►
From YouTube: Meshery Development Meeting (June 23rd, 2021)
Description
Meshery Development Meeting - June 23rd, 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/
B
B
It
well
there's
the
standing
tradition
we
have
to
you
have
to
embarrass
yourself.
You
have
to
say,
hi,
tell
us
your
middle
name
and
then
say
you
don't
have
one
and
then
fine
say
that
tell
us
tell
us
about
yeah
introduce
if
you
would
what's
your
favorite
discussion.
C
Yeah,
actually,
I,
like
I
joined
this
community
like
this
week
only
and
like
I
I
I
wanted
to
like
contribute
to
the
some
like
service
mesh
related
projects
and
currently
I'm
working
with
vmware
or
tanza
service
mess
as
a
sre.
C
So
I
just
wanted
to
like
understand.
I
mean
about
this
domain
than
about
service
mesh,
and
so
I
thought
of
like
on
going
and
joining
this
community
will
help
me
so
looking
forward
working
with
you
guys,
yeah.
B
It's
hard
to
have
a
better
introduction
than
that.
That's
great!
Now,
that's
perfect!
It's
well
welcome
good
good.
B
Yeah
yeah
by
the
way,
neja
whereabouts
whereabouts.
Are
you
calling
from
what
city.
B
Okay,
been
there
been
there
yeah,
no
okay,
I'm
kidding,
I
haven't
been
there,
but
good
good
cool,
good,
okay,
you're,
already
being
put
putting
up
with
the
harassing
and
the
bad
jokes.
So
that's
that's
good!
We're
off
to
a
good
start
well
by
demi
by
demi.
Have
you
been
on
a
call
before?
Have
we
gotten
to
say
hi.
B
Oh
yes,
it's
your
first
time,
oh
beta
me,
okay
or
b
to
me,
be
to
me
or
be
the
may,
we'll
get
it
eventually
or
one
of
us
will
get
it
eventually
very
nice
to
have
you
thanks
for
saying
hi
in
the
chat.
If,
if
your
audio
is
on
feel
free
to
say,
hi
it'd
be
nice
to
hear
you
if
not
no
problem.
D
D
I
am
from
india
currently
pursuing
bsc
computer
science.
D
Interested
in
contributing
in
open
source
projects-
and
I
check
out
this
community
bridge
website
that
this
is
the
community
day
of
my
so
I
joined
it.
This
is
my
first
meeting.
B
B
Yeah,
are
you
what
year
are
you
in
university.
D
B
Okay,
you
might
have
said
that
I
might
have
missed
it
very
good
nice.
This
is
oh.
B
Okay,
oh
nice
kill
nice
to
have
you
thanks
for
coming
anybody
else.
First
time
on
the
call
anybody
else
want
to
say
hi.
B
B
See
the
thing
about
this
tradition
is
so
long
as
we
do
it
every
time
it
only
takes
so
long
because
we
we
chew
away
at
that
technical
debt
every
time
that
we
have
the
meeting.
So
this
is
great
all
right,
very
good
if
anyone
hasn't
had
it
so
jean
tiston
who's
she's,
your
old
news.
B
At
this
point
I
mean
gene-
your
name
jumps
out
at
me
because,
because
I'm
really
excited
about
your
pr
and
actually
gene,
it
might
make
sense
if,
since
you're
on
for
you
to
maybe
you
know,
put
your
name
down
and
show
some
of
the
work
that
you've
done
because
getting
getting
you
some
real-time
feedback
will
be
helpful
to
you
always
nice
to
show
off
your
work
too.
So
so
great,
okay!
Well,
thank
you
all
for
coming.
So.
B
So
first
topic
today
was
that
the
fact
that
mesheri
as
a
project
and
service
mesh
performance
smp
as
a
project,
if
you
haven't
so
all
of
you,
have
no
doubt
visited
the
mesheri.io
project,
page,
probably
like
you're,
all
in
various
ways
to
various
degrees.
Familiar
with
measuring
and
that's
it's
our
certainly
our
our
project
of
focus
in
this
call
in
the
layer,
5
community.
We
have
a
number
of
projects,
one
of
the
other
ones
that
we
spent
quite
a
bit
of
time
on
is
smp
service
mesh
performance.
B
If
you're
like
me,
you
think
specifications
are
horrifically
boring
and
so
fine,
that's
fair.
There's
another
specification,
that's
related
to
smp.
It's
called
smi
service,
mesh
interface,
smp
and
smi
are
interrelated
and
complementary
how
they
align,
how
they
underlap,
how
they
overlap
and
whether
or
not
the
two
projects
should
merge.
B
That
has
been
a
topic
of
discussion
over
the
last
couple
of
months
measuring
implements
smi
at
mescheri
and
the
venue,
if
you
cruise,
to
the
blog
actually
to
the
last
blog
post
on
this
site,
there's
a
blog
post
that
talks
about
mesherie
being
the
conformance
tool
for
smi.
So
so,
how
does
a
service
mesh
or
how
does
a
person
know
if
a
given
service
mesh
has
implemented
smi
service
mesh
interface?
B
How
does
everyone
know
if
those
specifications
have
been
implemented
to
to
spec
or
to
t
use
measurey
it'll
tell
you
ask
dhruv
he'll
tell
you
great.
These
projects
are
all
reinforcing
of
one
another.
They
are
all
also
so
they're
reinforcing
of
another
one,
another
they're
aligned.
They
are
they
slightly
overlap
and
slightly
underlap
and
to
what
degree
they
do
either
of
the
two
of
those
things
has
been
a
topic
of
discussion
in
preparation
for
measuring
and
service
mesh
performance,
the
two
projects
entering
into
the
cncf.
B
So,
as
of
yesterday,
it's
been
a
long
time
coming.
This
was
always
the
plan
for
the
projects
from
the
start.
As
of
yesterday,
they
are
both
cncf
projects,
so
that
is,
this
is
more
or
less
probably
the
first
announcement
of
that.
We'll
make
more
announcements
of
the
fact
that
that
is
true
and
we
will
have
more
to
discuss
about
whoa,
like
you
know,
celebrating
that,
but
also
what
does
that
mean?
What
does
that
mean
to
the
project?
B
And
what
does
that
mean
to
the
community
we'll
discuss
a
little
bit
of
that
today,
and
probably
even
more
of
that
on
friday
at
friday's
community
call.
So
for
most
of
you,
I
suspect
many
of
you
know
that
actually,
in
this
community
we
meet
every
day,
sometimes
multiple
times
a
day,
because
there's
a
lot
going
on,
because
there's
a
lot
of
you
that
have
an
appetite
to
dig
into
various
technologies
that
are
of
consideration.
B
Mescheri's
roadmap
is
well.
It's
it's
not
world
domination.
That's
for
sure
that
we
leave
that
to
aws
that's
their
gig,
but
it
is
large
and
we're
only
about
not.
