►
From YouTube: 2022-05-02 meeting
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
Okay
looks
like
most,
people
are
putting
their
name
on
the
attendee
list
and
I've
seen
a
whole
bunch
of
new
comments.
This
morning
I
haven't
gotten
a
chance
to
take
a
full
look
at
yet
okay,
so
let's
go
ahead
and
get
started
and
then
we
can
come
up
with
any
other
items
as
we
go.
So
I
already
have
three
proposed
agenda
items
so,
first
of
all,
welcome
to
the
demo
web
store
sick.
This
is
our
first
meeting,
which
is
super
super
exciting.
Thank
you
all
for
contributing.
A
I'm
sure
it
will
be
incredibly
valuable.
So
the
goal
of
this
effort,
of
course,
is
to
build
a
community
demo
that
can
be
used
by
you
know
a
wide
range
of
users
to
easily
get
up
to
speed
on
open,
telemetry
and
also
see
a
fairly.
You
know
real
world
realistic
example
of
what
open
telemetry
would
actually
look
like
in
a
real
architecture
so
for
our
agenda
items
today
the
first
is
just
to
surface
any
other
demo
applications
that
exist
that
we
could
potentially
use
as
a
fork.
A
I
know
I've
been
looking
a
lot
at
the
google
splunk
and
dynatrace,
which
all
kind
of
come
from
the
same
general
tree
of
the
google
app,
but
there's,
of
course,
other
apps
out
there
as
well.
So
please
call
those
out,
and
so
we
can
make
a
decision
not
today
necessarily
but
have
a
full
portfolio
to
make
a
decision
on
what
we
want
to
afford
going
forward.
A
C
Looks
good
to
me
so
far
to
me.
Thanks
for
the
introduction
carter.
A
Yeah
no
problem
nice
to
see
you
armin,
okay,
so
let's
go
ahead
and
kind
of
jump
into
it.
So
I'm
familiar
with
the
gcp
and
the
a
couple
of
the
other
demos.
Do
we
have
anyone
here
that
can
kind
of
speak
to
to
some
of
the
existing
demos
that
were
added,
maybe
like
the
tag,
app
delivery
or
tag
observability?
B
D
B
B
I
don't
know
exactly
how
I
feel
like
it's
maybe
a
year
and
a
half
from
when
we
like,
maybe
a
year
year
and
a
half
old
or
maybe
even
two
years
old.
Now,
the
I
don't
know
I
mean
it.
Would
you
know
it's
definitely
the
same
spirit
as
the
gcp
one,
I'm
not
like
super
happy
with
everything
we
did
with
it.
It's
a
real
mix
of
like
hotel,
datadog
jaeger,
a
few
other
things.
B
It
was
sort
of
focused
on
like
what
could
we
ingest
and
that
lift
that
kind
of
confused.
I
think
so.
A
E
E
I
was
studying
open
telemetry,
so
the
the
project
that
I
I've
shared
in
the
previous
document
is
not
related
to
data
trace.
What
is
actually
nice
for
our
purpose
at
the
moment.
So
what
I
did
I
cleaned
up
everything
that
I
had
from
that
the
project
had
from
google.
So
all
the
cloud
stack
and
cloud
operations
stack
driver,
open,
sensors,
basically
everything
that
wasn't
open
telemetry
and
then
I
instrumented
all
the
services
with
open
telemetry.
E
A
Okay,
great
yeah,
thanks
for
giving
a
current
state
of
that
and
then
is
morgan
on
the
call.
I
don't
think
so.
A
Okay,
no
well,
then
we'll
move
on.
So
we
also
have
some
sto
demos.
Do
we
have
anyone
that
could
speak
to
those?
I
know.
No
one
from
aws
is
here.
A
Okay,
that's
fine!
If
no
one
else
on
the
call
can
speak
to
any
demos.
I
was
thinking.
Maybe
we
could
you
know,
potentially
as
kind
of
a
target
outcome,
maybe
assign
some
of
us
to
go,
investigate
the
existing
demo
list
and
maybe
come
back
with
a
recommendation
next
week.
So
we
have.
B
B
Broad,
maybe
not
broad,
but
it's
just
it's.
It's
really
focused
on
black
box
telemetry
using
sdo
right
like
the
the
whole
concept.
Is
that
you're
doing
most
of
your
monitoring
everything
from
istio
itself
or
you're,
getting
traces
and
stuff
through
the
zipkin,
whatever
I'd
have
to
or
you're
getting
it
through
envoy
right.
