►
From YouTube: 2022-08-10 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
My
throat
this
morning,
I
see
martin's
out
nothing
on
the
agenda.
Santosh
will
probably
join
a
little
bit.
I
rams
out
today.
I
don't
see
anything
on
the
agenda
at
the
moment.
Is
there
anything
people
want
to
talk
about.
B
Nothing
too
specific.
I
was
just
looking
into
the
protobuf
support
in
the
browser
stuff,
so
just
a
minor
update.
It's
like,
I
think,
on
the
issue
that
martin
posted,
I
started
looking
into
it
and
it
looks
like
we
just
need
to
target
the
browser
for
a
couple
of
those
packages,
so
I'm
hoping
to
have
a
pr
up
for
that
soon.
B
B
Yeah
and
also,
I
think
we
need
there's
some
folder
structure
stuff
so
like
I
think
you
have
like
a
platform
folder.
That
then
has
like
node
in
the
browser
and
then
there's
some
stuff
that
goes
in
package.json
and
maybe
ts
config
as
well.
I'm
still
figuring
it
out.
A
D
I
just
I
mean
I
didn't
join
for
a
specific
reason.
I
just
wanted
to
check
in.
A
Yeah,
one
of
the
things
to
talk
about
yesterday
is
with
ted
being
out
for
another
couple
of
months.
We
were
looking
to
have
a
tc
member
included
to
help
us
drive
some
of
the
things
specifically
this.
I
can
never
print
out
the
ephemeral
resources
that
ted
started.
D
Oh
that's
why
he
hasn't
been
around
okay.
I
was
thinking
the
other
day
I
was
like
I
haven't
seen
ted
in
a
while.
I
can
assist
with
driving
that
the
challenge
is
I'm
going
to
be
on
holiday
for
about
four
weeks
starting
next
week,
so.
D
Yeah
it's
my
honeymoon
stuff
yeah,
so
I
can
help
up
until
up
through
next
week,
but
then
I
will
be
out,
unfortunately
yeah,
but
I
can
also
ask
around
at
the
gc
meeting
on
thursday
see
if
there's
anyone
else.
A
Yeah
I
was
going
to
sync
with
riley
yesterday
because
at
least
he's
familiar
with
the
space
that
we
play
in.
But
yes,
some
stuff
just
came
up
and
I
I
didn't
get
it.
D
Okay,
yeah
riley
would
be
good
and
and
sorry
I
couldn't
hear
on
the
through
microphone.
If
you
wanted
a
gc
governance
committee
or
tc
technical
committee
member.
D
Yeah
riley
might
be
the
best
of
all
the
people
that
you've
mentioned.
A
A
So
effectively
I
talked
we
talked
about
this
last
week.
This
is
a
document
that
ram
is
starting
and
is
going
to
drive,
because
I
haven't
had
the
time
to
be
able
to
do
this,
where
one
of
the
challenges
that
we've
had
is
we're
going
to
be
defining
the
set
of
events,
so
we've
already
got
the
basic
structure
that
santosh
should
be
driving,
but
now
we
need
to
start
drilling
into.
A
What
does
this
look
like
so
from
an
hotel
perspective,
every
vendor
can
say
I'm
going
to
generate
the
x
event,
and
it's
going
to
look
like
this
and
we're
all
going
to.
Basically
do
it
the
same,
but
to
do
that,
we
need
to
understand
what
those
events
look
like.
So
this.
A
A
Hopefully,
you
can
all
see
that
yeah
yep
cool,
so
with
this
document
here,
it's
we've
got
the
steps
that
to
process,
but
simplistically
we're
looking
for
the
major
vendors
who
want
to
be
involved
in
generating
rum
events
want
to
nominate
people
who
who
to
be
the
contact.
B
A
And
then
what
we'll
do
is
go
through
and
say:
okay
well
for
the
specific,
so
this
will
be
the
hotel
domain,
an
event
name,
but
for
each
individual
company
we'll
go
down
and
actually
put
well.
What
do
you
currently
call
it
just
so
we
can
identify
and
we
can
start
having
a
the
same
frame
of
reference.