You
know
like
halfway
there
ish.
There
are
a
couple
of
major
components
that
have
yet
to
be
added,
and
I'm
excited
for
those
us
collectively
to
start
on.
Some
of
those
components,
so
one
of
those
is
a
workflow
engine.
B
I
just
know
that
you
are
there
are,
which
also
is
an
appropriate
time
to
call
out
a
couple
of
things
one
some
of
you
have
been
here
for
a
long
time.
Some
of
you
have
given
well,
I
know
you've
probably
hurt
your
fingertips
on
the
keyboard
or
on
the
mouse
for
how
much
work
you've
been
trying
to
do
so,
a
little
bit
of
blood
and
sweat.
Some
of
you
become
have
taken
on
different
roles
on
the
project
and
in
the
community
and
to
it's
all
to
your
grant.
B
You
know
like
the
to
your
benefit,
so
I
hope
that
measuri
and
the
projects
continue
to
grow
so
that
so
yesterday's
or
today
yesterday
is
a
significant
step
for
the
projects
as
they
enter
into
the
cncf.
So
great.
So
some
of
you
are
already
maintainers
of
a
cncf
project
or
a
couple
of
projects.
B
So
awesome
most
of
you
are
already
contributors
to
the
project
and
actually
I'll
call
this
out,
because
this,
like
really
pisses
me
off
actually-
and
that
is
that
I
think
in
general-
open
source
communities
have
gotten
better
about
this
particular
concept,
but
the
concept
that
what
it
means
to
contribute
doesn't
necessarily
mean
code
at
all.
In
fact,
it
doesn't
mean
code
at
all.
B
It
often
is
what
people
are
referring
to,
but
just
by
by
showing
up
asking
questions,
that's
contributing
as
a
matter
of
fact,
there's
a
couple
of
folks
that
have
shown
up
for
about
a
year
on
the
calls
and
don't
say
anything,
so
they
don't
even
ask
questions.
B
They're
still
contributing
really
nice
to
see
you
all
here,
like
it's
helpful,
it's
no
matter
what
you're
doing
it's
it's
helpful
and
so
like
so
just
to
pick
on
neha,
because
I
want
her
to
feel
very
welcome,
like
the
rest
of
you
is
if,
as
neha
jumps
into
like
there's
a
whole
set
of
meetings
dedicated
to
devops
centric
things
to
ci,
to
cd,
to
workflows,
to
automation,
to
deployments
and
how
mesher
works
in
different
systems
and
all
this,
and
not
always,
you
know
most
of
the
time
the
activities
that
go
on
in
there.
B
There
aren't
necessarily
code
per
se,
but
they're,
extraordinarily
important
they're,
just
as
important
as
the
code
as
a
matter
of
fact,
in
some
cases
more
important
than
the
code.
So
so
that's
not
this
community.
If
you're,
not
writing
code
that
doesn't
matter
that
doesn't
mean
you're
not
contributing
if
you're
not
a
maintainer,
it
doesn't
matter
that
doesn't
mean
you're,
not
contributing.
B
So
there
are
many
other
communities
to
go
to
where,
if
you
disagree
with
that,
like
you
can
almost
stumble
into
any
other
community
and
they'll
agree
with
you
about
like
what
it
means
to
contribute
and
not
having
to
be
code.
That's
not
here
so
jean
got
hammered
on
and
loved
on.
I
guess
in
the
best
of
ways
when
she
joined,
because
she
was
gonna
focus
on
some
front,
front-end
front-end
things
and
oh
what
a
blessing.
B
That
is,
because,
oh
how
much
we
need
of
like
people
helping
the
rest
of
us
interface
with
these
with
unexposed
back-end
things
so
good.
Okay,
I'm
totally
digressing
into
what's,
hopefully
a
a
word
of
encouragement,
a
word
of
celebration
about
the
two
projects
entering
into
the
cncf.
B
It
does
mean
change,
but
not
a
lot
like
like
supportive
change.
A
lot
of
the
things
a
lot
of
the
changes
will
will
go
unnoticed
by
the
vast
majority
of
you,
I'm
we're
happy
for
my
part
and
for
the
maintainers
and
others
that
are
representatives
here
like
happy
to
itemize
every
single
one
of
what
one
of
those
changes
are,
but
most
of
them
are
don't
mean
it.
B
B
One
of
one
of
the
things
that
is
meant
and
we
sort
of
started
to
work
on
last
night-
is
that
the
measuring
repository
and
by
the
way
I
need
to
shut
up
and
let
you
guys
continue
so
the
measuring
repository
as
one
example,
you
know,
there's
like
30,
something
in
the
layer,
5
io,
github,
org,
there's
like
40
something
repositories
repos,
I
don't
know,
there's
like
it's
about
30
of
them
ish
or
something
25
or
something
that
are
for
mescheri
they're,
like
of
the
mystery
project.
B
It's
either
adapters
or
mastery
itself
or
various
related
supporting
components.
And
so
you
know
a
lot
of
them
will
migrate
from
layer,
five
io
org
to
measure
org.
In
the
same
way,
service
mesh
performance
will
migrate
to
service
mesh
performance
org
and
that's
just
a
matter
of
helping
ensure
that
the
projects
themselves
are
sort
of
their
own
entity
unto
their
own
right.
B
Which
unto
his
own
is
a
concept
that
needs
to
be
explained,
and
I
won't
do
that
now,
but
I
will
also
say
that
there's
a
service
mesh
patterns,
github
org,
so
the
meshery
is
an
awesome
thing.
Service,
mesh
performance,
an
awesome
thing,
but
service
mesh
patterns
are
potentially
like
a
standalone
thing,
as
well
or
kind
of
like
some
of
the
web
assembly
filters.
The
wassum
filters
that
some
of
you
are
working
on.
They
too
are
kind
of
a
standalone
thing,
so
I'm
just
foreshadowing
some
potential
things
to
come
kind
of.
B
A
Thanks
lee
sounds
awesome,
looking
forward
to
hear
more
about
it
on
the
community
call.
So
with
that
announcement.
Let's
move
on
to
the
topics
for
today,
so
first
we
have
pierce
you.
Should
you
like
to
talk
about
the
system
check
command
and
what
it
represents.
F
F
F
F
F
But
we
want
to
make
it
work
like,
depending
on
the
platform
the
person
wants
to
run
machine
like
if
it
is
he's
running
it
on
docker,
then
we
should
only
run
docker
checks.
F
So
this
is
the
command
that
we
are
using
as
a
page
execution
execution,
and
so
we
are
running
docker
health
checks
when
it
is
not
a
premium
execution.
That
means
we
are
running
system
check.
That
means
in
every
time
we
run
system
check.
Docker
will
be
always
test
it
out
if
it
is
running
or
not
or
if
the
platform
is
docked-
and
this
is
a
premium
execution
that
is
user,
wants
to
run
it
on
much
more
and.
