►
From YouTube: wasmCloud Community Meeting - 20 Sep 2023
Description
Welcome to the wasmCloud community! Tune in live where we discuss the latest developments in the wasmCloud ecosystem, WebAssembly standards, and break out sweet demos.
Agendas for wasmCloud community meetings can be found at: https://wasmcloud.com/community
A
All
right,
hello,
everyone
welcome
to
the
wasm
cloud
community
meeting,
Wednesday
September
20th
we're
going
to
start
off
with
not
one
but
two
demos
today,
which
is
very
exciting
and
then
get
into
some
some
pretty
Lively
discussion
around
metrics
and
then
the
0.78
official
release,
so
I'm
going
to
go
ahead
and
hand
off
to
Vance
who
is
actually
live
at
the
TM
Forum
dtw23
I'm
talking
about
wasmcloud,
of
course,
a
little
bit
but
is
heading
up.
The
webassembly
canvas
Catalyst
there
in
that
effort,
so
fans
I'll.
B
All
right,
thank
you.
Well
yeah
I'm,
here
on
the
show
floor,
the
show's
closed
for
the
day
it's
7
P.M
now
I'm
in
Copenhagen.
This
show
is
a
it's
put
on
by
the
TM
Forum,
which
is
a
regroup.
The
managers,
the
the
software
systems
that
telephone
companies
use
to
run
their
business.
So
it's
a
very
Telco
show
I'm
gonna
share
my
screen
here.
I'm
I
think
I'll
just
show
you
this
one
slide
and
then
I'll
I'll
show
you
the
what
we're
demoing
we've
got.
A
we've
got
a
live
demo.
B
We,
the
the
group,
came
together.
We've
got
orange
and
Vodafone
and
it
tis
a
lot
and
NBN
code.
These
are
all
national,
Community
or
Global
Communication
service
providers,
so
mobile
phone
operators
and
that
sort
of
thing
the
these
guys
partnered
with
us
and
they've.
They
all
had
a
great
interest
in
in
wasmcloud
and
using
wasm
sorry
webassembly
and
using
Walzem
Cloud
to
to
realize
that
the
our
colleagues
here
at
cosmonic
have
been
working
with
orange
already,
but
we've
introduced
these
other
folks.
B
So
most
people
on
the
floor
here
either
had
an
inkling
of
webassembly
or
or
I
hadn't
heard
of
it
at
all.
So
it's
been
great
fun.
Introducing
everyone
to
that.
So
you
can
see
this
slide.
I've
got
up.
B
Yep
looks
good
okay,
so
this
is
showing
the
the
canvas
as
the
the
before
using
in
the
terminology
here
of
the
applications
that
we're
doing
so.
This
we
have
the
two
blocks:
production
block
and
core
Commerce
blocks.
So
core
Commerce
is
like
that's
the
the
where
we
have
all
the
billing
software
for
for
a
service
provider.
Production
is
where
the
network
is
and
the
the
things
to
that
are
needed
to
run
the
network
in
real
time.
So
this
thing
at
the
bottom
is
a
converged
charging
system.
B
So
if
you
have
a
prepaid
service,
you
need
a
real-time
charging
system
and
that's
what
this
does
so
this
this
block
at
the
bottom
will
need
to
interact
with
these
with
these
other
applications.
So
we've
implemented
all
of
these
blue
blocks
as
as
actors
in
wasm
Cloud.
So
we,
you
know,
wrote
rust
code
and
and
are
orchestrating
those
using
wadam,
and
you
know
it's
quite
an
eye-opener
for
folks,
because
basically,
you
know
you're
easily
looking
at
a
million
dollars
worth
of
software
there.
B
So
people's
idea
about
these
things
is
that
they're
very
large
and
expensive
and
come
from
providers
like
Nokia
and
Erickson
Etc,
and
so
we
have
demonstrated
you
know
minimum
viable
product
proof
of
concept,
but
we
have
demonstrated
running
this
stuff
in
wasm
Cloud.
So
let
me
go
over
here.
B
Share
well,
there's
the
this
is.
Actually
this
is
a
public
public-facing
I
can
share
the
URL.
It
is
behind
a
registration
wall
which
is
a
bit
of
a
pain
in
the
ass,
but
I
thought
they
do
things.
They,
like
everything,
I'll
show
you
the
UI.
So
if
you
can
see
that
there's
our
our
complement
of
actors
running
and
We've
I
I
shared
a
picture
earlier,
we've
got
these
this
little
cluster
of
arm
servers.
We've
got
this
metered
power
that
shows
everyone.
B
There's
only
a
couple
watts
of
power
going
on
here
and
we've
implemented
all
of
those
things
and
orchestrated
them
with
wasm
cloud
and
people
have
been
really
blown
away,
but
by
what
they're
seeing
here.
So
it's
been
fun.
A
Vince,
this
is
awesome.
I
I
I've,
seen
like
little
Snippets
of
you,
like
you,
know,
sending
some
sending
messages
about
what
you're
working
on
I
know
that
we've
talked
about
like
the
the
wadham
Daemon
scaler,
which
is
that
we
are
using
here
to
run
like
the
different
copies
of
actors.
