►
From YouTube: 2022-11-01 meeting
Description
Instrumentation: Messaging
A
B
E
A
E
I
haven't
seen
any
further
activity
either.
You
know
pretty
much.
Everyone
was
at
with
a
cubecon.
Last
week
my
PR
ended
up
getting
tagged.
The
sale
and
I
closed
it
and
said
you
can
reopen
it
if
you
like,
except
I,
don't
have
the
permissions
to
reopen
so
I
I,
linked
it
to
tigrants,
because
they're
pretty
much
the
same
thing
right,
but
based
on
the
376
being
out
there
for
multiple
years,
I
think
it's
we're
still
on
the
slow
path
for
that.
E
E
Box
I
I
thought
it
was
going
to
be
an
easy
fix.
I've
been
working
on
it
for,
like
the
last
three
working
days
and
I'm
still
trying
to
get
it
to
function.
I
think
what
the
problem
is
with
my
git
merge
scripts
when
it
merges
the
JS
and
the
contrib
and
all
the
history
back
into
a
single
repo.
The
contrib
must
have
been
forked
from
Js,
so
the
git
merge
goes.
Oh,
these
are
the
same
thing
and
actually
ends
up
overriding
things
in
the
wrong
way
around.
E
So
confrib
files
end
up
in
the
JS
repo
because
they
were
changed
afterwards
and
likewise,
when
some
of
the
contrib
stuff
gets
copied
around,
it
looks
like
it
thought
they
were.
You
know
the
same
file
just
being
moved
and
that's
causing
my
script
group
as
well
so
I'm
trying
to
resolve
that
in
a
nice
automated
fashion.
So
we
can
keep
this
thing
going,
but
it's
a
lot
more
painful
than
everyone.
E
D
So
I
think
the
goal
is
right,
because
we
have
this
repo
you're
gonna.
You
want
to
keep
it
updated,
based
on
like
what's
what
new
commits
in
in
the
JS,
SDK
and
contrary
yeah,
so
you
want
to
keep
pulling
things
in
and
we're
making.
You
know
we'll
be
putting
things
be
making
commits
to
in
this
sandbox
and
we
want
to
keep
it
up
to
date
with
the
other
repos.
E
E
Once
all
that
that's
the
bit
that
I'm
now
working
on
to
get
to
get
that.
So
we
end
up
with
correct
copies
once
that's
there
I'm
going
to
have
another
script,
which
will
then
pull
out
all
the
web-based
stuff
and
Munch
it
in
into
the
equivalent
of
what
I
had
in
my
first
PR,
where
I
did
all
manually
so
that
we
have
Maine
will
be
a
buildable
web-based
repo,
which
will
have
all
the
browser
tests
and
it'll
be
generating
browser
bundles
with
rollup.
E
I
would
like
other
people
to
you
know
if
you've
got
something
you
want
to
play
with,
we'll
create
a
branch,
and
you
can
own
that
branch
and
we'll
yeah
do
whatever
we
want
once
we
validate
with
the
whole
idea
of
the
sandboxes,
we're
just
like
hacking
and
making
it
work
once
we
validate
that
this
is
a
would
be
a
good
thing.
We'll
then
create
PR's.
So
you
have
to
manually
move
that
code
back
into
the
jsr
country,
repo,
depending
where
it
is
once
it
gets
merged
back
into
there.
D
E
Yes,
so
so
there's
a
I've
already
got
an
action
on
the
thing
where
I
can
just
like
trigger
the
script.
To
do
it,
but
like
the
first
one
worked,
the
second
one
for
the
API
worked.
So
then
I,
then
we
came
to
the
contrib
and
then
it
stopped
working.
So
that's
what
I'm,
where
I'm
at
now
so
I'm,
making
that
working
in
for
the
country
repo-
and
that
was
because
the
API
was
moved
from
its
own
repo
back
into
Js.
C
E
D
E
The
the
only
the
only
change
that
I
want
to
make
domain
is:
is
this
automatic
pipeline
to
get
in
there?
Okay,
so
yeah
I
I
don't
want
to
make
direct
changes
to
Maine
unless
we
have
to
yeah.