A
So
these
are
proposed
names.
We
haven't
settled
specifically
on
these,
but
these
are
most
likely,
I
guess,
and
then
just
go
through
and
say:
okay
for
your
company
and
then
the
the
name
that
we
come
up
with.
This
is
what
is
currently
there
so
once
we
have
the
the
set
of
fields
and
what
it's
used
for,
we
can
then
say:
okay
well,
based
on
the
common
set.
So
here
we
have,
like
you
know
two
so
cisco,
and
that
and
microsoft.
A
We
currently
have
a
page
view
event
and
that's
probably
very
likely
that
everyone
will
have
a
page
view
event.
We
can
then
go
through
those
set
of
fields
and
say
this
is
it.
This
is
the
common
set
of
things
that
we
have.
This
is
how
we
go
about
defining
what
each
individual
component
is
like.
The
referring
uri,
there's,
actually
a
bunch
of
extra
stuff.
A
We
have
in
here
that
I
don't
see
but
yeah,
that's
that's
really
the
the
intent
and
then
we
go
through
so
we've
got
some
examples
here
that
rams
already
put
in
and
then
this
is
the
intent
of
saying
okay.
A
So
this
is
what
the
hotel
event
will
then
look
like,
but
until
we
have
the
set
of
events
and
the
idea
is
once
we've
got
these
events
listed,
we
can
then
break
up
into
smaller
groups
and
say:
okay
well
for
user
action,
we'll
have
you
know
aws
dynatrace
and
whoever
else
work
or
not.
So
that's
the
intent
of
this.
There
is
already
a
spreadsheet
that
was
created.
Previously
ram
was
going
to
go
through
that,
but
he's
actually
now
on
holidays
himself
for
a
little
bit
but
yeah.
C
Yeah
now
I
started
looking
at
this
document
yesterday
and
I
have
one
question
so
I
I
guess
like
previous
to
open
telemetry,
we
we
all
used
to
call
everything
as
events
and
now
we
have
spans
and
events
so
the
base
page,
what
we
used
to
have
like
a
page
view.
What
we
used
to
have
a
page
view
is
now
a
document
load
span.
A
Well,
it's
experimental
document
load
span
in
in
there
at
the
moment,
because
that
was
the
only
way
of
getting
events.
I
guess
that
is
an
open
discussion.
Do
we
want
to
convert
it
into
an
event,
or
do
we
want
to
just
have
as
a
as
a
span
that
will
be
as
part
of
defining
what
the
event
is.
A
Yeah
and
that's
really
the
definition,
so
some
people
will
want
spans.
Some
people
won't
so
effectively.
We
can
go,
define
the
fields
and
then
we
can
say:
okay.
Well,
the
consensus
is
that
from
an
hotel
perspective,
this
should
be
a
span
or
maybe
a
span
or
can
have
a
span
associated
with
it.
Yeah.
C
Yeah,
no,
I
I
only
meant
that
just
a
clarification
you,
you
might
want
to
add
at
the
top
of
this
document
that
we
want
to
list
here.
A
A
To
we
need
to
go
collect
the
fields
first,
but
I
I
I
see
where
you're
coming
from,
but
that's
really
for
me.
That's
that's
like
the
reason
we're
using
logs
is
for
the
transport,
as
opposed
to
spain
events,
because
spain
events
doesn't
support
everything
we
need
from
a
microsoft
perspective.
A
C
Yeah,
no,
actually,
the
question
really
is
that
once
we
collect
this
schema,
let's
say
we
use
it
for
the
purpose
of
data
validation
at
the
receiver
right.
So
so,
whatever
validation,
we
would
be
doing
on
events
we
we
could
do
that
for
the
spans
as
well.
A
A
Yeah,
which
the
whole
point
of
defining
the
individual
like
these
fields
like,
if
we
say
okay
name,
is
here.
One
thing
I
did
notice
he's
missing
is
we
say
it's
a
string,
but
we
don't
specify
a
maximum
size.
We
probably
shouldn't
do
that,
but
I
said
not
necessarily
in
ours
here,
but
in
the
hotel
in
in
down
the
bottom
as
part
of
defining
what
these
fields
should
be.