B
Yeah,
that's
right,
so
we
we're
Distributing
10
replicas
of
everything
across
all
the
Linux
hosts.
So
there's
a
there's,
a
Mac
Mini,
that's
inside
the
kiosk.
Here
that's
running
the
providers
and
all
of
the
actors
are
just
being
distributed
across
these
little
tiny
servers.
And
again
you
know
it's
really
people.
Actually
there
was
one
other
thing
I
wanted
to
share.
Let
me
let
me
do
that.
B
I've
got
one
other
slide
and
this
is
really
been
an
eye-opener
for
a
lot
of
people
and
and
that's
really
how
you
build
a
a
large
system
like
this.
B
Once
we've
implemented
everything
in
wasm
Cloud,
putting
it
all
together
becomes
quite
interesting.
So
there's
the
side,
and
let
me
just.
B
Just
play
from
current
slide
and
then
I'll
just
share
that
oops
seems
to
have
moved
I'll
fix
that
in
a
moment.
B
Oops
right
here
so
once
we
realized
all
of
that
all
those
actors
and
I
showed
you
how
they
interact
with
each
other.
Well,
everything
in
yellow
is
all
of
the
the
open
apis
rest
apis
that
we
usually
use
to
integrate
between
these
different
software
components.
But
what
we
did
in
this
project
is.
B
We
highlighted
the
fact
that
that
two
actors
within
the
same
canvas
as
we
call
it
or
wasm
Cloud
instance
having
to
to
interact
using
rest,
is
really
awkward
and
unnecessary,
and
so
we
took
one
of
these
rest
open
apis.
The
Service
inventory
API
and
we
reverse
engineered
it
to
create
a
service
model.
B
We
we
implemented
a
contract
in
Smithy,
they
did
all
the
same
things
and
and
then
we,
you
know,
used
the
SDK
to
create
an
interface
for
that
and
we
created
another
another
actor
which
implements
the
actor
doctor
interface
instead
of
the
the
normal
Tim
form
standard
interface,
and
so
we
were
able
to
take
all
the
yellow
and
replace
it
with
the
green
and
and
what
we
say
to
people
is
everything
in
yellow
represents,
represents
carbon
emissions,
energy
cost
latency
and
Technical
complexity.
B
And
if
we
have
a
lot
of
of
implementation
of
of
software
systems
that
communicate
with
each
other,
we
can't
ignore
that
cost
and
and
our
partners
agree
and
they
and
and
they
want
to
see
that
eliminated.
And
so
we've
demonstrated
how
that
can
be
done,
which
has
added
a
perspective
to
folks
who
really
only
saw
kubernetes,
and
they
really
only
saw
open
apis
as
as
the
way
to
do,
integration
and
deployment.
And-
and
so
we've
we've
opened
eyes
for
a
lot
of
people.
A
Yeah
vansight
I
really
like
that.
That
last
slide
that
you
showed
like
talking
about
the
unnecessary
hop
that
you
could
take
over
HTTP
and
and
kind
of
how
you're
solving
that,
which
is
really
cool,
I
would
love
it.
I,
don't
know
if
you
have
any
of
this
in
a
public
publicly
accessible
place
or
if
you
can
share,
but
I
would
love
to
take
a
look
at
at
your
slides
and
also
the
the
demo.
A
If
you
have
an
open
GitHub
repo
for
that,
but
only
if
that's
something
that
is
fine
for
you
to
share
and
you
have
in
the
open.
B
Yeah
I
can
I'll
I'll
drop
the
URL
in
it's
behind
a
registration
wall,
but
I
can
find
another
way
to
share
it
as
well.
But
all
the
source
code
and
all
the
slides
and
all
the
the
stuff
is
is
available
publicly
if
a
little
bit
awkwardly.
A
Okay,
yeah
well
I
hope
you're
not
getting
kicked
out.
Quite
yet.
Does
anybody
have
any
questions
for
Vance,
okay,.
A
All
right,
well,
I
think
we
can
go
ahead
and
move
on
then
thank
you
so
much
Vance
for
for
coming
in
and
talking
a
little
bit
about
what
you're,
showing
off
and
and
doing
a
demo
I
hope
the
is
the
conference.
This
was
the
end
of
the
conference
for
today.
Is
it
still
going
like
through
the
end
of
the
week
or
is
it?
Is
it
done
now.
A
Awesome
yeah,
please
please
post
in
the
awesome,
Cloud
slack
or
let
us
know
when
you
bring
home
that
award
super
super
exciting
good
luck
at
the
rest
of
the
conference.
A
All
right
folks!
Well
as
we
look
forward,
the
next
thing
on
our
agenda
for
today,
like
I
said,
is
another
awesome.
Demo
Kevin
is
here
and
going
to
be
demoing
what
what
he
coined?
Building
Nats
microservices
in
wasm,
Cloud
I'm
super
excited
to
see
this
one
and
so
we'll
go
through
the
demo
and
then
get
on
to
discussion.
So
Kevin
you
should
have
co-host.
Hopefully,
so
you
should
be
free
to
go.
D
New,
can
you
say
my
terminal
window
now.
A
D
A
Yeah
yeah
I
can
see
that
part
I
just
want
to
make
sure
you
weren't
just
trying
to
do
that.