There
will
be
this
initial
setup
to
get
a
building
and
I
suspect
there
will
be
some
changes
we'll
have
to
do
for,
like
you
know,
updates
of
different
packages
that
break
things
but
yeah
in
terms
of
major
changes.
E
E
F
A
D
A
F
B
F
A
F
F
F
A
F
B
E
E
E
F
A
C
A
F
A
F
A
F
Yeah
I
I
get
it
so
the
top
of
the
things
that
we
Define
once
you
know.
Let
me
take
this
example
screen
width.
Where,
where
do
you
think
we
are
going
to
Define
this
in
semantic
Dimensions?
You
know
I
think
we
established
that
we
need
to
Define
these
things
in
semantic
Dimensions,
correct
yeah.
So
where
will
we
Define
it?
Let's
say
we
pick
a
place
now.
It
obviously
is
not
going
to
live
under
today's
yeah.
F
Okay,
so
you
build
a
resource
now
somebody
looking
through
these
screens
if
they
want
to
understand
the
definition
right
for
you-
and
you
know
probably
a
few
here-
it's
obvious
some
brand
new
person
coming
through.
They
look
at
all
these
foods
and
go.
How
do
they
know?
How
do
they
know
to
navigate
into
bunch
of
these
things
and
end
up
at
a
particular
field?
Right
and
there's
got
to
be
some
if
you're
looking
at
a
trace
and
actually
do
the
instrumentation,
it's
pretty
obvious.
F
A
E
E
A
A
A
F
F
E
Yeah
I
think
what
we're
really
defining
here
is.
This
is
we're
defining
what
our
proposal
would
be
and
therefore
the
Deans
that
will
they
will
have
for
that
proposal
for
these
particular
ones.
For
the
varying
attributes.
Let's
say
we
are
assuming
they'll
go
in
the
thermal
resources,
assuming
a
federal
resources
goes
in.
F
C
D
B
C
C
E
E
F
Okay,
cool
I,
I
think
this
makes
sense.
We
can
we
can,
you
know,
put
this
thing
out.
You
know
anything.
That's
already
existing.
Somehow
we
need
to
say
HTTP
URL
from
whatever
that
location,
I
don't
know,
we
can
link
to
different
dots
and
stuff
right.
So
that
way
somebody
happens
up
on
page
view.
We
know
exactly
where
to
find
the
definitions
and
things
and
don't
have
any.
You
know
we're
not
going
to
have
a
have
ambiguity,
but
at
least
should
be
clear,
I'm
ready
to
find
definitions
for
things.
F
F
B
A
E
Comes
back
to
our
discussion
on
data
and
like
I
I
would
really
like
to
avoid.
Having
you
know,
deeply
nested
names
because
of
the
duplication
so
I
would
I
would
prefer
to
have
it
as
event.data
and
then
we
say
we
have
event.data
for
a
page
name.
We
need
a
screen.
Width
screen
height,
not
turn
around
and
say:
okay,
we're
going
to
call
all
of
these
browser.screen
web
browser.screen
height,
because
that's
just
going
to
get
us
into
the
same
pain
point.
A
E
E
Well,
when
we
say
everything's
in
event.data,
we
can
say:
okay,
we
have
no
data,
no
need
for
dotted
names,
because
the
event.data
is
unique
to
the
event,
so
we
could
just
say
event.data.
Instead
of
page
dot
URI,
we
could
just
say
it's
now,
your
eye
or
URL.
It
doesn't
need
to
conform
to
the
general
attribute
semantic
convention,
because
it's
not
it's
it's
Unique.
To
that
event,.
E
B
D
So
I
was
originally
thinking
that
that
could
be
the
case
like
the
way
you
describe
like
page
page
URL
or
versus
URL
I.
You
know
I
originally
thought
that,
even
if
the,
if
the
attributes
were
top
level
they
would
for
for
logs,
it
would
would
not
have
to
conform
to
the
semantic
conventions,
because,
because
of
the
convention
of
if
it's
an
event,
it
has
event.name
about
dot
domain,
then
everything
else,
that's
not
even
doubt,
is
just
specific.
To
that
event.