B
So
and
also
I
don't
know
what
istio
and
envoy's
timeline
is
on,
like
I'm
sure
at
some
point
like
maybe
even
now
I
haven't
looked
deeply
but,
like
I
don't
know
how
they're
doing
on
upgrading
to
hotel
and
w3c,
but
I
know
that
needs
to
land
an
envoy
stable
and
then
it
would
have
to
land
in
istio
and
istio
seems
like
they're
kind
of
in
the
middle
of
some
stuff,
with
this
whole
cncf
adoption
or
donation.
So
whatever
I
would.
B
Maybe
I
I
feel
like
we
can
kind
of
cut
out
a
lot
of
this
by
just
saying
like
what's
our
actual
priority,
I
think
it
needs
to
be
like
something
where
it's
like
polyglot
micro
services,
white
box,
mostly
white
box,
instrumentation
question
mark
is
like
the
third
point,
but
that
knocks
out
the
istio
one
I
think,
and
maybe
the
yeah
thumbs
up
from
riley
cool
cool,
the
jaeger
ones.
The
hot
rod
stuff
from
jaeger,
like
I
think,
is
good
at
showing
tracing,
and
it's
actually
like
an
interesting.
B
You
know
I've
used
it
before
it's
an
interesting
enough
demo,
but
it's
all
go
with
my
recollection
and
doesn't
satisfy
the
polyglot
requirement.
I
did
haven't
looked
at
the
logs
one.
B
B
Basically,
I
don't
yeah.
I
think
I
think
that
and
the
gcp
was
maybe
the
least
offended
the
smallest
defender
here,
but
it
still
has
quite
a
few
like
gcp
integrations
real
world.
I've
tried
I've
screwed
around
with
real
world
stuff,
and
I
know
some
of
my
friends
over
at
honeycomb
did
as
well
trying
to
add
in
hotel
there
and
that
I
feel
like
is
there's
a
lot
of
the
best
way
I
can
describe
it.
Is
it
felt
like
a
lot
of
work
to
not
really
get
anything
like
it?
B
It
felt
too
small
right,
like
it's
really
a
pretty
like
the
actual.
What
you
get
is
kind
of
a
simple
enough
client
server,
and
it
was
like
a
lot
of
work
to
sort
of
deal
with
the
scaffolding
and
everything
and
what
you
got
at
the
end
wasn't
maybe
super
interesting
from
like
an
observability
point
of
view,
because,
yes,
you
can
like
mix
and
match
and
plug
things
in,
and
I
thought
that
part
was
cool,
but
it
it
wasn't
super
useful
from
like
an
instructional
point
of
view.
B
I
had
more
questions
than
answers
and
I
haven't
seen
the
app
delivery
one
pod
tato
head,
but
points
for
creativity
on
the
name.
A
Yeah,
that's
a
good
one.
Okay,
awesome
thanks
for
that
pretty
high
level
overview,
so
it
sounds
like
we'll
probably
go
with
the
gcp
app
in
some
form,
so
I
think
we
still
need
some
sort
of
recommendation
between
you
know
either
splunk
julianna's
or
the
actual
gcp
one
itself,
or
also
lightstep.
I
guess
as
well.
A
So
I'm
not
sure
if
we
once
again
it
might
be
better
just
to
split
into
a
smaller
group.
So
does
anybody
like
a
couple
of
us,
want
to
take
it
on
and
maybe
make
a
specific
recommendation
from
among
these
giuliano?
I
know
you're
already
familiar
with
your
fork,
so
would
you
maybe
be
willing
to
dig
into
the
other
ones
as
well.
A
Sorry
need
we
need
a
formal
sorry.
I
probably
should
have
clarified.
We
need
a
formal
recommendation
for
among
the
gcp
apps,
which
one
we're
going
to
fork.
So
could
you
and
maybe
some
others,
would
you
mind
taking
a
look
at
the
three
or
four
existing
ones
and
maybe
make
it
a
recommendation
on
which
one
we
should
go
with.
G
Thanks
james,
I
have
a
question
so
once
we
once
we
figured
out
which
repo
are
we
going
to
work
on
like
I,
I
guess
we're
going
to
fork
it
or
we
just
copy
the
code
somewhere
like
under
the
open,
telemetry
demo.
