►
From YouTube: 2023-03-14 meeting
Description
Instrumentation: Messaging
B
B
C
Yeah
I
haven't
had
a
chance
to
do
that.
Yet
I
was
working
on
this
hand,
box
which
I'm
just
putting
a
details
in
so
I
think
it's
it's
all
up
to
date
and
merging
it's
up
to
date.
As
of
five
days
ago,
I
have
some
other
PRS
out
to
update
to
like
yesterday,
which
then
once
that
one's
merged
I
can
do
a
another
one,
so
it'll
be
current.
C
I
only
ask
is
I've
got
a
bunch
of
components
in
there.
What
other
components
do
people
want
to
see?
C
Now,
if
you
go
straight
to
the
sandbox
and
have
a
look
at
the
the
main
branch
now,
you'll
actually
see
them
under
pkgs
for
packages
that
I've
converted.
Not
everything
is
there,
but
if
you
don't
want
to,
let
me
know
what
you
want.
We
can
add
them,
I'm,
pretty
much
ready
to
create
a
branch
for
the
minification
stuff
based
on
that
and
I
guess,
from
what
we
talked
about
last
week
with
the
minimum
minimal
web
SDK,
there's
probably
enough
there
for
that
too,
to
start
work
on.
A
I
think
the
minimum
packages
we
need
are
those
few
instrumentations
for
the
web.
C
A
B
A
But,
but
what
what
does
it
look
like,
as
in
so
the
the
pack?
What
what
is
the
sandbox
you
know
give
us
like?
Does
it
give
us
a
package
bundle
at
the
end
that
we
could
run
as
a
as
a
agent
that
can
be
embedded
in
the
web
apps.
C
Yeah
it
it
doesn't
publish
them
but
yeah.
If
you
basically
clone
it,
do
an
npm,
install
npm,
run,
compile
or
Rush
rebuild,
which
is
what
uncompile
does
that
will
effectively
rebuild?
C
Actually,
the
compile
won't
rebuild
the
runt
if
you
rush,
rebuild
it'll,
compile
and
run
all
the
tests.
Otherwise,
if
you
do
npm
run,
compile
npm
run
tests
that
run
the
tests
as
part
of
the
compile,
as
well
as
the
build
folder.
It
also
generates
and
do
I
have
it
here,
a
bundle,
folder
I
think
I
called
it.
Was
it
I,
don't
have
it
built
here?
Okay,.
C
And
where's
the
roll
up,
let's
go
into
the
roll-up
it'll,
tell
us
where
it
puts
it.
C
Yeah
it
effectively
creates
every
single
instance
possible
that
you
can
think
of
when
it's
Gonna
Roll
Up.
A
And
and
how
do
we
want
to
take
up
the
discussion
with
Danielle,
based
on
this,
to.
C
Answer
I
was
yeah,
I
was
going
to
mention
it
tomorrow,
so
when
I
say
that
the
sandbox
is
open
for
any
branches,
people
want
to
create
and
we're
also
thinking
about
creating
a
branch
from
you
know,
creating
a
minimal
web
SDK
which
only
relies
on
the
API
and
the
minimum
and
we'll
see
how
that
goes.
A
D
Great
great
news,
I
I
was
talking
to
abernath
yesterday
and
I
went
I'm
wondering
if
no
I
told
him
to
to
create
prototypes
of
the
of
the
instrumentations
in
the
contribute.
Repo,
but
I,
know
I
wonder
if
you
should
do
it
here
in
the
in
the
sandbox.
C
Yeah
we
can
go
to
branch
that
because
I
I
don't
want
any
check-ins
apart
from
the
auto
scripts,
to
go
into
main.
Otherwise
it's
going
to
screw
up
the
auto
merging
scripts,
so
yeah
we
can
create
a
branch.
In
fact,
you
should
be
able
to
create
a
branch
as
well
Martin,
so
you're.
C
C
Okay
and
yeah,
it
drops
it
into
so
in
the
build
folder
you'll
end
up
with
all
your
other
artifacts.
You
do
normally,
but
there'll
also
be
a
bundle
folder
and
in
there
you'll
have
loads.
You'll
have
every
format
known
demand
in
ES
five
and
six
ready
to
be
loaded
into
a
browser
directly.
