
reference
to
that,
he
was
mentioning
that
we
were
not
able
to
upload
two
type
of
application
with
the
same
file
name
as
well
as
the
same
name.
So,
but
that
thing
is
there.
We
are
able
to
upload
the
two
applications
with
the
same
file
name.
So
I
have
just
added
that
thing,
so
that
is.
B
F
C
That
out,
okay,
thank
you,
ashish!
Please,
please
get
some
rest,
but
car
of
to
re
characterize.
The
challenge
is:
if
you
can,
we
are
being
smart
and
introspecting
the
name,
that's
included
inside
the
application
file,
and
so
what,
if
you
take
the
same
file
and
upload
it.
F
C
What
am
I
trying
to
say?
I
think
you
could
take
you
know
if
you
take
a
file
yammer
file,
you
have
a
name
like
this.
You,
you
leave
the
same
name
in
two
different
files.
F
C
Can
you
do
that
same
you?
Don't
don't,
please
don't
do
it
now,
but
can
you
do
that's
good?
I
think
that's
the
desired
behavior
that
you
know
that,
because
what
we
want
to
use
as
a
unique
identifier
for
a
given
entity
for
a
given.
C
Id
behind
the
system
so
or
within
the
system
can
you
do
the
same
thing
for
a
pattern
that
might
be
what
I'm
thinking
of
rather
than
applications.
I
don't
remember,
but
maybe
if
you
would
test
that
out
later
after
the
call
yeah
I'll
just
dip
it
up
on
your
top
of
the
board.
Oh
thank
you.
G
Yeah,
so
I
just
have
some
quick
updates.
Let
me
share
my
screen.
G
So
basically,
earlier
patent
support
was
was
available
only
in
the
sd
adapter,
but
I
have
recently
created
a
pr
in
osm
adapter,
which
basically
asked
support
for
patterns
the
patents
in
osm
adapter,
and
there
are
some
more
changes
in
those
personality,
but
they
are
not
in
specific
two
patterns.
The
second
thing
is
that
I've
also
added
support
for
the
same
in
linkery
right
now.
It's
it's
a
draft
fear.
The
reason
is
that
this
particular
pr
is
sort
of
relying
on
this
pr,
which
is
one
measure.
What
I'm
proposing
in
ministry
is
right.
G
Now
we
have
a
apply,
manifest
function
in
measuring
what
it
does
is
you
can
basically
it's
a
sort
of
like
applied,
so
you
can't
give
it
any
manifest
and
it
will
apply
all
of
them
one
by
one.
The
issue
that
we
were
encountering
was
that
illegally
installation
was
failing.
The
reason
was
that
one
of
the
crd
crds
crew
were
not
actually
getting
deployed.
The
reason
was
kubernetes.
Api
was
refusing
it
to
deploy.
G
The
reason
was
that
it
was
not
fitting
the
schema.
So
if
you,
if
you
would
actually,
instead
of
instead
of
applying
basically
instead
of
deploying
linkedin
from
mystery
adapter,
if
you
would
do
the
same
with
jupiter,
what
will
happen
is
that
it
will
deploy.
All
of
this.
All
of
the
resources
which
are
basically
perfectly
fine
and
leaf,
will
leave
basically
the
one
which
was
actually
erroneous.
So
that's
the
same
thing
that
I'm
proposing
to
add
in
this
particular
function.
G
That
is,
I'm
proposing
to
add
one
more
flag
so
that
the
default
behavior
doesn't
change,
but
if
a
user,
if
a
adapter
opts
in
so
basically,
the
proposal
is
to
add
this.
This
flag
ignore
errors.
If
this
is
true,
then
what
it
will
do
is
it
will
go
ahead
and
try
to
apply
all
of
the
manifest
if
any
manifest
apply
application
results
in
an
error.
It
will
skip
that
and
move
ahead
instead
of
throwing
error
right
away.
So
basically,
this
fixes
fixes
linkery
deployment
issue.
G
At
the
same
time,
it
is
also
more
like
cube
color,
so
maybe
it
is
more
of
so
it
is
more
of
it
is
more
intuitive
as
compared
to
what
we
had
earlier.
The
default
behavior
stays
a
while
we
I,
while
I'm
proposing
to
add
a
new
car.
This
particular
pr
is
actually
dropping
this
one.
That's
why
actually
it's
marked
draft
as
soon
as
this
one
gets
merged.
We
have
support
for
patents
in
malinkari
and
as
soon
as
this
one
gets
merged,
which
is
not
blocked,
it
can
be
merged.