Repo
and,
like
I
wonder,
like
what's
the
what's
the
end
of
the
game,
let's
say
the
end
of
the
game
is
there's
another
fork
or
the
end
of
the
game
is:
there's
a
repo.
G
It's
not
forked,
anywhere
from
anywhere
it's
just
a
copy
of
some
existing
code
and
after
that
google
and
other
folks
are
waiting
to
remove
their
ripple
or
they
want
to
retarget
and
create
a
fork,
because
I
feel
if,
in
the
end,
there's
another
fork,
maybe
it's
not
ideal
situation,
but
it
might
work.
It's
just
a
like
a
vendor
neutral
fork.
Eventually,
we
clean
up
everything,
but
the
problem
is:
if
the
upstream
has
changes.
What
do
we
do?
G
I
I
guess,
that's
a
hard
question
number
two:
if
the
end
goal
is
there's
no
fork,
there's
the
open,
telemetry
demo
that
works
and
all
the
vendors
are
happy
not
having
the
fork
that'll
be
ideal,
but
I
want
to
see
like
if
anyone
see
the
challenge.
The
third
situation
is
open.
Telemetry
demo
is
the
the
source
of
the
repo
and
then
the
vendors
still
have
their
form.
Then
it
seems
there's
no
huge
difference
versus
the
current
gcp.
G
B
I
just
ideally
in
my
mind
like
in
the
most
perfect
world,
in
my
mind,
riley
you're,
I'm
with
you,
there's
no
fork.
What
there
would
be
is,
let's
say:
if,
if
we
have
the
canonical
upstream
hotel
demo,
then
a
fork
would
really
just
be
a
config
fork
and
it
would
be
here's
a
collector
yaml
that
sends
data
to
you,
know
microsoft's
thing
or
amazon's
thing
or
whatever
right.
B
I
suspect,
that's
actually
not
going
to
be
feasible,
though,
because
I
feel
like
I
mean
I
think
abs
is
the
easy
example
here,
because
so
much
of
what
they
want
to
show
off
is
integrating
this
into
x-ray
and
their
stuff.
That
isn't
like
first
party
hotel
and
I
could
sort
of
see
that
with
like
stack
driver,
I
could
sort
of
see
that
with
like
splunk
any
you
know
any
anyone,
that's
kind
of
running
ahead
of
spec
and
running
ahead
of
upstream
and
like
producing
add-ons
to
hotel.
B
B
I
think
in
the
most
perfect
world
this
becomes
them
the
source
of
truth,
and
then
you
would
st.
You
would
probably
still
see
like
vendor
or
platform
specific
forks
that
had
you
know,
custom
features
and,
like
you
know,
integration
with
just
their
thing.
So
so
maybe
the
you
know
the
work
product
here
is.
We
are
helping
everyone
we're
helping
all
the
pms
at
all
these
places
you
know,
get
some
get
a
good
foundation
for
whatever
their.
G
G
B
B
This
is
sort
of
the
canonical
like
upstream
hotel
demo,
and
it's
always
going
to
be
like
hotel
guarantees
that
this
is
going
to
be
kept
up
to
date
with
whatever
is
like
bleeding
edge
and
hotel
right
or
not
even
bleeding
edge,
whatever's
stable
in
hotel.
So
if
there's
stuff
that
as
an
example
like,
maybe
we
want
to
show
off
our
terraform
integration
right
so
a
p,
you
know
an
engineering
team
might
fork
this
and
say
like
okay,
we're
basing
enough
of
this
and
then
we're
going
to
add
in
this
terraform
stuff
that
automatically
creates
dashboards.
B
B
We
can't
stop
people
from
forking
it
and
in
my
perfect
world
at
least,
there
will
be
a
pure
play
open
source
upstream
reference
implementation
that
people
can
go
out,
because
I
mean
in
my
mind
this
this
kind
of
gets
into
your
target
personas
the
the
third
question.
So
let
me
just
like
tackle
that
for
a
second.
In
my
mind,
the
real
thing
I
want
to
solve
for
here
is
like
someone
at
a
company
that
says
I'm
gonna
start
I'm
gonna.
Oh
here's!
B
This
open,
telemetry
thing
that
I
want
to
advocate
internally
right
and
I'm
gonna
do
a
brown
bag
session
and
I
need
a
demo
app
or
I
need
something
like.
I
need
content
so
that
I
can
go
and
show
this
to
my
team
or
to
my
co-workers
right
and
we
can
either
as
a
project,
provide
that
content
or
vendors
will
provide
it
for
us.