F
F
A
A
A
So
one
of
the
things
pious
was
mentioning,
was
system
check
and
it
also
has
something
called
a
preflight
check,
and
this
preflight
check
has
already
been
implemented
in
measures
right
now.
So
what
this
pre-flight
check
does?
Is
it
basically
checks
if
the
environment
is
ready
to
be
ready
so
that
we
can
deploy
measuring
so
it
checks?
A
If
the
user
has
docker
docker
compose
and
if
they
are
using
a
kubernetes
platform,
then
it
will
check
if,
if
you
can
initialize
a
kubernetes,
client
and
other
stuff,
so
the
the
preflight
check
is
already
been
implemented
and
the
system
check
is
is
a
superset
of
the
preflight
check
and
it
has
a
lot
more
checks
like
it
also
checks
for
yeah
as
push
is
showing
its
check
for
the
measuring
version
and
checks.
A
B
Yeah,
I
so.
B
Good,
so
what
you
should
explain-
and
the
vendor
explained-
is
relatively
straightforward
in
concept
and
it's
a
very
helpful
feature
to
have
both
before
someone
uses
mesherie
deploys
mesh
and
as
they
go
to
interact
with
an
already
deployed
measuring
environment,
that
these
checks
are
things
that
people
can
run
ad
hoc,
but
there
are
also
checks
that
run
implicitly
every
single
time.
Someone
invokes
a
command
like
these
guys
said
so
then,
but
okay,
so,
but
not
all.
B
B
So
I
think,
there's
a
call
to
action
is
making
one
he's
educating
all
everyone
on
the
work,
that's
been
done
and
then
two
a
call
to
action
for
people
to
well
one
go.
Try
the
utility
go,
try
the
commands,
but
to
the
fuchsia
is
linked
to
a
design
spec
that
I
encourage
everyone
to
review
and
think
about
what
of
these
check,
which
of
these
checks
are
universally
mandatory.
B
Some
of
these
checks.
They
should
always
pass.
Otherwise.
You
have
an
issue,
and
you
should
you
gotta.
You
have
to
address
that
issue.
Some
of
them
are
only
appropriate
in
certain
circumstances,
and
so
what
are
those
rules
like
that?
That's
what
we
need
to
reflect
on
and
for
my
part,
I
don't
have
something
more
intelligible,
something
more
specific
to
to
add.
But
but
but
all
of
you
do
all
of
you
who
are
on
the
call.
B
That
you
can
ask
yourself
this
question
when,
when
your
deployment
doesn't
work
or
you're
having
trouble
you're
trying
to
troubleshoot
it
like
here,
here's
here's
one
of
the
core
utilities
to
figuring
that
out,
which
means
so
I'll
end
on
this,
which
means
that,
like
hey,
there's
a
system
check
great
there's
a
system
check,
hyphen
hyphen
preflight.
I
can
pre-flight
check
great
a.
B
Maybe
there
should
be,
and
there
is
sort
of
an
early
design
spec
for
gosh
peach.
What
do
we
call
it?
Is
it
a
report
like
a
troubleshooting,
a
diagnostics
report
so
that
you
can
use
like
when
you're
facing
an
issue
and
you
need
to
open
a
bug.
B
B
Perfect,
just
because
your
voice
is
a
little
bit
hard
to
hear
repeat
that
last
bit
and
that
is
so
for
any
user
who's
having
an
issue
with
mesherie
that
they
can
go
run.
This
system
check,
hyphen
hyphen
report
have
a
diagnostic
bundle
packaged
up
and
make
it
very
convenient
for
them
to
send
in
reports
that
have
all
of
the
necessary
debug
or
a
contributor
to
look
at
it
and
work
on
it.
B
A
Thank
you
pierce.
Thank
you.
The
design
spec
is
on
the
chat.
The
link
to
the
design
spec
is
on
the
chat.
The
next
item
is
also
from
piyush.
It
is
about
the
perf
command.
The
buff
command,
basically,
is
as
lee
lee
mentioned
earlier,
like
we
use
service
mesh
performance
to
run
performance
tests
on
the
service
mesh,
so
system
perf
is
a
measuring
suitable
implementation
of
the
implementation
of
running
performance
test.
So
push
has
been
working
on
it
sometime
and
so
do
you
want
to
demo
it
and.
F
F
So
what
we
are
doing
right
now
is
we
are
running
these
tests
with
the
help
of
links.
So,
let's
see
how
this
works.
So
there
are
two
cases
if
someone
wants
to
create
a
new
profile
with
running
the
test
and
if
someone
wants
to
run
the
existing
profile
or
use
that
existing
profile
with
a
new
url
that
he
is
like
passing
with
the
flags.
These
are
the
three
cases.
F
So,
let's
see
what
like,
how
I'm
not
trying
to
transform
things.
So
if
there
are
no
arguments,
that
means
he
is
not
passing
any
profile,
so
we
create
a
new
profile.
F
We
create
a
new
profile
with
like
sending
a
request,
adding
all
these
parameters,
and
then
we
save
this
profile.
On
the
back
end,
we
get
the
profile
id
and
the
profile
id,
and
we
run
the
test
with
this
profile
id
and
the
url
that
has
been
passed
with
the
command
and
the
second
thing
that,
if
someone
is
using
the
existing
profile,
so
we
get
the
profile.
We
fetch
the
profile
first,
because
we
want
to
know
the
profile
id
of
this
profile.
F
Okay,
so
what
will
happen
is
we
do
not
have
unique
names
for
performance
profiles?
I
think
the
files
can
have
the
same
name.
That
is,
let's
say,
let's
say,
test
okay,
so
if
there
is
only
one
profile
which
that
name,
let's
say
there
are
no
profiles
found
with
the
name.
We
will
return,
no
profile
found
with
the
given
name
and
if
we
have
only
one
profile,
that
means
that
we
can
execute
the
test
with
it.
So
we
get
the
profile
id
and
we
execute
the
test.
F
F
F
F
F
F
F
F
H
Push
actually
I
I
do,
how
are
these
profiles
created?
Do,
do
you
have
to?
Is
it
done
manually?
Is
it
online?
F
F
B
It
turns
out
it's
a
it's
a
strategically
placed
question
because,
and
it's
a
good
thing
for
us
to
confirm
push
on
the
is
the
dash
f
flag,
dash
dash
file
flag,
that's
supported,
flag,
correct.
F
B
If
you
look
at
you,
look
at
the
description
of
that
flag,
so
what's
happening
here
is
like
it
from
the
command
line.
If
you
don't
want
to
specify
you
know
five
six
different
arguments,
you
can
very
simply
say
well,
hey.
I
have
an
existing
definition,
existing
performance
profile
that
I've
already
you
know
mouse
clicked
into
the
system.
I'd
like
to
just
up,
you
know,
run
that
profile
run
that
test
the
test
defined
by
that
profile.