D
All
right,
so
a
little
bit
of
background,
one
of
the
main
reasons
why
people,
one
of
the
many
reasons
why
people
like
using
mats
is
you
know.
Historically,
we
just
say
that
you
don't
need
service
Discovery
anymore.
No,
that
one
you
know
once
you're
on
NAS,
you
don't
need
service,
Discovery
and
that's
sort
of
true,
but
I.
D
Think
when
we,
when
we
talk
about
not
needing
service
Discovery,
most
of
the
time
we're
talking
about
not
needing
to
know
where
things
are
so
location,
transparency
or
you
know,
location
agnosticism,
but
we
don't
actually
have
facilities
within
Nets
or
things
like
service
Discovery.
Until
you
know
now,
yeah
over
the
past,
I'd
say
I
guess
a
week
or
two
well,
maybe
a
little
bit
longer,
some
of
the
clients
have
been
rolling
out.
D
Some
of
the
Nets
clients
have
been
rolling
out
with
a
new
feature
that
allows
you
to
expose
your
your
code.
That
is
following
the
request
response
pattern
as
a
service,
and
then
you
can
use
the
Nats
client
or
the
net
CLI
to
discover
and
get
information
on
those
services,
so
I'm
using
a
local
build
here,
because
I
think
this
CLI
version
that
has
the
the
micro
function
in
it
I
think
it
was
released
today.
D
D
A
couple
of
you
can
add
arbitrary
metadata
into
the
services
that
you
run.
So
these
are
added
by
Nats
and
then
each
service
has
one
or
more
endpoints,
and
so
this
particular
endpoint
is
called
Echo
and
the
subject
is
Bob
dot,
Echo
and
in
addition
to
all
of
the
the
regular
high-level
metadata,
it's
also
keeping
track
of
the
number
of
requests
processing
time
when
the
service
was
started,
the
number
of
errors.
D
So
if
I
do,
there's
micro,
I,
get
stats
on
the
Bob
service.
You'll
see
that
the
Bob
service
with
this
ID
and
this
endpoint
is
at
five
requests
and
it's
keeping
track
of
you
know
the
microseconds
for
the
time.
So,
if
I
were
to
do.
D
So
I
sent
a
message
to
bob.echo
and
now,
when
I,
do
the
stats
on
Bob
you'll
see
that
the
requests
have
gone
up
so
what's
interesting
about
this
is
again
this
allows
me
to
discover
microservices
and
get
like
a
service
catalog
of
stuff.
That's
running
on
my
enhance
infrastructure
or
in
my
Nets
account,
but
it
doesn't
sacrifice
any
of
the
the
good
things
that
are
that
we
all
take
for
granted
for
from
that.
So
it's
still
completely
location
agnostic.
D
When
you
get
the
information
on
these
things,
when
you
get
the
information
on
it,
there's
nothing
here
that
says
you
know
this
is
on
the
inside
or
the
outside
of
a
leaf
node
or
what
account
it's
from
or
anything
else.
It's
just
the
service
is
there
and
if
I
use
this
subject,
I
can
access
it.
D
So
obviously
this
is
wasm
cloud
Wednesday.
So
one
thing
that
I
want
to
show
is
you
can
list
all
of
the
available
services
or
that
are
you
know
within
that?
Can
that
are
within
communication
within
your
Nats
CLI,
and
so
we
saw
the
net
CLI
demo,
which
I
created
it's
called
Bob,
but
we
also
have
a
while
Some
Cloud
actor
running
using
nothing
but
the
wasm
Cloud
messaging
contract
that
has
advertised
itself
as
a
service.
D
So
I
can
now
get
info
on
a
calculator
and
you'll
see
calculator
has
endpoints,
add
subtract
and
multiply,
and
obviously
I'm
I'm
faking
some
of
the
data
here,
but
you
can
sort
of
get
the
idea.
It's
the
same.
It's
the
same
stuff
that
you
that
we
got
with
Bob,
but
I'm
just
faking
this
a
little
bit
so
right
now,
I'm
doing
this
manually
inside
an
actor.
D
But
my
plan
is
to
just
add
an
option
to
the
Nets
messaging
provider,
where
you
just
you,
can
supply
your
service
advertising
information
in
the
link,
Dev
and
it
just
magically
appears
as
a
Nats
service.
That's
the
plan.
I've
got
a
little
bit
of
design
work
to
do
on
it,
but
otherwise
that's
that's
pretty
much.
It
I
mean
being
able
to
have
this
kind
of
thing
and
have
you
know
first
class
service
catalog
to
support
inside
Nets?
For
essentially
no
cost
is
it's
pretty
fantastic
and
it
took
takes
very
little.
D
Effort
to
expose
was
and
Cloud
actors
as
Services
as
well.
No,
not
that
much
of
a
demo
but
I
wanted
to
to
show
it
off
a
little
bit.
Yeah
Brooks.
A
How
does
Nance
here,
I
guess,
first
of
all
discover
the
actor
as
a
as
a
service,
and
then
I
saw
you
have
kind
of
discrete
whether
it's
endpoints
or
operations
like
add,
multiply
subtract
that
kind
of
stuff.
What's
the
yeah
is
that
metadata
on
the
request?