So
I
think
having
an
opinionated
like
this
is
how
we
think
you
should
use
hotel.
This
is
how
we
think
you
should
integrate
hotel.
B
B
C
Yeah
and
even
someone
just
stumbling
upon
open,
telemetry
and
trying
to
understand
what
it's
about
and
studying
the
docs
could
benefit
from
from
just
having
a
simple
way
of
spinning
up
a
working
environment
where
they
can
dig
into
things
and
there
it
really
shouldn't
matter
too
much
what
the
what
the
backhand
is.
So
if
we
ship
it
with
with
jaeger,
sdko2,
open
source
solution
and
maybe
grafana
for
matrix,
then
they
can
already
see
the
the
bits
and
pieces
working
together
and
and
have
an
easier
time
learning
what
what
hotel
is
all
about.
B
B
So
I
think
it
also
helps
us
like
advocate
in
our
own
teams
and
our
own
organizations
for
like
how
do
you
use
open
telemetry
if
you're
going
to
someone
that
doesn't,
you
know,
isn't
on
the
open
source
side
of
things
right
like
if
you're
going
to
a
pm
or
an
engineering
team,
you
know
and
you're
trying
to
be
like
well,
here's
how
it
works
like
this
is
something
you
could
point
them
to.
That
can
kind
of
do
everything,
self-contained
and.
B
A
That'd
be
great,
anybody
else
have
any
input
on
some
sort
of
personas
we're
targeting
outside
of
maybe
an
enthusiast
at
a
company,
or
maybe
someone
looking
to
onboard
to
get
kind
of
a
bigger
picture
of
how
things
integrate.
H
I
think
it's
very
easy
to
forget
about
developers
in
that
discussion.
Think
about
a
new
developer,
that's
working
in
a
language
that
they've
worked
in
for
a
long
time
and
then
they're
like
oh
man.
I
want
to
see
how
this
works
and
plays
in
with
a
bunch
of
other
languages
and
frameworks,
and
all
these
other
things
to
develop
like
the
developer
mindset
might
be
important
here
too,
even
though
we
sit
so
close
to
the
problems.
Please.
I
A
Yeah
give
a
more
robust
example
out
there:
it's
not
just
hello
world,
but
with
a
trace
attached
to
it.
Yeah.
Okay,
great!
Well,
I
think.
Okay,
so
we
have
a
good
sense
of
developers
and
enthusiasts
they're,
they're,
probably
slightly
overlapped,
but
I
would
say,
like
enthusiasts,
probably
have
pms
in
that
bucket,
so
expect
a
slightly
lower
profit
technical
skill
set.
We
also
are
targeting
the
gcp.
A
C
The
topic
I
want
to
ask
which
personas
did
you
have
in
mind
when
you
initially
came
up
with
the
proposal
overall.
A
Yeah,
that's
a
great
question,
so
I
took
I
looked
at
it
from
an
enterprise
perspective
so
say
like
like
an
enterprise
is
trying
on
or
to
open
telemetry
what
could
be
like
a
real
world
example
for
them,
so
they
could
actually
see
a
realistic
architecture
and
with
the
open
telemetry
implemented.
So
if
I
was,
I
guess,
I'm
kind
of
looking
at
it
from
the
opposite
of
the
enthusiast
perspective.
It's
almost
from.
Like
the
you
know,
technical
requirements,
perspective
of
this
company's
considering
making
a
shift.
A
What
does
the
real
world
example
look
like,
but
I
also
come
kind
of
from
the
enterprise
space.
That's
always
kind
of
my
default
mindset
too.
G
Go
ahead,
yeah.
Another
thing
I'm
thinking
is
for
certain
apm
vendors
who
who
are
not
currently
involved
in
open
time,
should,
when
they
think
oh
open
time
tree
is
becoming
popular.
I
have
my
customers
asking
me.
I
want
to
see
if
my
product
works
well
for
open
telemetry.
This
serves
as
a
good
starting
point
for
them
to
evaluate
how
far
they
are
and
what
are
the
top
part
is
for
them.
A
Okay,
any
other
target
personas
for
the
demo.
We
should
keep
in
mind.
A
I'm
gonna
take
that
as
a
tentative.
No,
yes,
let's
go
to
the
application
requirements
doc
and
I've
seen
a
whole
ton
of
comments.
So
thank
you
all.
You
know
for
contributing
on
that
and
also
you
know,
working
to
clarify
things
as
well.