B
B
What
load
generator
do
you
want
to
use
all
the
same
things
and
as
it
turns
out
like
that
file
is
like
highly
related
to
the
service
mesh
performance
specification,
the
smp
spec,
the
spec,
calls
out
a
bit
more,
but
that
same
description
that
same
set
of
information
is
smp
compatible,
and
so
the
last
thing
piyush,
if
you
go
back
to
mesherie's
ui
briefly
inside
of-
and
you
take
a
look
at
the
details
of
the
results
of
a
given
test
and
when
you
view
the
test
results
on.
B
If
you
end
up
check
marking
the
box,
there
there's
a
new
yeah,
this
download
yeah,
if
you
click
that
what
you'll
get
is
well
you'll,
get
an
smp
compatible
file
you'll
get
that
same
file.
So
if
you
wanted
to
so
so,
then
yeah
we've
kind
of
come
full
circle
like
you
can
take
this
file
and
supply
it
as
input
to
define
a
new.
You
know
a
performance
test,
some
of
what
I'm
saying
right
now
we
have
to.
I
have
to
like
that's
the
intended
design.
B
It's
been
a
while,
since
anyone
has
really
focused
on
that
full
like
like
someone
needs
to
go
test,
that
what
I
just
said
is
remains
true,
that
you
can
technically
take
this
and
go
send
in,
because
mostly
what
this
is
showing
is
results.
But
it's
also
intended
to
fully
describe
the
profile
itself
so
that
you
can
reapply.
H
Actually,
if
you
don't
mind,
you
know,
I
have
a
quick
question
and
I
know
it's
easy
to
ask
questions
and
you
know
give
advice.
You
know
that
things
that
I
cannot
do,
but
I
was
wondering,
is
it
possible
that
you
could
have
applied
the
same
paradigm
to
the
previous
demo
instead
of
having
in
your
code
a
whole
bunch
of?
H
If
then
else,
if
this,
if
that
is
it
possible
to
have
a
test
engine,
you
know
and
and
and
you
add
a
test
to
this
test
engine
so
that
you
know
each
time
you
have
you
think
of
a
new
test.
Somebody
says:
hey,
you
know
what
is
this?
You
don't
have
to
code
and
recompile
the
thing
again.
B
Yeah,
the
vj,
that's
what
the
get
nighthawk
project
is
kind
of
about
like
yes,
if
you
tease
out
what
you
just
said,
you
well
here
here
I'll
present
this
to
to
all
of
you
like,
if
you're,
if
you're
running,
you're
running
your
infrastructure,
you've
got
some
workloads
that
are
running.
B
You
want
to
make
sure
that
the
changes,
the
compute,
the
way
that
you're
configuring
your
infrastructure,
the
way
that
you're
configuring
kubernetes
you're
configuring
your
service
mesh,
is
in
alignment
with
the
your
performance
requirements
or
you
want
to
understand
the
effect
you
know
you're
using
this
performance
management
right
you're
doing
these
things
well
part
of
the
question
that
you
begin
to
ask
yourself
is
like
am
I
doing
it
well
like,
like
I'd
turn
over
to
look
at
samir
and
I'd,
say
samir?
B
How
fast
is
yours
running
because
I'd
like
to
compare
that
to
mine
and
and
to
to
see,
if
I'm
you
know
slow
poking
along
I've
got
a
bad
config
like?
How
do
I
tune
this
thing?
To
be
just
perfect?
B
Okay,
well,
that
sort
of
leans
into
what
vj
was
just
asking,
which
is
so
I
can
go
run
a
test
and
it'll.
Tell
me
under
these
constraints,
with
this
configuration.
This
is
the
type
of
it'll
characterize
the
type
of
performance
that
I
would
expect,
or
that
I'm
not
that
I
expect,
but
that,
like
that,
I'm
getting
okay!
Well,
what
if
I
tweaked
this
configuration
over
here
like
what,
if
I
reduce
the
number
of
times
that
my
distributed
traces
are
being
sampled,
how
material
of
a
performance
effect
will
that
have?
B
Well,
I
don't
know:
okay,
we
have
okay,
let's
go
change
it
and
let's
run
a
test:
okay!
Well,
what?
If
I
lowered
it
a
little
bit
more?
Oh
geez,
okay!
So
like
okay,
so
like
there's
this
optimizat
this
question
of
optimization,
what's
the
optimal
number
of
time
of
like
what's
the
optimal
config
with
what
do
you
call
it
with
a
bunch
of
permutations
to
it?
B
Well,
that's
the
path
that
we're
on
is
is
to
having
this
system
do
that,
for
you
eventually
for
you
to
say,
here's
my
constraints,
constraint,
number
one.
This
service
called
the
payments
service
where
we
accept
credit
cards
and
we
process
credit
cards
and
we
need
to
make
sure
all
those
transactions
happen
like
that
or
whatever
the
service
is
right,
like
here's
my
service
and
it
needs
to
be
responded
under
500
milliseconds
or
it
needs
to
have
some
certain
performance
characteristic
to
it.
B
So
here's
my
constraint,
my
constraint,
number
one,
this
many
milliseconds
or
less
constraint,
number
two.
I
want
it
to
be
as
resilient
as
possible,
like
I
don't
want
to
the
user
to
try
to
process
the
credit
card
payment
and
it
to
fail
and
return
back
to
them
with
a
failure
they
might
just
get
upset
and
like
take
their
credit
card
somewhere
else
like
heaven
forbid
that,
so
I
want
the
system
to
be
resilient.
I
want
it
to
retry
if
the
connection
fails.
B
B
Don't
ask
me,
I
don't
know
like
can't
can't
the
system
figure
that
out
for
me.
Yes,
that's
what
we're
going
to
be
working,
that's
what
get
nighthawk
is
about
adaptive
load,
leverage,
nighthawks
adaptive
load
controllers
to
have
optimization
routines.
Programmatically
recursively
run
performance
tests
that
arrive
at
optimal
inflection
points
based
on
constraints
you
define
within
your
system.
G
A
Yeah,
thank
you
lee.
Anybody
has
any
other
comments
before
we
move
on
to
the
next
item.
A
Okay,
so
thank
you
pews.
Thank
you,
lee.
Let's
move
on
to
the
next
item.
Samir.
Do
you
want
to
yes,
so
samir
is
going
to
talk
about
configuration
of
virtual
services
for
missouri
in
istio
so
summer,
audios.
I
I
I
I
So
let
me
know
if
my
like,
you
know
this
figure
is
visible,
otherwise
I'll
just
make
it
bit
bigger
okay.
So
this
is
basically
the
infrastructure
which
we
are
following
to
be
a
very
like
you
know
resilient
or,
you
can
say,
secure,
secure
services.
I
I
Then
we
have
a
private
link
which
is
enabled
for
istio
gateway.