Is
that
something
that
you
have
to
specifically
design
for?
How
does
that
work?.
D
D
Replied
to
that,
and
so
that's
how
it
got
the
list
here
and
if
I,
if
I
do
calculate
your
info
and
I
do
Trace
you'll
see
it
did
info
on
calculator
and
got
one
set
of
data
than
it
did
stats,
not
calculator
Dot,
this
ID.
So
what
is
implicit
here
that
is
also
super
powerful.
D
In
here,
each
running
instance
of
this
would
have
its
own
unique,
ID
and
so
I
can
then
see
that
I
have
you
know
50
instances
of
the
calculator
service
running
and
it
can.
It
will
then
keep
track
of
the
the
usage
and
error
rate,
and
you
know,
processing
time
on
the
the
service
as
a
whole
or
the
service
with
this
specific
ID.
So.
D
It
looks
exactly
the
same
as
the
last
time,
because
that's
the
only
instance
of
calculator
running.
But
if
there
were
you
know,
10
different
instances
of
calculator
running
I
could
select
them
by
their
inique
ID
and
get
individual
stats.
And
if
I,
just
grabbed
info
on
calculator,
I
would
get
aggregate
stats
from
everything
that
replies.
D
Yeah
Dan,
what's
in
here,
what
I'm,
showing
is
everything
in
adr32
yeah
on
the
protocol
is
defined
there,
but
also
good
news
is
that
if
you're
using
the
go
or
rust
or
Java
or
any
of
the
major
clients,
you
already
have
the
ability
to
explicitly
publish
service
information.
D
So
if
you,
if
you're
using
rust,
the
Russ,
client,
async
Nats
version
0.31
or
later-
and
you
enable
the
service
feature
flag,
you
get
this
stuff
automatically
I
mean
you
still
have
to.
You
know
write
code
to
publish
your
service
metadata,
but
it
will
do
things
like
keep
track
of
stats
on
that
particular
subject.
A
Thinking
about
this
and
then
maybe
Turtles
all
the
way
down
way
if
I
was
writing.
I
know
this
kind
of
goes
against
the
contract
idea
for
for
writing,
model
actors
and
and
kind
of
tying
it
specifically
to
Nats.
Would
you
have?
D
I,
wouldn't
do
it
outside
of
the
awesome
Cloud
app
I
mean
you,
you
sort
of
could
but
yeah
for
a
couple
of
reasons.
It's
probably
not
the
best
idea
like
I,
said
I
think
what
I
want
to
try
doing
is
putting
this
as
a
feature
in
the
Nets
messaging
capability
provider.
D
And
if
your
actor
is
exposing
a
request
response
pattern
on
one
or
more
endpoints,
then
you
know
you
should
be
able
to
supply
some
metadata
in
that
link
definition.
And
then
the
the
capability
provider
will
just
automatically
take
care
of
publishing
your
stuff
as
a
service
and
it'll
keep
track
of
the
stats
for
you
and
all
that
stuff.
A
That
makes
perfect
sense.
Okay,
yeah
I
was
kind
of
looking
for,
like
like
the
right
integration
point.
I
think
at
the
provider
level
like
on
a
link,
diff
makes
sense.
E
F
It
kind
of
dovetails
with
our
next
topic,
which
is
talking
about
metrics,
which
is
a
bit
of
a
spoiler,
but
because
now
it's
microservices
seem
like
they're
keeping
things
like
per
request:
statistics,
I
guess,
as
somebody
who's
interested
in
operating
wasn't
Cloud
at
a
pretty
decent
scale.
If
we
end
up
or
think
about
kind
of
like
especially
on
the
Nots
messaging
provider
in
other
places,
kind
of
using
some
of
this
microservices
metadata
is
kind
of
more
of
a
first
class
like
it's
just
a
feature
of
Wallace
and
Cloud.
Do
we
like?
F
Would
it
make
sense
to
sort
of
pull
that
metadata
or
like
those
statistics
actually
from
like
from
Nats
itself,
like
I,
know,
I
think
surveyor
doesn't
have
microstats
probably
should
I
think
there's
actually
a
issue
about
it,
or
would
it
be
better
to
like
integrate
actually
to
deeper
Watson
Cloud
level.
D
I
think
it's
really
I
think
that
comes
down
to
preferences,
so
the
the
net
servers
are
not
keeping
track
of
the
stats
you
know.
So,
when
I
make
this
request
on,
you
know:
service.ping,
it's
just
a
straight
scatter
gather
and
the
services
that
are
running
will
respond
to
it,
and
the
services
that
are
running
are
the
ones
that
are
responsible
for
keeping
their
their
usage
stats.
D
Both
you
know
for
now,
given
that
this
feature,
at
least
in
the
CLI,
is
a
day
old.
D
F
Makes
sense,
I
think
it
was
more
just
I,
guess,
I
didn't
realize
offhand
that
the
server
side
wasn't
actually
keeping
track
of
these
statistics
is
actually
up
to
the
individual
service,
so
that
makes
sense
I
wonder
if
down
the
line,
I
mean
I
think
we
need
to
figure
out
and
again
we'll
discuss
more.