A
So,
let's
see
so,
I
think
the
kind
of
the
consensus
here
is
this
application
has
to
be
modular.
I
know
we
have
like
five
languages
represented
a
day.
We
potentially
have
six
that
will
need
to
be
represented
in
the
future.
I
think
armin
was
making
some
great
comments
around
you
know.
What
do
we
consider
a
supported
language
to
be
so
they
have
to
have
some
level
of
maturity
before
before
getting
adopted
into
the
the
demo
itself.
That
probably
contributes
to
a
reporting
question
too.
So.
A
So
I
think
one
thing
we
should
probably
focus
on
real
quick
is
so
how
are
we
going
to
make
this
both
a
modular
architecture
and
also
one
that
can
be
run
on
many
different
platforms
in
a
relatively
simple
way?
So
we've
talked
a
bit
about
you
know,
for
example,
using
a
docker
compose
and
then
austin
also
raised
the
point
of
that
they
have
a
lockdown
corporate
laptop
which
may
not
have
a
great
definition
of
what's
capable
on
it.
You
know
how
can
we
provide
some
sort
of
degraded
experience
too?
A
B
B
Familiar
with
like
the
cleverness,
if
that
makes
sense
so
like
let's
say,
we've
got
something:
that's
like
you're,
a
python
dev
you
jump
in
and
there's
like
one
python
service
and
you're
like
cool.
I
want
to
see
how
this
connects
to
another
python
service
and
it's
like
maybe
there's
only
the
one
python
service
right,
so
you
jump
in
it's
like
well,
you
kind
of
know
java,
and
then
you
jump
into
the
java
one,
but
there's
a
lot
of
like
weird
clever.
You
know
repository
pattern
or
whatever,
where
it's
hard
to
actually
understand.
B
What's
going
on,
because
we
had
to
write
it
in
this
really
modular
way.
I
feel
like
that's
something
to
avoid
is
like
how
how
far
can
we
go
like?
Maybe
even
it
means
like
we
write
it
in
a
week
and
maybe
in
not
a
great
way
where
or
maybe
we
write
it
in,
like
a
really
dumb
way
is
actually
where
I'm
kind
of
getting
this
like
what?
If,
instead
of
like
what
if
database
access
is
all
through,
you
know
it's
like
we
need,
so
we
need
something
to.
B
We
won't
be
able
to
swap
out
the
data.
The
data
layer
cool
one
way
you
could
do
that
would
be
to
have
services
just
maintain
their
own
databases.
Another
would
be
to
say
like
okay,
the
data
all
data
access
goes
through
a
completely
independent
service
and
then
we
own,
and
then
we
use
like
grpc
or
just
http.
B
H
I
love
I
love
that
statement.
Austin,
like
simple,
is
better
than
complex
here
because
of
all
the
intricate
moving
pieces.
The
other
thing
that
I
think
should
be
a
goal
of
this
project
is
to
use
the
most
language
independent
tools
and
frameworks
that
we
have
available
like
using
a
make
file
for
a
lot
of
the
things
probably
makes
a
lot
of
sense
using
grpc
for
communication
makes
a
lot
of
sense.
These
things
that
to
try
and
use
things
that
are
independent
of
the
languages
that
implement
the
process.
H
It's
important
because
we
want
to
separate
like
we
don't
want
to
have
specific
things
that
we
have
to
do
and
install
python
libraries
to
work
on
a
python
thing
and
then
install
specific
versions
of
php
to
do
a
specific,
php
thing,
and
I
think
that
you're,
just
getting
like
a
very
strange
dependency
held
very
quickly,
so
anything
we
can
do
to
avoid.
That
is
probably
super.
Smart.
B
I
think
that
we
should
probably
figure
out
like
the
one,
the
one
thing
that
is
kind
of
the
like.
What
is
you
know
and
I
think
for
like
the
hipster
shop,
it's
grpc
right.
So
if
you
say
like
okay,
well,
grpc
is
sort
of
like
the
one
thing
that
has
to
be
the
same
everywhere.
Then
at
least
that
gives
you
options
in
terms
of
like
regenerating
your.
You
know
you
you
can
just
like
build
new
protos.
You
know
build
a
new
server
for
whatever
language
and
using
the
tooling.
B
I
don't
know
if,
like
maybe
grpc,
is
too
heavy
or
maybe
it's
one
of
those
things
where
people
like
I.