So
in
this,
what
exactly
happens
depending
upon
your
request,
so
this
request
will
be
forwarded
to
the
this
particular
like
you
know,
you
can
see
a
private
link,
ip,
which
is
heating
here
so
now
request
from
the
user
goes
through
this
many
thing.
So
now,
what
is
the
error
which
is
happening
currently?
What
I
like
you
know,
currently
I
see
is
so
when
I
go
to
the
message
and
click
on
the
provider.
I
B
B
And
just
in
case,
it's
not
obvious
for
everyone
else.
In
this
environment,
smear
is
running.
Measury
on
istio
right
measuring
is
on
the
surface
mesh
as
well.
Yes,.
I
Okay,
okay,
so
now,
if
you
see,
I
have
a
gateway
and
I
have
a
message
which
is
running
so
let
me
just
go
up
to
my
browser
okay,
so
this
is
my
dns
or
my
application
gateway
url,
wherein
I
am
just
forwarding
it
with
slash.
Okay.
Now,
when
I
go
here.
F
I
think
you
can
save
the
file
when
you
change
it
to
slash
yeah
save
it.
I
didn't
save
the.
I
Okay,
so
you
can
see
the
error
so
instead
of
it
goes
to
my
app
gateway.
It
is
redirecting
me
to
the
like,
you
know,
messaging
or
my
what
I
would
say
my
ist
or
get
it
so
this
is
not
the
desirable.
That
is
why
it
is
saying
the
site
can
be
reached,
because
this
is
in
the
private
id.
So
you
cannot
reach
this,
so
this
is
one
thing
which
is
kind
of
not
supporting
my
architecture
yeah,
so
it
is
like
when
I,
when
it
is
redirecting.
I
B
One
this
one
brings
a
smile
to
my
face
samir.
This
is
I'm
sorry
to
say
that
unless
budkash
or
abhishek
or
someone
else
on,
the
call
corrects
me
need
to
change
a
little
bit
of
code
here
to
to
to
help
to
enable
this
use
case
like
what
a
great
use
case.
B
As
a
matter
of
fact,
we've
seen
this
one
before,
but
we
didn't,
but
in
your
case
we're
committing
to
making
you
successful
like
we
really
want
you
to
be
successful,
and
so
we
postponed,
like
you,
really
really
updating
the
system
to
facilitate
this
kind
of
a
thing
like
it
is.
I
And
instead
of
the
application
gateway,
there
are
some
scenario
which
people
are
using
like
front
door
here
as
your
front
door
or
some
of
the
similar
kind
of
gateway
from
aws.
Also,
so
what
my
thing
is
we
may
need
to
enable
some
sort
of,
like
you
know,
verification
code
based
mechanism,
so
where,
in
like
you
know,
we
can
just
talk
with
the
api
of
any
particular
services
or
any
particular
identity
provider.
Here,
yeah.
B
Yeah
totally
yeah.
This
is
that's
a
great
suggestion
that
I
near
as
I
can
tell
that
suggestion
would
circumvent
the
system
like
our
circumvent
the
need
for
like
from
these
the
the
social,
the
social
authentication
that's
happening
here.
The
oauth
that's
happening
is
that
once
a
user's
identity
has
been
confirmed
by
the
identity
provider,
twitter
github,
google
linkedin
they
are
in
a
hard-coded
way.
B
The
the
mescheries
remote
provider,
the
thing
the
page
that
you're
on
now
it
will
receive
the
jw
well
it'll,
receive
an
access
token
back
and
no
no
hold
on.
We
will
generate
an
access
token.
B
There
may
be
a
point
of
manipulation
right
here
where
we
might
be
able
to
address
this.
This
is
not
something
that
I
don't
think
it's
something
we'll
be
able
to
find.
The
answer
for
on
the
call.
The
thing
that
you
suggested,
though,
does
like
doesn't
rely
on
callbacks
from
the
way
that
oauth
2
works.
It
there's
callbacks
and
things
that
happen
in
rerouting
and
redirecting
that
the
which
I
think
is
can
be
addressed,
but
also
the
approach
that
you're
that
you've
suggested.
B
B
I
B
Sure
yeah,
actually
that's
a
great
question
right
there
is
there
any
so
there's
no
active
director
yeah,
there's,
there's
no
directory
services
at
all
that
you
guys
are
using
whether
from.
I
We
can
go
for
a
zero
authentication,
but
because
that
is
our
kind
of
authentication
system.
I
see
okay.
B
B
Good
good,
I
would
crush
any.
I
Yes
and
the
next
thing
which
I
wanted,
that
is
very
straightforward
like
like
you
know,
we
I'm
not
able
to
run
it
behind
this
reverse
proxy,
because
there
is
multiple
redirection
happening
so
which
is
causing
like
it
in
like
in
the
same
way
how
grafana
happened
in
grafana
it
happens,
so
maybe
we
have
to
have
some
kind
of
a
adjustment
or
configuration.
So
this
is
in
this
way,
maybe
domain
or
root
url
or
something.
Let
me
just
show
that
quickly.
I
I
I
H
B
Multiple
redirections
yeah,
but
well
I
also
wonder
right.
I
mean
if
so
in
this
case
the
user
doesn't
have
an
active
token
and
so
mesherie
like
while
the
first
request
it
looks
like
it's
getting
rerouted
to
me
to
mastery
properly.
Mastery
will
then
probably
redirect
from
root
from
slash
to
slash
login.
B
I
I
I
You
were
telling
about
something.
Can
I
try
this
out.
B
B
I
B
So
I
need
to
I
need
to
look.
I
need
to
refresh
my
member
there's,
so
I
guess
they're
probably
a
couple
of
things.
I
mean
one
one
vendor
just
for
in
terms
of
the
the
meeting
itself.
One
thing
to
that.
B
I
might
suggest
to
everyone
who
wasn't
able
to
show
their
work
today
and
is
looking
for
some
feedback
to
do
a
quick
recording
of
your
work
and
just
and
share
it
in
the
slack
or,
if
you
like,
share
it
on
the
issue
that
you're
on
or
share
it
in
discuss
dot
layer,
five
that
io
that
will
help
solicit
some
feedback
from
from
people
and
if
it's
like
well,
that's
not
really
helpful,
because
you
need
like
live
feedback
and
you
want
to
engage
in
a
discussion.
B
For
my
part,
I
have
time
to
hang
out
for
the
remainder
of
the
hour,
for
anyone
who
wants
to
you
know
get
feedback
on
their
stuff,
but
navenu.
I
know
we're
over
time
so.
A
Yeah,
thankfully,
like
so
thank
you
everyone
for
coming.
We
are
out
of
time,
so
we
have
missed
a
couple
of
items
and
some
of
these
some
of
these
has
been
updated
in
the
meeting
minutes
and
some
of
these
will
be
demoed
or
showed
in
slack,
and
we
have
some
continuous
integration
or
ci
items.