G
After
review,
we
will
have
support
for
osm
in
patents,
which
includes
support
for
core
resources
for
western
resources,
as
well
as
installation
of
osm
service
management
cycle
injection.
Those
kind
of
things,
though
all
of
them,
are
actually
this
vr.
So
these
were
the
quick
updates.
G
The
third
thing,
the
another
thing
is
that
there's
call
for
volunteers,
so
we
have
already
added,
not
we,
I
should
say
reduction
since
then
have
already
added
support
for
basically
they
have
already
added
workload,
definitions
and
those
kind
of
things
which
you
can
actually
see
so
yeah.
They
have
added
workflow
definitions
in
all
in
kumar
doctor,
so
there
is.
This
is
called
for
voluntary
that
if
someone
wants
to
propagate
patent
support
in
kumar
after
they
can
just
go
ahead
and
do
so,
how
would
you
proceed?
G
Actually,
you
can
basically
refer
this
particular
pr.
What
this
pr
is
doing,
this
pr
is
achieving
exactly
the
same.
The
changes
that
you
would
have
to
do
in
kumar
adapter
would
be
very
similar
to
what
the
changes
have
been
in
western
doctors.
If
anyone
is
interested,
they
can't
ping
me
and
they
can
probably
create
an
issue,
or
maybe
I'll
basically
guide
you
how
you
can
actually
proceed
ahead
with
the
process,
so
this
was
this
was
update
and
called
for
what
I
do.
C
A
quick
question
is
the
the
thought
that
well
the
ignore
flag
or
the
force
flag-
maybe
you
know
sort
of
maybe
one
in
the
same
that
that
be
potentially
exposed
to
the
user
in
the
ui.
Should
they
want
to
take
their
life
into
their
own
hands.
G
Yeah,
no,
no,
actually
it's
more
about,
and
so
no
so
right
now
we
have
only
three
flags
in
so
no
it's.
It
won't
be
available
to
the
user,
it's
more
of
a
availability
to
the
developer.
So
right
now
we
have
these
three
things:
namespace
update
and
delete,
I'm
just
adding
one
more
ignore
errors,
because
I
know
that
are
these
installations
sometimes
errors
out
and
even
without
that
crd?
So
actually
one
thing
actually
goes
into
crash
crash
loop
crash
back
crash
back
off
loop,
but
other
than
that
actual
inspiration
succeeds.
G
The
same
result
is
also
produced
if
you
would
actually
go
with
cubicle
installation.
So
the
thing
is
that
so
it's
it's
more
of.
We
can't
say
that
hey.
We
proceeded
with
the
deployment
now
it's
a
lingerie's
issue
because
they
have
something
messed
up
in
this
theory.
C
C
If
you
try
to
go,
install
open
service
mesh,
you
can't
not
through
messaging,
I
just
mean
straight
up
through
osm,
which
is
I
don't
know
if
it's
comforting
that
other
you
know,
large
projects
have
egregious
bugs
as
well,
but
but
they
do
and
no
but
the
the
hair,
the
single
hair
that
continues
to
stand
up
on
the
back
of
my
neck,
about
the
ignore
errors
is
isn't
that
there
isn't
more
errors,
like
that's
a
that
sounds
like
an
entirely
appropriate
behavior
to
enhance
and
add
so
that
I
think
this
is
net
positive.
G
Yeah,
actually,
we
can't
expose
it,
but
the
reason
I
haven't
exposed
it.
So
definitely
we
can't
do
it
via
patents,
but
that's
the
caveat
that
we
can
expose
it
only
via
patents.
So
definitely
when
we
have
new
management
ui,
you
would
have
access
to
that
thing
in
the
e1,
but
with
the
current
ui
it
won't
be
very
explicit
to
the
user
and
listen
to
their
writing
pattern
for
different
linguistic
service
mesh
or
any
service
finish
nice
cool.
C
Now,
for
those
that,
by
the
way,
just
a
quick
note
that,
like
getting
each
of
these
adapters
supporting
pattern-based
operations
like
they
call
for
help
on
kuma,
that's
that's
wonderful.
If
you
are
raising
your
hand
there,
please
note
that
you
will
be
hounded
to
execute
and
if
you're
raising
your
hand,
don't
this
isn't
the
time
for
you
to
be
taught
and
trained.
C
C
A
We'll
try
it
thank
you.
Everyone
for
joining
we'll,
see
you
next
time.