We
should
actually
have
you
know
maximum
length
or
ranges
for
that
validation,
purpose.
A
Okay
and
likewise
in
terms
of
the
total
size
of
event
there
is,
we
do
have
limits.
So
I'm
sure
we
want
to
specify
limits
here
as
well
to
stop
people
from
trying
to
do
a
denial
of
service
attack
by
sending
megabytes
of
events
in
a
single
event.
That'll,
it's
a
good
way
to
kill
a
server.
E
Another
question
for
you
now
is
the
goal
to
list
out
every
single
event
that
we
have
as
vendors,
or
is
it
just
to
sort
of
take
those
three
that
you've
said
so
that
we
can
at
least
attack
those?
Because
we've
tried
to
do
this
in
the
past.
I
think
with
the
sig
correct
me.
If
I'm
wrong,
like
we
have
like
three
different
or
four
different
documents
running
around.
A
Correct
yeah
yeah:
this
is
an
intent
to
try
and
consolidate
those
but
yeah.
We
want
to
list
all
the
events
that
we
want
to
have
hotel
defined
like
if
you
have
a
custom
event
that
you
don't
want
hotel
to
generally
have
to
do
then
don't
list
it.
A
Correct
yeah
anything
from
a
from
a
client.
So
while
I'm
very
browser-centric.
A
I
don't
have
direct
access
to
anyone
who
does
android
and
ios
apps.
We
do
have
some
react
users,
but
that's
still
browser-centric.
C
Yeah,
I
I
would
say
we
should
start
with
some
basic
events,
say
four
or
five
and
and
then
see
how
this
effort
shapes
up
and
then
automatically
we'll
use
it
for
the
any
other
events
that
we
create.
C
E
So
again,
can
we
decide
which
four
or
five
those
are
we
have
three
over
here,
nev
would
d?
Do
you
want
to
specify
two
more
then
we
can
at
least
focus
our
efforts
on
those
five.
E
A
A
We
specifically
use
it
the
navigation
for
page
view,
so
we
have
a
combination
of
page
view
and
page
view
performance.
So
that's
our
centric
approach
so.
B
A
But
it's
like,
if
you
have
an
event
that
is
just
added
like
the
intent
is
to
rather
than
limit
to
a
particular
set.
Let's
just
enumerate
what
we've
got
then
find
common
ground.
C
E
I
did
have
a
brief
conversation.
Unfortunately,
we
still
need
to
figure
out
like
see
at
the
end
of
the
day,
as
I
mentioned
to
you
santosh,
it's
all
about
identifying
the
right
maintainer
yeah,
so
we
probably
need
to
talk
to
the
hotel
javasig
as
well.
There's
a
reasonable
amount
of
overlap.
I
haven't
made
an
action
plan
around
that.
Yet
give
me
a
couple
of
weeks.
C
C
Given
that
morgan
is
here,
we
could
ask
him
on
what
he
thinks.
Morgan.
We
have
a
question
for
you,
so
splunk
has
an
android
sdk,
which
I
think
long
ago
john
had
mentioned,
that
they
are
willing
to
contribute
to
open
telemetry
yeah,
but
we
haven't
acted
on
it.
So
now
I
have
some
interest
and
we
were
wondering
you
know
how
to
go
about
it
and
says
you
know
the
it
requires.
Some
maintenance
you
know
has
has
has
a
certain
criteria,
so
I
was
wondering
what
is
that.
D
Yeah,
for
I
mean
first
off,
like
it's
open
source,
we
can
just
start
contributing
it.
However,
you
see
fit
to
be
honest.
We
have
a
team
working
on
it.
I
think
this
upcoming
quarter
is
going
to
be
one
person
committed
full
time
to
it.
D
So
we'll
need
I
mean,
like
splunk,
can't
do
all
the
lifting
of
porting
things
in,
but
if
it's
a
thing
that
you're
willing
to
use
as
sort
of
the
basis
for
this
we'd
love
to
assist
with
that
and
we're
supportive
of
it,
I
don't
know
what
you're
asking
specifically
but
like
basically
centers
like
if
you
want
to
use
any
part
of
it,
feel
free.