F
What
some
of
the
metrics
data
even
looks
like
in
the
first
place,
but
I
do
wonder
if
at
some
point
you
know
once
this
features
had
a
little
bit
of
time
to
bake
and
we
we've
had
some
time
to
think
about
it
if
it
would
make
sense
to
actually
integrate
that,
like
at
a
deeper
level
within
was
Some
Cloud.
Just
have
that
like
be
part
of
the
provider
SDK,
for
example,
I'm
not
sure.
D
It
would
be
yeah,
I
mean,
theoretically,
you
could
do
something
where
every
actor
that
sits
on
a
lattice
could
also
expose
itself
as
a
service,
but
not
all
actors
use
Nas
messaging,
which
means
they,
wouldn't
it
wouldn't
make
any
sense
for
them
to
appear
as
a
Nats
microservice
yeah,
like
you,
said,
the
the
NAT
servers
are
completely
unaware
of
any
of
this
stuff.
The
net
Server
doesn't
know
what
a
microservice
is.
The
only
thing
that
knows
is
the
Nats
client.
A
Well,
I
feel
like
this
is
a
good
well
before
I
before
I
segue,
any
other
questions.
G
C
Oh
I
was
just
gonna
say
no
questions,
but
thanks
for
sharing
Kevin,
that
thing
is
awesome.
Yeah
I
think
I
can
really
see
a
future
there.
That
is,
that
is
very
nice.
I
know,
I
was
playing
with
Nat's
a
bit
and
a
little
bit
of
wasmcloud
and
going
gosh.
I
really
would
like
to
be
able
to
publish
some
sort
of
extra
discovery.
On
top
of
you
know
just
the
subjects,
and
that
was
it
just
fits
right
in
there.
So
I'm,
looking
forward
to
playing
with
that.
G
D
Can't
inject
metadata
into
anything
and
exposing
your
code
as
a
service
is
something
that
your
code
has
to
do
explicitly
using.
D
You
know
either
just
sort
of
manually
subscribing
to
the
service
topics,
or
you
know,
using
one
of
the
apis
that's
available
in
a
language,
specific
SDK,
but
the
usage
pattern
for
the
services
described
in
this
is
at
least
at
the
moment.
Currently,
it's
just
request
reply.
D
G
A
A
Let
me
go
ahead
and
bring
up
the
original
issue
here,
just
so
that
we
can
actually
get
wow.
Do
we
close
it
just
so
that
we
can
get
the
get
the
right
context?
A
A
There,
it
is
okay,
just
didn't
know
the
title,
so
Steve
filed
an
issue
the
other
week.
Thank
you
so
much
Steve
talking
about
RPC
invocations.
We
have
an
ADR
which
you
can
actually
go
look
at
which
talks
about
separating.
Oh,
maybe
that's
just
the
the
wrong
link
separating
out
RPC
events
onto
a
different
topic
compared
to
our
regular
Cloud
events,
and
that's
just
because
they
were
a
little
bit
different
and
in
the
refactor
to
the
rest
host.
A
We
don't
have
the
wasmbus.rpc
EVT
events
getting
published
for
every
invocation
that
succeeds
or
fails
in
the
system.
This
was
really
nice
beforehand
to
be
able
to
subscribe
to
the
RPC
EVT
topic.
Whenever
you're
doing
debugging
or
watching
things,
you
can
see
the
you
can
see
the
status
of
all
the
invocations
to
flow
through
a
system,
but
really
the
the
main
point
of
these
events
were
to
be
able
to
create
kind
of
an
aggregate
metric.
A
You
can
watch
the
number
of
invocations
succeeded
or
invocation
failed
over
time
and
get
an
idea
for
if
your
service,
the
thing
that
you're
running
in
wasmcloud
is,
is
succeeding
in
what
it's
doing
you
can
create
a
you
know,
essentially
a
a
percentage
of
successful
requests.
This
is
kind
of
like
what
a
lot
of
service
mesh
products
offer
and
we're
kind
of
doing
it.
We
were
doing
it
straight
up
over
Nats
and
over
the
course
of
the
discussion
which
I'll
go
ahead
and
send
in
the
chat
here
and
I.
A
Think
us,
like
Steve
and
us
on
the
maintainer
side,
kind
of
got
to
the
point
where
we're
thinking
that
really
this,
it
feels
a
lot
more
like
a
useful
metric
than
it
needs
to
be
a
discrete
Cloud
event
per
se
and
while
we're
thinking
about
one
metric,
we
might
as
well
think
of
other
things.
That
would
be
great
to
instrument
with
metrics
in
the
host
and
I.
A
Don't
think
Patrick
is
actually
here
today,
but
thank
you
so
much
Patrick
for
filing
this
RFC
over
about
hostmetrics
really
just
talking
about
implementing
Prometheus
metrics,
and
you
know,
there's
some
various
design
aspects
in
here
that
I
won't
go
too
deep
on
I'll.
Just
try
to
send
this
out
now.
Dan
I
know
that
you
have
talked
about
implementing
Prometheus
metrics
and
in
wasmcloud
before
I.
Think
it's
been
on
your
your
wish
list
for
a
little
while
as
a
infrastructure
person
who
runs
wasm
Cloud.
A
Would
you
be
interested
in
talking
a
little
bit
more
about?