A
Yeah,
but
look
at
the
original
reason
why
these
are
part
of
the
spec
repo
right
I,
think
it
is
to
like
I,
think,
I
think
what
never
suggesting
that
your
event
name
should
be
the
primary
key.
Now
it
should
serve
as
a
reference
to
what
the
other
fields
are
about
yeah,
but
that
is
not
there
in
the
you
know
spec
today,
so
so,
until
we
get
that
understanding,
agreed,
yeah
yeah
for
sure.
E
Yeah,
which
is
why
I
think
having
event.dater,
is
an
easier
sell
than
saying
okay.
Well,
when
we
have
a
long
event
and
we
haven't
happened
to
have
event.name
so
therefore
we
can,
then
the
attributes
now
ignore
the
existing
semantic
conventions.
I
think
that
would
is
going
to
be
a
really
hard
sell
for
the
same
reason
that
they
have
namespaces
today
and
that
you
could
end
up
with
a
complex.
E
A
E
Well,
it
comes
back
so
if
there's
no
support
for
having
our
own
data
value,
so
this
is
like
it
goes
back
to
the
proposal.
I
put
in
the
now
closed
PR
of
having
the
schema
and
then
yeah.
This
is
tigran's
event,
name
the
main
thing,
I
I
think
part
of
it.
That
needs
to
play
out
and
come
to
a
conclusion,
and
we
didn't
do
that
last
week
because
typical,
you
know,
the
log
scene
didn't
happen.
Yeah.
A
F
E
Like
if
we
continue,
if
we,
if
we
say
okay,
it's
going
to
stay
and
we're
going
to
have
event
name
event
domain,
then,
for
me
at
least
event
data
is
a
logical
extension
of
that,
and
then
we
just
say
we
have
all
our
Fields
living
inside
of
that,
so
that
you
have
a
generic
event,
any
any
back-end
system
can
say:
okay,
I
have
event
these
three
event.
Prefix
Fields.
So
therefore
I
know
everything
event.
E
E
F
E
C
F
A
F
E
A
E
E
But
I
think
tigrett
has
this:
the
schematized
logs
I
think
are
still
out
there.
I
think
that's
that
might
be
what
he's
must
have
proposed
for
it
which
sort
of
works
as
well
but
depending
how
that
plays
out,
but
yeah
I
I
agree.
Let's
just
move
on
and
send
us
if
you
want
to
drive
the
extension
that
would
work.
Okay,.
C
F
C
F
F
D
C
D
From
the
definition
perspective
like,
so,
if
you,
if
you
don't,
if
you
put
everything
in
the
in
event
data
and
then
let's,
let's
assume
that
we
have
the
the
way
we
write,
the
conventions
is
by
like
we,
we
write
like
a
one
document
per
event,
type
correct
listing
all
the
different
attributes
in
that
event
type
so
like.
If
we
have,
if
we
have
events
that
share
attributes,
even
in
this
like
either
in
the
same
domain
or
even
across
domains,
possibly
yeah.
E
F
F
E
Can
I
change
that
required
to
expected
because
it's
going
to
depend
on
the
event,
so
some
you
will
always
accept
like
the
page
title
you're
not
always
going
to
have
a
document.title
so,
but
we
should
definitely
Define
a
fallback
like
not
not
the
specified
or
something
but
yeah
it'll
be
a
case
of
saying
these
are
the
fields
that
will
be
expected
to
be
there
and
we'll
have
some
that
we
might
tag
as
required
and
some
that
will
say
are
optional.
Foreign.
F
We
will
open
one,
you
know
look
at
that
and
we
are
going
to
follow
what
okay.
So
then
I
think
that
answers
most
of
the
questions
here.
I
understand
your
logic
behind
using
HTTP,
URI
and
stuff.
What
we
had
in
the
spreadsheet
it'll
be
ideal
if
the
spreadsheet
name
matches
what
we
eventually
end
up
with,
but
I
think
that's,
okay,
but
this
is
a
working
doc.
Yeah
I
think
that's
fine!
F
Going
back
here!
The
next
one!
Fetch
timing
is
marked
as
done
in
spreadsheet,
but
has
several
blanks
touch
timings
you
have
a
star
in
a
prefix.