I
would
assume
that
it
would
be
fine,
because
that's
like
one
of
that's
something
that
everyone's,
I
think
every
language
supports
grpc
instrumentation,
but
maybe
we
even
say
like.
Oh
that's,
too
much,
let's
drop
down
and
say
everything's
going
to
use
us
in
http
json.
B
You
know
api
and
we're
going
to
use
swagger
to
kind
of
build
all
those
defs
out,
and
then
you
can
just
use
swagger
cogen
to
create.
You
know
a
copy
of
every
language
but
php
for
servers.
B
Yeah
I
mean,
like
I
said:
http
might
be
the
the
actual
lowest
common
denominator
here
and
if
that's
the
case
then
cool
like
I
don't.
I
think
that
would
be
a
big
change
from
how
hipster
shop
works,
so
that
would
require
a
lot
more
work.
So
maybe
maybe
we
can
two-tier
this
and
we
say
like
okay
for
the
new
stuff,
it's
all
http.
F
Just
to
maybe
add
something
here
and
I
work
with
giuliano
and
we
use
the
hipster
shop
internally
to
add
some
extra
services,
one
being
php,
which
we
had
the
lovely
experience
of
discovering
that
the
grpc
wasn't
working
so
well
with
the
server
side
of
things.
But
we
did
exactly
as
austin
just
said.
There
just
used
http
when
we
wanted
to
make
it
a
bit
faster
to
kind
of
get
it
out
and
show
our
pms
the
the
language
with
the
hipster
shop.
F
So
yeah,
it's
also
a
nice
one,
as
has
been
previously
said
in
terms
of
code
complexity.
It's
it's
quite
nice
to
work
with
as
well
and
as
well
just
to
say
that
the
grpc
stuff
as
well
is
easily
extendable
for
for
the
services
there
too.
So
that'll
be
something
nice
to
have.
A
H
Yep,
I
think
that
php
is
annoying
compared
to
a
lot
of
other
languages.
For
this
reason,
so
I
don't
want
to.
I
don't
want
to
be
prescriptive
because
my
language
is
goofy
compared
to
the
rest,
but
we'll
we'll
slap
through
it
with
php
just
wanted
to
make
sure
that
people
need
that
caveat.
D
Okay
cool:
let's
oh.
B
A
I
think
we
were
pretty
settled
on
the
connector,
so
how
do
we
make
the
demo
as
easy
to
set
up
as
possible,
so
that's
kind
of
some
foundational
stuff?
I
think
everyone's
kind
of
a
plus
one
to
the
collector
at
this
point.
So
if
anyone
has,
I
guess,
a
point
against
using
the
collector
instead
of
direct
exporters.
You
know
now
is
your
opportunity
to
raise
it.
A
A
A
We
can
probably
clarify
that
as
well.
Would
you
put
that,
under
the
open,
telemetry
requirements
to
like
have
collector
yeah.
J
Exactly
and
that's
generally,
the
one
I
think
most
of
us
usually
recommend
to
people
is
use
the
collector
to
capture
host
metrics,
use
it
to
also
pull
in
stuff
from
your
other
data
sources,
including
the
hotel,
sdks.
B
I
J
Yeah,
I
would,
I
would
do
both,
like
apologies
for
joining,
laid
out
a
conflict
earlier,
but
like
honestly,
this
is
one
of
the
few
circumstances
I
think
is
software
engineers,
where
in
some
sense,
our
goal
is
to
almost
create
like
some
disgusting
mess,
because
that
better
represents
like
this.
This
is
being
this
demo
is
being
used
to
show
people
like
this.
This
is
an
example
of
hotel
used
in
a
pseudo
production
service
production
services
are
often
disgusting,
messes
and
so
yeah.
B
You
missed
it
morgan
I
I
did
yeah.
One
of
my
preface
points
was
like
we
should
like
try
to
avoid
being
clever
like
this
does
need
to
be
pretty
exactly
yeah.
It
doesn't
need
to
be
clever,
it
doesn't
need
to
be
polished
like
it
can
be.
We
can
do
things
the
wrong
way
or
the
bad
way
as
long
as
they
exactly
achieve
a
goal,
because
yeah.
C
B
Variety
of
reasons,
the
one
thing
I
do
want
to
comment
on
this
is
I
put
this
in
the
dock,
but
I
think
it's
really
important,
because
I
get
a
lot
of
questions
about
this
is
we
need
to
make
sure
that
we
have
services
that
are
using
a
mix
of
auto
and
manual
because
there's
a
lot
a
lot
of
people
like
I
installed
the
spring.