C
Yep,
so
yeah,
so
if
anyone
has
a
problem
actually
doing
the
compiler
test,
yeah,
just
let
me
know
so,
you
shouldn't
do
it
should
work,
because
all
the
CI
tests
are
working.
So,
unlike
the
issues
Martin
had
with
running
the
scripts
locally,
that
shouldn't
be
an
issue
to
compiling
this.
The
sandbox.
C
The
things
that
are
missing
it
doesn't
have
any
of
the
examples
at
this
point
and
I'm,
probably
not
going
to
get
to
that
this
month,
maybe
late
next
month.
In
theory,
they
should
work.
It
should
just
be
a
case.
We
update
the
config.ts
and
they
should
come
over
touch
wood,
but
I
haven't
tried
it.
So
now,
I'd
only
picked
the
major
ones
like
at
the
moment.
We
have
the
the
exporter
base,
but
we
don't
have
any
exporters,
we
don't
have
the
otlp
or
the
crpc
brought
over
again.
C
They
should
just
be
a
case
of
we
update
the
config
TS
and
it
should
work
unless
there's
compile
errors
between
the
different
versions,
typescript,
which
means
we
then
have
to
fix
them
in
JS
or
contrib
wherever
they
were
occurring
once
they're
pushed
in.
We
just
do
a
double
sink
sink
into
the
the
merge
branch
and
then
sync
the
merge
Branch
into
me.
C
Now.
The
pr
I
put
in
the
notes
is
the
pr
de
effectively
sync
JS
and
contrib
into
the
merge
branch.
So.
C
C
I
think
that's
it
for
me,
but
no,
it's
good
to
have
that
done.
I'm
gonna
I
have
an
internal
task
for
this
I'm,
going
to
mark
that
as
done
today.
So
I
can
then
move
on
to
the
minification
and
any
other
bits.
If
you
look
back
at
the
pr
for
like
even
adding
the
instrumentation,
it
was
just
a
case
of
updating,
config
TS
to
say
bring
these
in
now
so
that
worked
really
smoothly.
Once
I
fix
the
bugs
in
Json
country,
which
were
typing
issues.
D
B
C
There's
the
the
pr
that
I
put
in
there
number
71.
I've
already
signed
it
off,
because
we
only
have
one
sign
up
required,
but
I
thought,
as
of
yesterday,
it
looked
like
Daniel
had
published
some
updates,
so
I
thought
I'd.
Get
you
guys
to
look
at
that
one
before
I
just
push
that
one
in.
D
So
are
those
going
to
be
in
general,
safe
to
merge,
I
mean
I'm,
just
gonna.
Look
it
over,
but
no
need
to
like
get
into
like
specific
details
right,
correct.
C
It
should
just
work
like
the
the
script
even
goes
to
the
point
of
manually
comparing
every
file
and
if
the
file
is
different
from
the
merge
into
the
merge
Branch
from
what
it
was
on
JS
country,
a
bit
overwrites
the
one
on
the
merge
management
with
every
byte
that
was
in
the
original.
So.
D
Okay,
how
often
does
this
does
this
run.
C
It's
manual
all
right,
so
if
you
go
to
the
actions
tab,
there's
there's
two
actions.
Unfortunately,
it
cuts
off
the
name,
but
the
first
one
is
the
hotel
JS
country
to
merge
staging
that's
the
the
pr
71
that
got
created
and
then
what
the
the
other
one
is.
The
repo
staging
Branch
domain,
that's
the
one
that
takes
that
merge,
staging
branch
managers
into
the
pkgs
format
and
creates
a
PR
to
push
it
into
Main.
D
C
C
So
yeah
at
this
point,
they're
manual
both
scripts,
only
like
having
one
PR
open.
So
if
there's
already
a
PR
open,
it'll
bail
and
it
won't
create
another
PR.
C
Any
time
we
know,
there's
something
major
in
JS
contribute
that
we
want
to
drag
in.
So
if
there's
like
the
reason
I
was
doing,
it
recently
was
the
typing
issues
which
were
blocking
the
compile
I
had
to
fix
them
in
Json
contrib,
once
they
got
checked
in
I,
then
merged
that
using
the
the
first
script,
approve
that
and
then
ran
the
second
script,
to
bring
it
in
wow.