A
G
B
So
samir
that
might
be
a
that
might
be
enough
for
for
the
well
us
collect
it's.
Not
it's
not
like
not
like
everyone's
gonna
go
work
on
this
in
the
secret
or
we're
happy,
but
like
there's
a
spec,
there's
a
sequence
diagram
that
kind
of
shows
the
flow
of
what
happens
today.
B
So
to
your
point,
like
an
email
address,
with
a
pat
like
an
email
and
a
password,
a
username
and
password
like
that,
and
so
maybe
a
you
know,
verification
token
like
that.
Also
then,
just
in
your
case,
I
don't
think
it
helps
as
much
but
to
support
directory
integration
or
open
id
or
other
forms
of.
B
Credentials
and
so
yeah,
it
says
so:
let's
initiate
a
thread.
Let's
initiate
a
I'll
go,
get
the
the
sequence
diagrams.
We
can
start
to
brainstorm
sure
yeah
samia.
This
is
great.
It's
really
great.
This
is
this.
Is
nice
jean
miss
tiston?
Did
you
since
you're
still
here
and
dedicated
to
staying
up
to
the
wee
hours
of
the
morning?
B
B
Yeah,
if
I
recall
kind
of
where
we
left
off,
is
you're
enhancing
both
the
toaster
notifications
and
the
notification
center
itself
inside
of
mesheri
and
there's
sort
of
you're
seeking
out
an
opinion
and
a
perspective
on
just
how
many
buttons
should
be
there
to
control
the
opening
and
closing
of
the
notification
center.
The
drawer,
as
well
as
like
what
should
be
the
exact
behavior
of
some
of
those
buttons.
J
Okay,
can
you
see
my
can
everyone
see
my
screen
and
can
you
hear
me
hi
everyone.
J
Okay,
so
right
now
the
notifications,
the
toaster
notifications,
look
like
this.
It's
currently
in
dark
mode
and
my
changes
involve
making
them
in
night
mode
and
another
one
is
some
changes
on
the
door
and
the
buttons
here.
So
here
is
my
changes
right
now.
J
The
toaster
notifications
are
already
in
light
mode
and
then
the
notification,
so
the
door
sort
of
got
a
bit
of
a
change.
Look
it's
a
bit
wider
and
then
different
buttons.
J
So
this
one
does
the
clearing
of
all
the
notifications-
and
this
is
the
back
button.
But
I
think
there
was
a
comment
about.
J
D
D
J
Oh
yeah,
it
clears
all
the
notification
and
it
also
closes
the
drawer
at
the
same
time,.
D
J
So
should
it
be
just
clearing
everything
and
not
closing
it,
or
should
we
have
the
button
up
here.
B
Yeah
yeah
that
sounds
yeah.
I
would
give
a
plus
one
on
that
so
gene.
If
you
go
back
to
the
current
system,
the
current
the
the
icon
here,
the
like
slanted
hamburger,
you
know
the.
B
It,
I
think,
the
challenge
the
conundrum,
the
challenge.
The
reason
for
the
part
of
the
reason
for
the
requested
change
in
the
first
place
is,
is,
I
think,
the
potentially
non-intuitive
nature
of
what
that
symbol
means
and
for
me,
as
a
user,
if
I'm
facing
these
errors,
like
oh
istio,
didn't
install
oh
geez,
oh,
what's
going!
Okay,
what's
going
on
like
that's
that's
an
issue,
I
gotta
and
it's
gonna
tell
me
the
details
and,
but
you
know,
there's
it's
like
cluttered
with,
like
some
other
stuff,
like
oh
yeah,
something
else
was
successful.
B
B
They
click
on
things
and
they
are
expecting
a
behavior
and
they
get
that
behavior
and
it
feels
great
in
this
case,
it's
like,
like
the
only
change,
might
have
neces
like
the
only
necessary
change
potentially,
and
I
think
that's
what
we're
arriving
at
that's
what
samir
is
suggesting
and
basically
what
ayush
is
suggesting
as
well
is
he's
like:
hey
leave
the
behavior
the
same
just
change
out
the
icon.
B
I
think,
because
that
icon,
the
check
check,
I
think,
I
think,
is
slightly
more
intuitive
and
yeah.
If
you
don't
have
any
notifications
like
it
makes
sense
to
to
close
it.
For
me
it
always
when
I
had
first
started
to
use
it,
it
was
holy
hell.
What
just
happened.
I
just
clicked
the
button.
All
my
notifications
had
gone
and
the
drawer
went
away
and
it's
like.
I
want
my
stuff
back
and
I
can't
like
they're
gone.
B
It's
like
well,
I
didn't
that's
not
fair,
I
didn't
know
what
was
gonna
happen
and
you
know
like
it.
Just
I
felt
cheated
and
and
and
okay
fine.
Maybe
we
should
put
a
little
warning
that
says:
are
you
sure
you
want
to
clear
them
all?
B
You
know
something
and
we
could
leave
the
icon,
but
but
unless
I'm
mistaken,
I
think
the
double
check
is
a
little
more
intuitive
as
to
what's
going
to
happen,
especially
if,
if
you
go
back
to
the
new
design,
I
think
that
this
is
especially
true
if
there's
a
check
mark-
oh
okay,
see
so
this.
This
is
what
we
could
do
is
like
see.
B
Yeah,
I
think
so
because
actually
you
know
what
this
is.
This
is
great.
I'm
glad
that
we
had
this
conversation
and
you
didn't
just
do
a
recording
and
then
we
didn't
have
because
well,
when
you
think
about
it,
the
toaster
notifications
that
come
out
they
have
a
small
little
x
on
them
and
yeah.
If
you
hit
the
x
or
if
you
hit
the
notification
like
it
goes
away,
it's
it's
quite
ephemeral
like
it's
just
and
great.
That's
what
the
x
the
x
the
x
has
that
context
to
it.
B
Maybe
it's
like
it's
just
dismissed,
like
you
know,
get
out
of
my
way.
The
check
mark
here
is
indicated
is
slightly
different,
like
if
we
use
the
double
check
mark
in
a
single
check
mark.
It's
hey.
This
thing
isn't
gonna.
If
you
waited
long
enough
on
the
toaster
notification,
it
would
just
go
away
by
itself
because
it's
like
it
wasn't
important
enough
to
stick
around
these
other
ones.
No,
no.
B
They
have
a
small,
an
emergent
life
cycle
to
them
where
they're
gonna
stick
around
until
you
acknowledge
them
and
then
they'll
go
away
and
later
in
gene.
I
expect
that
you'll
be
working
on
this
and
that
is
like
later
there's
a
little
more
life
cycle
that
needs
to
happen
like
well.
There
was
an
error
here.
B
You
can
dismiss
it
with
the
check
mark,
fine
or
there'll,
be
like
a
recycle
arrow.