Maybe
what
you
think
about
this
RFC
and
what
you
think
would
be
useful
as
a
Prometheus
metric.
F
Yeah
sure
I
think
there's
a
couple
of
different
things
that
sort
of
are
wrapped
up
in
this
RFC.
So
we
do
already
have
Hotel
support
right
for
tracing,
which
is
great,
I,
don't
know
how
many
people
on
this
call
have
tried
that
out,
but
like
it's
really
really
nice
to
be
able
to
get
that
data
I
think
some
of
the
things
that
are
missing
from
that
are
kind
of
more
to
an
aggregate
level,
especially
now
that,
or
rather
for
historical
comparison
right
in
the
old
washboard
and
The
Elixir
host.
F
We
used
to
have
like
a
slash
metrics
endpoint,
which
gave
you
some
decent
summary
statistics,
but
now
that
that's
sort
of
not
really
a
thing
anymore
and
we've
transitioned
to
the
rest
house,
like
we
lost
some
of
that
introspection,
and
so
things
that
are
useful
are
kind
of
more
just
like
at
a
higher
level.
Things
information
about
a
particular
house
like
how
many
actors
is
it
running?
F
How
many
providers
like
what
are
how
many
invocation
has
a
particular
actor
or
provider
had
because
once
you
have
that
you
can
start
calculating
some
interesting
summary
statistics,
particularly
around,
like
your
usage
rates
of
individual
actors,
like
do
you
need
to
scale
up?
Do
you
need
to
scale
down
like
it
gives
you
a
little
bit
better
information
there?
F
So
that's
kind
of
at
a
host
level
out
of
I
think
the
follow-on
on
here
is
also
at
a
provider
level
like
it
could
be
really
great
at
some
point
for
individual
providers
to
be
able
to
expose
metrics
and
stuff.
You
know,
for
example,
HP
server,
HTTP,
client
right
it'd
be
great
to
have.
F
You
know
the
counter
for
the
number
of
500s
or
400s
or
whatever
that's
pretty
common
in
pretty
much
any
other
ecosystem
out
there
and
I
think
we
want
to
be
able
to
do
the
same
thing
in
wasm,
Cloud
I
think
there's
a
lot
of
implementation
to
sort
of
figure
out,
especially
with
how
does
the
host
grab
those
metrics
from
Individual
providers?
Do
we
surface
them
purely
through
otel?
Do
we
just
have
you
know
a
slash,
metrics
endpoint
on
a
particular
Port
like
we
normally
do?
F
Do
we
support
both
should
go
over
Nats
I.
Think
the
where
Patrick
in
particular,
was
leaning
towards
is
probably
just
having
that
slash
metrics
endpoint,
but
if
anybody
has
opinions
on
how
to
do
observability
in
metrics,
like
Now's,
the
Time
that
RFC
is
open,
you
know,
feedback
is
definitely
more
than
welcome.
F
I
know
again,
like
Brooks
mentioned,
someone
who
is
operating
was
on
cloud
in
production
every
day,
like
I'm
really
excited
to
actually
start
getting.
Some
of
this
data
in
place.
A
Yeah,
this
is
just
to
invite
everybody
explicitly
definitely
open
floor
time
for
for
discussion.
Questions
requests
all
that
stuff.
G
A
You're
just
you're
a
little
a
little
quiet
on
your
end.
Sorry,
if
you're
I
don't
want
to
really,
you
know,
make
you
well
that's!
Okay,
that's
better!
Now,
actually
that
that
really
that
you
just
said.
G
Would
you
mind
repeating
your
yeah?
No,
the
like
the
open,
Telemetry
semantic
conversions.
They
they
have
a
lot
of
predefined
keys,
you're
supposed
to
hook
into
it
to
have
a
particular
meaning.
So
a
lot
of
tools
are
built
around
that
nowadays.
G
F
Yeah
I
think
I
like
to
or
I
just
I
mean
it's
kind
of
just
the
overall
event
model
right,
I.
Think,
since
we're
already
doing
hotel,
we
may
as
well
keep
going
with
that.
We've
already
made
that
decision
right
for
tracing
there's,
no
reason
not
to
do
it
for
metrics
at
the
moment,
so
yeah
I
think
is
I
would
prefer
to
see
us
follow.
You
know
Community
practices
and
the
specs
as
much
as
possible.
I
think
we
just
need
to
make
that
clear.
A
Yeah
I
think
that
that's
that's
definitely
the
goal
here,
as
as
we
think
of
was
on
cloud
as
something
that's
in
the
the
cloud
native
landscape
and
you
know
going
to
be
used
by
people
who
are
building
native
apps
now,
whether
it's
a
totally
new
use
case,
they're,
trying
to
use
webassembly
or
just
trying
to
you
know,
take
something
that
we're
they're
currently
doing
in
kubernetes,
for
example,
and
doing
it
in
in
wasmo,
Cloud
I
think
our
use
of
cloud
events
and
Nats
and
otlp-
and
you
know
oci
all
those
things
is-
is
all
under
the
goal
of
using
the
open
standard
and
Community
standard
I.
A
A
Great
well
I
have
I,
did
drop
the
RFC
into
the
community
meeting
chat
here
and
on
the
YouTube
live
stream.