C
So,
okay,
so
like
at
the
moment,
I
don't
have
like
the
APA
logs,
for
example,
if
we
did
have
API
logs
and
you
wanted
to
pick
up
an
API
logs
change
that
was
already
checked
into
to
JS,
then
that
would
be
the
point.
You'd
run
the
script.
C
Yeah
they're
ad
hoc
at
this
point,
the
the
first
one's
safe.
It
really
just
does
sync,
you
know
those
two
into
the
into
that
branch.
The
second
one
is
the
one
that
does
the
full
CI
test
pass
and
that's
the
one
that
has
been
breaking
every
time.
I've
been
doing
this
because
of
changes
made
in
Json
country,
okay,.
C
And
most
of
that's
because
it
attempts
to
make
the
components
when
possible,
run
node
browser
and
web
worker
tests,
even
if
they
don't
have
those
tests
in
Json,
contrib
Ute.
Some
of
them
didn't
pass
that
so
in
the
contrib
TS
I
had
to
say,
don't
try
and
do
this
in
the
browser
of
the
node
and
web
worker.
D
D
D
It's
been
a
long
project
good
to
see
it.
Finally,.
C
Finally,
here
yeah
and
we
have
full
history,
so
it's
perfectly
where
I
was
when
we
had
the
three
things
like
was
it
80
months
ago,
but
we
didn't
have
history,
it
was
all
manual,
but
now
we
have
history,
so
it
means,
if
you
want
to,
you,
can
actually
do
stuff,
and
it
also
means
if
we
want
to
rename
things
like
one.
One
thing
I
had
been
thinking
of
I'd
already
checked
in
the
Fetch
and
the
xxhr.
C
Is
we
probably
want
to
have
a
fetch
two
and
an
xhr2
with
the
changes
we're
talking
about
with
the
events
at
the
moment
they
get
dropped
into
packages
instrumentations
web
Fetch
and
xhr,
which
of
course,
if
you
want
to
have
a
version,
two
we're
gonna
have
to
call
them
fetch
two
and
next
hr2
or
something
so,
but
we
can
just
do
you
know
git
moves
and
move
them
around
and
update
the
script,
and
we
can
get
around
that
so
or,
if
there's
something
specific
in
it.
C
So
we,
if
you
create
a
branch,
for
example,
we
create
our
minification.
You
know
minimal
web
SDK
branch
and
there's
something
that's
not
there
that
auto
merge
folder
for
Json
contract
contains
everything
from
JS
and
Contra.
So
you
can
always
just
do
it.
A
git
move
in
your
own
Branch
to
bring
that
forward.
D
C
The
only
thing
would
be
is,
as
we
create
the
branches
we
should
document
them
in
the
main
readme,
so
I
think
I
have
the
intention
for
the
minification
branch
yeah.
So
I've
got.
You
know
main
the
staging
minification
TBD,
which
is
the
one
that
I'll
create
shortly
and
release
TBD
that
if
we
ever
decide
we
want
to
publish
this,
which
is
why
the
packages
will
get
renamed
as
sandbox
Dash,
so
they
don't
clash
with
them
anything
else,
so
yeah.
C
So
if
we
create
a
minimal
web
SDK
one
then
we'll
update
the
Ruby
as
well.
That
would
be
the
only
exception
to
checking
directly
into
the
main
branch
ourselves.
Hopefully.
D
Okay,
so
we
should
I'll,
probably
then
create
a
range
for
the
instrumentation
prototype
yeah
prototypes.
D
So
that's
all,
if
that's
all
then
I
can
just
quickly
show
you
and
update
the
events
SDK.
A
Martin
is
the
logs
SDK
part.
All
merged
is
all
completed.
B
A
D
I'll
look
at
it
one
more
time,
then
I
can
bring
it
up
tomorrow
and
I
yeah.
Let's,
let's
see
like
I
haven't,
let's
see
how
it
kind
of
depends
how
quickly
the
person
can
respond
to.
A
Okay-
and
it
doesn't
include
the
exporter
as
well
right,
it's
just
the
SDK,
okay,
okay,.