That's
like
no!
I
think
I
fixed
the
issue.
Why
don't
you
go
ahead
and
retry
that
same
operation?
Oh
great?
Okay?
So
then
there
becomes
a
life
cycle
around
what
this
is
to
do,
or
maybe
it's
like.
Oh,
I
want
to
send
that
to
slack
or
I
want
to
create
a
like.
There
may
be
other
actions
to
take.
J
B
Yeah
well
a
couple
other
like
hey.
If
I
could
get
really
nitpicky
with
you
for
a
moment,
and
that
is
one,
do
you
do
you
think
that
we'd
be
successful,
sliding
the
the
chevron
and
what
will
be
the
single
check
mark?
All
the
you
know,
all
the
way
all.
B
And
then
let
me
ask
you
this,
and
this
is
this.
Is
I
don't?
I
don't
have
an
answer,
I'm
asking
everyone:
what
does
the
notification?
What
does
the
bell
do
for
us
here.
B
It
should
be
it
should
it
be
inanimate,
or
should
it
be
like
an
auto
and
expand
all
or
collapse
all
along
as
people
think
about
that
understand.
That
gene
is
also
looking
at
predefined
filters,
because
the
errors
that
people
are
seeing
here
eventually,
they
will
have
severity
assigned
information.
B
B
D
If
we
just
have
a
drop
down
for
all
types
of
severity,
so
that
we
could
select
what
all
we
want
to
see,
just
like
our
console.
B
B
Yep
yep
the
check
that
yep,
that's
a
good
here's,
the
issue
with
the
drop
down
they're
ugly,
one
sorry.
B
To
be
actually
it's
not
as
much
of
that,
as
is
it
is
it
takes
two
clicks
to
do
the
same
thing
you
could
otherwise
do
in
one
click
now,
look
if
you
have
ten
different
types
of
notifications
like
yeah,
please
don't
show
me
all
10
buttons
at
once
like
let
give
me
a
drop.
Give
me
some
other
way
to
manage
this
information
in
this
case.
If
it's
just
a
handful.
D
F
B
Anybody
else
having
an
opinion
like
that
that
could
it
would
be
less
intuitive
there'd,
be
less,
it
would
be
less
discoverable.
B
D
Like
when
we
click
on
the
icon
that
bell
icon,
it
will
just
just
like
our
cards
are
expanding
downwards.
It
could
expand
right
words
and
have
the
options.
H
You
can
have
a
radio
button
all
or
show
all
comments,
or
you
know
just.
H
And
I
actually
have
another
question,
so
it's
nice
that
you
have
if
what
if
there
is
something
that
I
have
to
see
right
so
is,
is
there
something
that
indicates
hey
wait,
a
second
whatever
it
is
you're
doing
you,
you
have
to
see
this
notification
right
so
now.
I
know
that
there
is
a
notification,
so
I
have
to
click
on
that
notification
to
see
what
notifications
I
have
right.
So
it
is
that
that
is
dependent
upon
my
action,
so
it
it
it.
H
It
is
triggered
by
you
know
my
initiative,
but
what
what?
If
something
did
you
do
you
see
what
I'm
trying
to
get
at
or
what?
If
there
is
something
that
that
I
have
to
be
notified
of,
you
know,
how
does
it
do
that.
H
Just
as
an
example,
so
I'm
struggling
I'm
trying
to
do
something.
You
know,
and
it's
trying
to
tell
me:
hey,
listen,
you
you're
not
connected
to
the
network,
but
here
I
am
I'm
trying
to
run
all
kinds
of
tests.
I'm
trying
to
do
all
everything.
Nothing!
Actually
this
that's
not
a
very
good
question,
but
suppose
you
know,
let's
say
this
is
a
more
life-threatening
application.
You
know
and
say:
there's
a
patient
out
there,
who's
in
who's
who's
who
who's
who's,
got
a
severe
condition
they
have
to.
H
J
Yep
well.
I
H
That
is
correct,
but
the
problem
is
what,
if
I'm
not
looking
there
see,
I'm,
maybe
I'm
not
I'm
making
this
too
complicated,
you
know.
So
the
thing
is
I'm
working
on
it.
I'm
looking
at
something
that's
of
interest
to
me.
It's
at
the
top.
You
know,
I'm
I'm,
I'm
I'm
just
looking
at
the
test
result
or
something
like
that.
You
know
performance
test
or
something
like
that,
and
this
comes
at
the
bottom
of
the
screen.
It
just
pops
up
for
two
or
three
seconds
and
seconds,
and
then
it
goes
away
right.
H
So
is
there
something
that
you
know
that
that
will
listen
you?
You
really
have
to
look
at
this
there's
something
that
says.
D
Something
like
having
a
solution
solution
that
you
could
do
this,
because
it
is
not
here,
so
you
cannot
perform
this
task
too.
Something
like
that.
B
H
Something
like
that,
but,
but
you
know
again,
a
model
is
very,
very
crude,
so
so
the
notification,
if
it
has
a
flag,
you
know
or
yeah.
I
guess
you
know
if
it's
really
that
that
serious,
yes,
you
have
to
have
a
model,
but
but
if
there
is,
if
there
is
over
the
notification
area,
when
things
like
that
come
up,
you
know
you.
You
have
a
visible
red
flag
that
always
pops
up
whether
the
notification
screen
is
on
or
not.
E
B
B
Actually
we
used
to
anyway,
there's
teams,
there's
teams,
teams
of
teams
that
work
on
uniform
user
interface,
uniform
ux
and
user
research
and
etc,
and
so
so
there's
been
a
top
like
don't
take
for
granted
that
there
hasn't
been
a
ton
of
discussion
in
the
team
that
has
implemented
this
product
behind,
like
what
it's
doing
and
how
the
you
know
like
it's
fairly
well
designed
it's
what
I'm
getting
at
and
actually
it
looks
kind
of
good.
B
I
have
some
comments
about
how
it
looks,
but,
but
it
looks
pretty
and
what
I
wanted
to
help
us
reflect
upon
is.
This
is
a
system
that
manages
infrastructure.
What
does
measuring
do?
Well,
it
manages
infrastructure,
okay.
Well,
it's
pretty.
Okay,
pretty
similar
good
comparison;
okay,
fine,
fine,
fine,
fine!
So
here's
a
dashboard
and
we
just
land
by
the
way
I'll
give
you
guys
all
the
link
to
this.
You
can
come,
try
it
out
and
play
with
the
system
and
think
around,
which
will
be
helpful.
B
Just
don't
dink
around
too
long
like
go
back
and
do
some
coding
after
that.
That
is
so.
They've
got
a
dashboard,
there's,
there's
tons
of
things
to
look
at
and
then
go
through
and
it's
fun.
You
should
go
look
but
there's
there's
two
things
you
know
relevant
to
or
there's
three
things
I
think
relevant
to
what
we're
talking
about
right
now,
all
right,
so
we've
got
a
couple
of
things
in
your
face
about
the
health
of
this
system.