So
please
feel
free
to
go
and
you
do
not
have
to
be
a
wazzle
god
contributor
or
a
member
of
the
orc
to
comment
on
that
issue.
A
A
That's
still
proposed
we're
not
planning
on
taking
any
actions
yet
until
we
go
through
kind
of
the
RFC
process
which
we
have
outlined
in
one
of
our
more
recent
adrs,
but
please
feel
free
to
leave
feedback
on
this.
One
I'm
looking
forward
to
hear
what
you
all
think
and
then
think
it's
going
to
be
pretty
awesome
to
add
metric
support
and
soon
to
come
version
of
of
wasmcloud,
so
very
fun,
quick
last
check
if
anybody
has
any
questions
or
comments.
Otherwise,
we'll
move
on
to
the
last
thing.
A
A
We've
included
binaries
on
the
release
and
then
also,
of
course,
published
some
oci
images
that
you
can
pull
down
and
run
yourself
and
we've
gone
through
like
one
or
two
RCS,
as
we
just
wrap
up
this
release
and
get
it,
you
know
fully
ready
for
people
to
use
and
fix
some
last
bugs
and
all
things
like
that,
but,
as
we've
been
going
through,
we
are
essentially
hitting
the
essentially
right
now
we
are
feature
Frozen
for
future
things
going
into
the
0.78
release.
A
There's
always
going
to
be
more
work
to
do,
of
course,
but
we're
pretty
happy
with
where
this
host
is
now,
after
really
getting
a
chance
to
really
getting
a
chance
to
test
it
out
with
our
existing
workflows,
and
things
like
that.
So
this
is
really
just
a
call
out
for
everybody
to
keep
an
eye
out
either
today
or
tomorrow,
for
an
official
0.78
release
and
then
from
there
we'll
be
able
to
get
into
actually
milestoning
out
future
features
for
wasm
cloud.
We're
going
to
be
following
a
strict.
A
You
know
the
the
rust
December
the
thing
same
thing
that
we
do
in
all
of
our
other
projects
in
the
Watson
Cloud
ecosystem,
but
just
wanted
to
note
that
that
release
is
coming
pretty
soon
and
we're
going
to
be
able
to
get
that
out
into
wash
as
the
regular
version
that
runs
when
you
run
wash
up
as
well.
As
you
know,
we'll
probably
have
an
Associated,
Autumn
or
wood
dam
release
to
go
with
it
as
well.
That
just
takes
advantage
of
all
of
the
new
things
that
we've
been
putting
into
wasmcloud.
A
So
this
is
really
sweet.
I'm,
really
looking
forward
to
I'm.
Looking
forward
to
showing
you
all
the
posts
that
we've
got
together
for
the
0.78
release,
because
it
kind
of
touches
on
you
know,
we've
talked
about
it
in
a
community
call
but
wiser
at
odd
78.
You
know
what
we're
doing
with
versioning
going
forward,
especially
thinking
of
wasm
cloud
as
a
proposed
incubating
project
in
the
cncf
and
something
that's
working
to
a
1.0
release
as
we
reach
some
major
milestones
in
the
webassembly
community
and
just
generally
operating
as
a
project
for
the
last
three
years.
A
So
there
is
there's,
there's
not
really
any
explicit
callouts
for
version
78
other
than
if
you've
got
an
application,
would
love
it.
If
you
tried
it
out,
it
is
a
drop-in
replacement
for
version
63,
which
is
the
Elixir
OTP
host
of
wasmcloud.
So
everything
should
work
exactly
the
same
and
that's
how
it's
been
for
the
last
couple
of
RCs
that
we've
tested,
but
always
happy
to
take,
take
more
feedback
all
right,
any
any
questions
there
on
on
that
release.
It's
really
just
a
just
a
general
call
out.
A
Awesome
well,
I
think
that
that
is
kind
of
the
end
of
the
community
meeting
at
least
the
scheduled
agenda
for
today.
A
So
since
we've
got
a
little
bit
of
time
left
in
the
call
now
is
a
great
great
time
for
just
general
q,
a
things
about
the
broader
webassembly
Community
things
like
that,
so
I'm,
just
gonna
go
on
mute
for
a
second
and
anybody's.
Welcome
to
take
the
floor.
E
I
would
just
call
out
that
the
wasem
day
as
schedule's
been
released,
there's
still
a
couple
talks
that
are
moving
around
for
folks
that
could
or
cannot
make
to
Chicago
wasmocon.
Obviously
we
talked
about
that
last
week.
E
Everything
is
still
live
up
on
YouTube
on
the
Linux
Foundation
YouTube
channel,
and
we
still
have
some
work
to
do
to
process
some
of
the
video
recordings
from
bacon
the
pi
code
Alliance
conference,
but
I'd
like
to
pull
out
at
least
demo
the
demo
hour
and
a
few
of
the
talks
and
get
those
on
there
to
go
from
there.
Those
are
the
big
Community
things
that
are
happening.
E
H
A
H
I
The
at
the
moment,
it's
it's
for
cm4s
on
the
touring
pie
to
that
little
four
cluster
board,
but
yeah,
two
of
them,
two
of
them,
have
a
gigs
in
memory.