D
Yeah,
so
so,
when
I
was
so,
my
my
my
events,
SDK
prototype
right
now
is
is
basically
on
top
of
this
logs
SDK
Branch.
So
it
can.
You
know.
B
A
D
A
Test
it
yeah
yeah,
no
I,
think
I
was
thinking
that
before
the
events
SDK
it'll
be
good
to
work.
On
the
you
know,
exporter
log
exporter.
That
way
you
know
we
can
test
things
end
to
end.
D
A
D
D
So
I
have
this
react
application
that
I
put
under
examples
just
for
testing
as
part
of
my
branch,
and
so
this
hotel
page
view
file
it
just
simulates
an
instrumentation.
D
So
here,
like
I,
showed
you
like
last
week,
I
showed
you
the
Prototype
that
was
that
was
had
a
dependency
on
the
SDK
directly.
This
one
does
not.
This
one
has
depends
on
the
API,
so
it
gets
gets
to
events
emitter
from
the
API
like
this,
and
then
you
know,
that's
that's
pretty
much
it
so
like
from
the
instrumentation
perspective
like
it's,
it's
very
simple:
the
application
owner,
so
this
Auto
config
file.
D
This
is
what
they
would
have
to
do
to
configure
it,
so
they
have
to
because
the
events
SDK
depends
on
the
logs
logs
one
logs
SDK.
Then
both
need
to
be
configured
right.
So
so
here
we're
creating
a
provider
and
registering
Global
provider,
blog
provider,
right
nine
and
then
the
same
thing
for
the
events,
emitter,
creating
a
provider
and
registering
Global
events,
submitter
provider.
C
C
C
D
Yeah,
okay,
yeah,
and
so
that's
that
and
then
the
events
SDK
it
has.
You
know
the
way,
I
have
it
it's
just
a
sketch,
but
the
provider
just
creates
a
payment
emitter
and
then
the
event
emitter
when
it's
in
the
Constructor,
it's
getting
a
logger
through
the
logs
API.
D
It
takes
the
event
interface
and
then
it
creates.
You
know
it.
It
creates
the
event
name
event:
domain
attributes
on
the
log
record
and
then
admits
the
log
record
using
blogger.
C
C
B
C
B
C
A
And
in
your
previous
example
of
your
react,
app
I
think
there
you
use
the
same
term
navigation
for
both
the
event,
emitter
name
as
well
as
the
event
name
like
line
three
and
line
18
the
the
parameter
navigation
they
they
refer
to
different
things
right
like
in
line
three.
It
refers
to
the
event
emitters
in
the
scope
name,
whereas
in
line
18,
it
is
the
event's
name,
so
maybe
use
something
something
else
to
avoid
any.
D
D
Data,
so
we
have
right
now
in
the
API
there's
the
event
interface
so.
D
That's
right
what
I
wanted
to
ask
is
so
on
on
currently
on
the
event
interface,
we
have
just
single
attributes
property.
So
so
we,
what
I
was
we
were
trying.
The
question
we
had
yesterday
was
so
everything
all
the
attributes
here
would
get
put
in
into
the
event
data
attribute.
Is
that
correct
or.
A
But
if
we
were
to
you
know
Skip
asking
for
the
permission
and
then
go
ahead,
then
I
would
split
this
into
two
separate
Fields,
one
called
data
and
one
called
other
attributes
with
both
their
types
being
attributes,
yeah
and
and
then
you
know
they
said
them
accordingly.
On
the
lock
record
in
the
SDK.
D
I'll
put
a
link,
put
a
link
to
that
and.
B
A
Martin,
if
you
could
also
you
know,
review
my
PRS
and
you
know,
put
any
comments
or
you
know,
put
your
approval,
then
it'll
help
me
push
the
assignee
to
in
a
review
further.
D
Okay,
yeah,
so
that's
that's
the
the
resource
attributes
and
the
the
events.
Events,
yeah
yeah.
A
Resource
attributes,
I
I,
think,
should
be
straightforward.
A
There
might
be
questions
on
specific
attributes,
but
largely
I
think
it
should
not
be
any
controversial.
It's
the
only
thing
is
now.
There
are
I
counted.