B
One
of
those
is
a
perpetual
navigation
bar
that
isn't
isn't
entirely
dissimilar
from
what
we
have
like
you
know
what
mastery
has
is
actually
fairly
similar
layout
in
many
respects,
so
gene.
Actually,
you
can
see
some
things
that
are
a
lot
easier
to
understand
over
here
that
there
are
categories
of
menus,
and
I
use
like
it's
not
as
easy
to
understand
that
there
are
categories
of
menus
in
mastery,
nor
that
they're,
expandable
or
collapsible,
and
so
that's
something
that,
but
over
here
there
is
a
notification
center.
Okay,
this
is
where
alarms
come
immediately.
B
We're
seeing
that
between
your
critical
and
your
warning,
these
two
severities!
This
is
what
you
got
going
on
good.
If
I
click
anywhere
on
here,
it
activates
deactivates,
the
sidebar
wait
a
second.
If
I
click
that's
a
little
weird
anyway,
if
I
click
on
it,
it
shows
this.
Oh,
it
has
filters
just
like
what
we
were
talking
about.
Oh
if
I
filter
just
on
critical
there's
six
of
them.
B
If
I
go
back
to
all
okay
great,
so
here
you
know
like
again
like
this
would
be
our
bell
all
and
then
these
would
be
the
icons
for
the
others.
Okay,
all
right!
Here's
an
alarm
now!
Can
I
dismiss
it?
No,
that's
just
the
signal
for
the
fact
that
it's
critical
is
when
it
happened.
Okay,
great.
What
do
I
do?
Well,
I
can
view
all
so
that's
maybe
another
way
of.
B
Maybe
it
takes
me
to
a
different
page
yeah,
where
there's
a
table
full
of
these,
so
just
a
different
like
different
information
architecture
for
the
same
data,
but
is
providing
it
to
you
in
a
different
way
in
which
you
can
provide.
You
know,
do
bulk
actions,
great
dismiss
them
all.
Maybe
is
my
bulk
action
or
do
that?
Okay,
fine,
maybe
there's
ones
that
I
have
acknowledged.
Oh
you
mean
the
check
mark
that
we
were
talking
about
yeah,
okay,
so
you're
here,
one
nice
thing
and
we'll
get
to
this.
B
B
We
don't
have
to
walk
through
all
this.
Just
a
good
thing
to
reflect
upon
how
another
well-designed
system
has
approached
it.
They've
got
this
little
x.
What
does
the
x
do?
Oh,
it
closes
the
bar
okay,
fine,
so
I
can
close
the
bar
with
the
x.
I
can
close
the
bar
by
clicking
on
it
again.
Can
I
close
the
drawer
sorry
by
clicking
somewhere
else,
yeah?
Okay,
it's
fairly,
it
wants
to
go
away
like
by
default.
It
doesn't
want
to
be
shown.
Okay,
fine.
B
There
is
a
cousin,
a
sibling
drawer
called
requests.
What's
a
request,
a
request
is
our
operation
when
you
go
to
install
istio
service
mesh
like
the
error
that
gene
is
seeing
for
us,
that's
that's
a
request
that
that's
an
operation
like
hey.
I
want
you
to.
I
want
this
management
system
to
perform
this
action.
The
actions
are
they
don't
always
happen
like
that,
like
they
may
take
four
steps
to
execute
the
request
it
may
take
an
hour
or
two
hours
or
a
day,
those
requests.
Those
operations
may
fail
anywhere
along
the
way.
Okay.
B
So
there's
this
concept
that
there's
a
life
cycle
for
what
they're
calling
an
alarm,
what
we're
calling
a
notification,
a
small
life
cycle
that
they're
facilitating
to
dismiss
them
or
export
them,
or
do
you
know,
sort
them
and
do
whatever
then
there's
also
requests
that
you
know
like
I
don't
I'd
have
to
here's
all
the
requests.
This
one
was
successful
at
this
one.
They
were
trying
to
turn
on
a
virtual
machine.
It
failed
if
I
click
on
it.
It
does
deep
link
me.
B
B
It
sounds
like
we'll
see
how
well
this
works.
It
sounds
like
you'll
have
like
thank
you.
Nutish
you'll
have
like
two
or
three
icons
that
are
filters,
one
for
errors,
one
for
successes,
one
for
infos
you'll,
probably
leave
the
notification
bell
to
perform
a
show.
All
action,
basically
to
clear
the
filter.
B
B
And
it's
kind
of
yeah,
then
maybe
we
don't
need
it
like.
I
would
suggest
that
we
don't.
The
other
thing.
Is
you'll
you'll
write
a
line,
the
you
know
all
the
things
you
will
put
a
single
check
mark
instead
of
the
x,
and
you
will
also,
hopefully,
maybe
add
a
little
bit
of
like
bolding
or
italics,
or
something
to
more
crisply
set
apart.
B
J
J
Yeah,
I
think
I
I
think
I
had
a
discussion
with
michael
about
that
that
the
events
that
are
being
captured
are
not
in
the
same
format,
not
in
the
new
format.
Yet
so
the
details
about
the
other
details
are
not
yet
part
of
the
object
that
was
being
passed
onto
those
notifications.
B
I
know
abhishek
is
tracking
toward
changing
those,
but
so
maybe
we'll
leave
the
hyperlink
out
as
a
separate
concern,
but
I
so
then
I
think
yeah.
That
was
all
the
does
that
get
you.
What
you
need
is
that.
J
Yeah,
thank
you
so
much
to
everyone
who
commented,
and
I
already
listed
out
the
things
that
I
need
to
change
so
yeah.
Thank
you.
So
much.
B
It's
looking
great,
it's
looking
good,
but
it
just
has
a
lot
as
one
last
statement
for
me.
I
know,
ayush
had
been
sort
of
you
know,
considering
and
working
toward
consideration
of
how
we
might
ultimately
have
have
themes
like
have
a
light
mode,
have
a
dark
mode,
something
to
be
chatting
with
him
about
or
sort
of
thinking
on,
as
you
go
through
this.
B
Actually,
so,
actually,
where
we
actually
are,
it's
just
sorry
pranav
thanks
for
coming,
we
we
just
we're
just
ending
the
call.
B
B
K
It's
pretty
much
done
just
one
limitation
is
that
I'm
not
yet
handling
types,
so
jason
can
have
num,
integers
or
strings
and
stuff
like
that.
I
need
to
modify
the
the
config
for
the
for
allowing
sami
to
mention
or
specify
what
type
the
header
is.
So
once.
B
Thank
you
all
very
good.
It's
been
like
a
almost
a
two-hour
marathon
for
some
of
us,
so
thanks
a
bunch.
Thank
you
for
hanging
out
talk
to
you
guys
soon.