Two
of
them
have
two:
all
of
them
are
working
really
well
and
I.
Have
three
of
the
nodes
are
running
wasmclow
one
of
the
nodes
or
is
running
with
them,
and
yeah
I
mean
I'm,
really
I'm,
really
having
no
issues
with
it
at
the
moment.
To
be
honest,.
I
Here
in
about
45
minutes,
Google
Coral
will
be
arriving
at
my
door
and
that
has
a
spot
on
the
board
as
well.
So
we'll
see
how
that
works.
H
I
was
gonna,
say,
I,
don't
know
how
much
you'd
get
for
just
you
know
just
regular
computer.
You
can
use
along
with
the
with
the
yeah
stuff
on
the
coral,
but
that's
super
interesting.
I
Allegedly,
it
has
a
its
own
little
slot
on
the
host
board
that
connects
directly
to
one
of
the
nodes,
so
fingers
crossed.
A
I
think
I
actually
have
a
Dev
board.
I,
don't
know,
I
think
it
was
a
they
might
have
like
older
versions.
I'm,
not
sure
it
was
at
least
a
few
years
ago,
but
the
I
have
a
Dev
board
and
then
an
old
or
a
coral,
and
then
an
old
Nvidia
Jetson
and
then
a
Raspberry
Pi
and
like
I,
always
wanted
to
get
like
a
little
rack
to
like
slot
them
all
into
and
and
connect
them
into
a
lattice
I.
C
But
the.
J
The
corals
themselves
are
really
good
for,
like
open,
CBE
workloads,
the
instructions
that
their
chips
kind
of
come
for
and
the
amount
of
memory
they
have
make
them
good
for
like
streaming
data.
So
the
latest
and
greatest
there
is
stuff
with
like
liquid
neurons,
but
it's
like
image
segmentation,
opencv
jobs
like
object,
recognition.
Those
are
all
out
of
the
box,
things
that
you
should
be
able
to
get
running
pretty
quickly.
H
Yeah
another
use
case
actually
that
I
think
a
friend
gratitude
was,
you
know
he
just
tried
to
manage
his
like
at-home
cameras
and
stuff,
so
I,
just
like
that's
like
a
good
camera
feed
to
run
opencv
on
or
like
if
you
have
like,
like
an
outside
sort
of
camera
or
stuff
like
that.
That
sounds
like
would
be
pretty
good
with
that.
In
this
case,.
J
A
I
consider
doing
like
a
shoot
is
that
still
an
RFC
I
know
that
there
was
a
proposed
sensors
interface
that
Kevin
came
up
with
a
few
months
ago,
where,
like
it,
would
be
cool
to
just
get
kind
of
constant
readings
about
air
quality
and
then
have
it
do
something
like
send
you
a
text
message
or
something.
I
was
specifically
wondering
about
the
difference
between
air
quality
inside
of
the
house
versus
outside,
and
that
would
be
I.
Think
that'd
be
a
pretty
cool.
A
I
was
pretty
sure
that
that
was
how
we
were
actually
displaying
something
on
this
little
or
how
we
were
interpolating,
like
the
input
from
the
little
buttons
on
this
little
Raspberry
Pi,
but
I'll
have
to
double
check,
I
think
that
was
Conor
or
Taylor.
That
wrote
some
of
that
stuff
wave
share.
That's
what
it
is.
Okay,
so
it's
the
OLED
provider,
but
it
does
that
using
gpio.
E
I
was
going
to
mention
that
there's
some
work
going
on
in
and
for
a
proposal
for
YZ
sensors,
which
should
be
interesting.
You
know,
Jordan
I,
know
that
you
have
all
of
this.
Like
you
know
these
couple:
demos
like
your
ITC
and
the
the
LED
one
they're
all
written
in,
go
I,
believe
and
now
that
we're
on
the
rust,
toast,
I'm
wondering
and
we
can
load
components.
The
Wazi
components,
I'm
wondering
what
the
what
the
possibility
of
of
getting
those
running
is.
E
If
they,
if
they
work,
I'd
love
to
get
them
into
the
examples,
repo
would
be
I
think
the
ideal
place
for
them
to
go.
I
A
Folks,
I
think
we're
we're
hitting
right
at
the
end
of
time,
so
to
avoid
going
over
I
think
I'll,
probably
we'll
just
say
we
can
call
it
here
today.
Thank
you
so
much
I
guess
Vance
is
off
the
call
now,
but
Vance
and
Kevin
for
doing
our
awesome
demos
at
the
beginning
of
the
call
and
everybody
for
participating
in
the
metrics
in
0.78
discussion
and
then,
of
course,
getting
to
hang
out
a
little
bit
and
talk
about
a
little
tiny,
embedded
devices.
A
It's
really
awesome
to
see
was
on
cloud
working
on
that
so
I
hope
we
can
just
turn
this
discussion
into
a
really
fruitful
demo,
hint
hint
nudge,
nudge,
Jordan
or
anyone
who's
got
the
little
devices,
because
it's
such
a
cool
use
case
for
wasm
and
I
will
get
off
my
soapbox
before
I
get
started.
So
thanks.
Everybody
definitely
looking
forward
to
seeing
you
next
week
and
I'll
talk
to
you
later.