A
Yeah
absolutely
yeah
that
makes
sense.
Okay
yeah,
so
we
have
auto
instrumentation
and
manual
instrumentation
covered
we're
using
the
reference
collector
architecture.
I
think
so
some
of
the
open
questions
we
have
right
now
or
do
we
want
to
bring
in
platform?
You
know
generic
components
that
we
have
to
think
about
in
architecture
that
could
potentially
accommodate
that,
and
we
also
need
to
think
about.
You
know
what
sort
of
technologies
we
want
to
move
forward
with.
That
would
make
it
extremely
easy
to
run
locally.
A
B
B
I
mean
it
feels
like
we
should
target
like.
We
can't
expect.
We
can't
be
universal,
but
I
think
if
it's
something
like
okay,
this
works
with
kubernetes.
It
works
without
kubernetes
through
docker
compose.
You
can
point
the
compose
file
at
a
kubernetes
thing
you
can,
we
can
have
a
makes
file
to
run
this.
Like
I
love
tilt,
I
love
a
lot
of
these
things,
but
everyone
theoretically
can
run
kubernetes
and
also
like
this
is
cloud
native.
So
I
feel
like
we
lose
our
funding
if
we
don't
involve
kubernetes
in
some
way,.
B
Yeah,
like
I
think
kubernetes
I
mean
it's
2022
I
feel
like
kubernetes
is
at
least
you
know
enough
of
a
general
thing.
That
also
is
gonna
and
also
it's
gonna
make
everyone
else
happy-ish,
because
everyone
like
cloud
wise
can
do
it
and
you
can
also
see
a
like.
Oh
you,
do
it
the
hard
way
where
you
spin
up
a
vm,
and
then
you
install
kubernetes
in
your
vm
or
mini
cube.
B
B
Oh
there's
a
lot
of
variables
there?
Maybe
we
just
say
like
okay,
my
sequel,
everyone
has
a
my
sequel.
Everyone
has
some
sort
of
blob
storage.
Everyone
has
some
q
or
redis.
You
know,
reddish-shaped
object,
kafka
shaped
object.
H
Oh
you're
good,
I
was
going
to
say
that
also
probably
should
be
a
goal
of
this
project.
In
the
beginning,
we
should
probably
set
some
maximum
system
requirements
so
that
developers
can't
easily
spin
this
up
on
their
machines.
It's
easy
to
like
slap
all
these
things
in
and
then
you
need
64
gigs
of
memory
to
run
this
hotel
example
application.
Then
nobody
will
do
it.
So
maybe
we
should
make
sure
that
we
capture
that
requirement
early.
J
K
B
B
A
B
A
But
I
think
we're
about
at
times
so
thanks
everyone
for
joining,
I
think
we've
matured
the
requirements
a
good
bit,
so
once
we
get
the
application
donated,
I
think
we
can
start
making
both
some
docker
compose
changes
and
then
also
doing
some
more
architecture
work.
So
I
think
throughout
the
week
we
should
just
try
and
firm
up
this
document
and
then
try
and
work
on
the
actual,
maybe
like
architectural
planning,
so
potentially
need
a
separate
dock,
for
that
is
anybody
super
super
passionate
about
the
architecture.
A
B
Help
yeah,
I
can
help
I'm
just
like
this
is
a.
I
know
this
is
a
bad
month
for
everyone
because
of
kubecon.
Oh
it's
bad
month
for
some
of
us
because
of
kubecon.
B
But
after
after
that,
my
schedule
should
clear
up
a
little
bit.
A
Okay
cool
well,
I
appreciate
that
I'm
sure
other
people
will
be
more
than
willing
to
contribute
as
well
so
morgan
and
austin
kickstart
architecture,
design,
okay,
cool
yeah-
so
I
guess
just
to
review
giuliano
and
james
will
be
expecting
some
sort.
A
Among
the
gcp
existing
apps
and
then
morgan
austin
will
kick-start
their
architectural
design
and
we'll
also
kind
of
continue
to
have
the
conversation
and
the
demo
application
requirements
document
as
well.
A
Yeah,
of
course,
it's
at
a
great
level
of
support,
so
thanks
to
y'all
as
well,
and
with
that,
I
think
we
can
go
ahead
and
call
it
have
a
nice
day
afternoon
evening
morning.
You
know
what
have
you
and
sarah.