Was
a
done
marker
I
think
there
are
some
blank
Fields
here
should
probably
like
these
guys
here.
I,
don't
think
we've
gone
through
these
at
all
or
no
I
think
what
was.
E
C
E
C
E
F
E
Yeah
yeah
and
we
can
enumerate
what
exists
today,
but
we
just
need
to
in
in
the
specification
we
just
need
to
say
if
this
other
document
changes,
those
additional
Fields
should
also
be
included
and
they'll
be
optional
Fields,
because
not
all
browsers
will
support
them
like
most
of
these
are
optional
as
well,
because
it
depends
on
the
request.
You
don't
have
the
the
domain
start
and
end
unless
the
browser
to
look
up
and
when
it
does
look
up
it
cases
it
for
whatever
the
DNS
okay
life
cycles,.
F
E
E
F
D
F
B
F
C
E
Because
yeah
device,
ID
and
type
are
really
for
the
device
that
they
would
really
belong
in
the
yeah.
Does
it
would
they
change?
No,
they
they
shouldn't
change,
they
could
be
resources.
The
SDK
version
might
change
the
operation,
ID
is
actually
the
trace.
Id
operation
name
is
actually
the
page
title
and
then
yeah
session
ID
news
right.
You've
already
got
there.
E
F
E
E
If
we
have
an
event,
there
will
be
a
trace,
ID
associated
with
it.
So
it's
like
for
application
insights.
We
always
you
know.
We
don't
have
an
ongoing
span
because
we
don't
have
spans,
but
but
we
do
identify
the
fact
that
this
page
is
this
operation
ID.
So
this
Trace
ID,
which
lives
the
lifetime
of
the
page.
C
E
E
E
F
E
E
F
E
C
C
B
F
C
E
F
A
F
F
F
A
D
E
C
C
F
D
It's
a
timing
for,
for
us,
it's
a
timing
event.
So,
like
we
have
we
capture
a
bunch
of
different
timing
like
actually
at
least
the
way
like
we
presented
on
the
back
end
represent
it
in
the
back
end.
Is
we
capture
different
timing?
Events,
which
can
be
you
know
things
like
paint
timing,
but
it
can
also
be
like
the
first
first
interaction
so
and
it's
something
that
you
would
then
like.
We
thought
be
useful
like
to
display
like
on
on
a
timeline.
C
A
E
I
think
it's
more
okay,
so
when
do
we
want
to
find
another
in
this
case,
is
the
gorilla
to
find
them?
If,
if
we
don't
Define
them,
someone
will
request
it
to
Define
them
because
I'm
getting
growing
pressure
internally
as
well
like
we
don't
Define
it
today
and
at
the
moment,
teams
go
and
do
this
themselves,
but
it
would
be
nice
to
have
it
as
a
standard
thing.
D
A
E
C
E
E
F
F
So
if
portal
changes
happen,
we'll
count
them
as
four
page
views
that
save
it
as
though
they
are
for
different
previous.
That's
basically
it
I,
don't
think
there
is
a
mean
for
finding
out
if
they
were
virtual
or
physical
type
changes
question
is:
do
you
guys
have
anything
any
you
know
experiences
built
around
the
the
infrastructure
that
which
is
basically
what
it
is.
A
E
Yeah,
that's
always
the
hottest
thing.
It
depends
on
the
framework
they're
using
whether
it's
in
your
angular
react
or
you,
because
they've
got
to
hook
their
the
router
changes
so
yeah
today
we
just
say
the
duration
is
always
zero,
and
but
we
provide
the
ability
to
for
you
to
give
us
the
duration,
because
of
that
in
the
open,
Telemetry
World.
E
In
theory,
someone
can
go
and
build
a
standard,
router
extension
a
router
instrumentation
for
each
of
those
and
calculate
it,
but
yeah
the
the
question
is:
do
we
want
to
have
an
additional
flag
that
says
this?
Is
a
route
change
versus
a
PG,
or
do
we
don't
have
it
as
a
you
know,
affecting
a
separate
event
saying
from
Microsoft
perspective
we
just
treat
them
as
page
views,
but.