There
are
total
11
resource
attributes
overall.
Are
they
too
many.
A
A
The
direction
I
was
going
with
this
question
was:
does
it
make
sense
to
further
Nest
them
in
a
in
a
tribute,
but
but
I
guess
that's
complicating
thing,
so
maybe
we
can
leave
it,
but
otherwise
yeah
these
attributes
were
all
for.
Based
on
our
past
discussions.
C
A
Yeah,
so,
okay,
let's
yeah,
that's
let's
talk
about
that
part.
Where
else
do
they
go,
I
mean
I,
don't
know
from.
A
A
Or
you
know
in
another,
whatever
mechanism
we
come
up
with,
but.
C
C
And
it's
not
necessarily
just
browser.session
ID
like
these
are
all
going
in
the
browser.
Namespace
screen,
width
and
height.
That's
fine!
Url!
That's
fine!
The
user
agent
is
a
good
change.
That's
fine!
C
The
visitor
ID
and
the
impression
ID
one
could
argue:
they're
not
browser
specific,
but
we
don't
really
have
enough
representation
from
non-browser
people
here
so
but
yeah
the
session
I
used
the
one
that
concerns
me.
I,
guess
more
than
the
other
ones.
C
A
Okay,
could
you
could
you
quickly,
you
know,
put
them
as
comments.
If
you
don't
mind,
yeah,
okay,
yeah
I,
think
the
the.
B
A
Attributes
I
can
I
can
completely
remove
them
from
this
PR.
If,
if
you
want
to
get
to
them
later,.
A
Like
what's
the
big
deal
in
repeating
the
same
attribute
under
two
name
spaces?
Ideally.
B
C
Which
is
the
whole
the
part
that
they're
going
through
at
the
moment
with
the
HTTP
semantic
inventions
they're,
actually
rationalizing
it
down
and
effectively
undoing
all
the
duplicates,
which
is
why
the
user
engine
is
now
called
user
agent
original.
So
introducing
more
is
just
going
to
make
more
work
later.
To
do
the
same
thing.
A
Okay,
yeah,
maybe
you
know
once
once
you
put
the
comment:
I
I
can
just
remove
it
for
now
and
and
then
we
I'll
create
another
issue
to
track.
It.
I'll
create
an
issue
with
all
the
resource,
varying
resource
attributes
and
and
then
use
that
for
tracking.
D
So
if
it
takes
a
long
time
to
to
agree
on
something
like
the
ephemeral
resource
attributes
and
equally
Define
it
as
a
just
a
regular
attribute
for
now,
that's
repeated
that.
A
That's
what
I
tried
doing
here
but
I
think
the
there
is
an
additional
concern
from
Nev
that
how
do
we
name
it?
Given
that
it's
going
to
be
common
with
the
mobile
I?
Think
if
we
can
answer
that
part,
then
I
think
we
can
go
ahead.
C
One
of
the
things
that
came
up
was
ECS
strictly
has
a
client
and
a
server
namespace
to
identify
where
it's
coming
from
so
ideally,
this
would
probably
be
like
client
session
IDs,
okay,
yeah
I,
don't
I,
don't
know
whether
ECS
currently
has
that
it's
probably
okay,
so
look
see
what
see
if
there's
an
equivalent
in
ECS,
and
maybe
we
try
and
adopt
that,
because
it
is
looking
more
and
more
likely
that
they
will
merge
they're
just
in
discussions
about
who
merges
what
so
for
new
stuff
like
this,
that,
if
we
can
just
say
yeah,
we
can
we're
just
copying
what
ECS
does
it'll
probably
get
the
nod
a
lot
faster.
C
C
I
did
have
get
a
chance
in
the
past
week
to
look
at
the
ECS
events.
They
don't
look
like
realm
events,
so
they
don't
seem
to
Define
ROM
events
anywhere
they're,
more
just
server-side
events,
so
that
doesn't
negate
the
event
data
discussion,
which
is
why
I
was
looking
for
it.
So
I
think
at
the
moment
we
just
keep
going
I
haven't
still
haven't,
had
a
chance
to
comment
on
Jack's
PR
about
well
when
event
s
should
be
used
or
not.
So
it's
been
very
time-conscious.