E
C
Guess
somebody
will
need
to
you,
know,
split
the
repo
or
separating
out
splunk
specific
code
and
and
make
it
generic.
So
that
is
one
effort
and
I
think,
a
bigger
question
that
sirath
is
bringing
up
is.
It
needs
some
folks
to
be
listed
as
maintainers,
and
unless
that
is
identified,
we
we
cannot
proceed.
D
Got
it
so
I'm
trying
to
think
the
best
way
to
go
about
this
then
so
one
thought
is,
we
go,
create
a
copy
of
it,
put
it
temporarily
in
a
hotel
repo,
either
a
folder
inside
of
an
existing
repo
that
we
own
or
a
new
repo,
doesn't
really
matter
which,
and
we
go
from
there
in
terms
of
like
refactoring
it
and
everything
else.
As
a
group
and
we'll
certainly
have
some
small
people
to
assist
with
that,
that's
probably
the
most
sensible
way
to
proceed.
D
I
suspect,
unless
you
have
other
ideas
in
mind
yeah.
I
have
to
check
like
on
on
like
timing
for
the
team
like
how
much
time
we
can
get
dedicated
for
this
effort,
for
this
upcoming
quarter
might
not
be
a
huge
amount,
but
it'll
definitely
be
some,
and
then,
if
we've
other
people
in
this
group
who
are
interested
in
assisting,
we
can
proceed
with
that.
D
I
suspect
I'm
saying
this
with
no
actual
facts
backing
it
up,
but
I
suspect,
but
the
factoring
it
of
it
is
already
such
that
the
splunk
stuff
is
probably
pretty
splunk.
Specific
functionality
is
probably
pretty
contained
just
because
everything
else
we
work
on
is
open,
telemetry,
and
so
that's
the
pattern
that
we
usually
follow.
D
Anyways
in
splunk,
observability
cloud
gdi
or
getting
it
in
so
I
suspect
it's
already
well
factored,
but
we
will
need
someone
to
validate
that
yeah
yeah,
but,
like
I
mean
we
can
do,
we
already
have
a
repo
for
this
group
within
the
opentselementary.github.org.
D
Okay,
so
why
don't
we
do
this?
Let's
create
a
rum,
repo
or
client
instrumentation
rather
repo
within
observe
within
open
telemetry,
and
then
we
can
start
by
just
pasting
over
this
code
and
then
going
from
there.
D
Rather
sorry,
yes,
thank
you
yeah
yeah.
Okay,
do
you
want
me
to
follow
like
go,
get
a
repo
created
for
the
android
sdk.
C
But
before
that,
I
think
see
that
maybe
we,
it
might
be
helpful
to
have
one
discussion
with
the
current
maintainers
to
see
what
they
think.
What
is
the
approach,
and
then
you
know
I
can
get
some
more
folks
from
my
side,
maybe
from
other
companies.
If
anybody
is
is
willing
to
participate
in
that
effort.
You
know
please
come
forward
and
then
we
could
have
a
combined
meeting.
D
B
A
I
can
probably
just
give
an
update
on
the
on
the
web
sandbox
so
for
any
web
developers,
so
wow
right
for
it's
really
playing
up
today,
so
with
the
merging
pr.
So
I
got
daniel's
got
some
comments.
A
I
did
an
update
to
that
yesterday,
so
the
intent
is,
the
sandbox
will
automatically
merge
from
the
js
and
the
js
api
repos
regularly
that
pr
is
almost
about
to
go
in
and
then
I'll
have
to
have
another
piano
to
effectively
push
that
into
a
form
that
we
can
then
start
compiling,
which
I
have
an
older
pr,
which
was
just
manually
copying,
but
we're
still
probably
a
few
weeks
away
from
getting
that
in
once.
That's
done
we'll
then
have
effectively
a
mostly
automated
chain
of
getting
changes
from
the
js
repo
into
the
sandbox.
A
So
we
can
then
start
splitting
off
and
working
on
different
streams
of
work,
so
I'll
be
concentrating
on
the
this
initially
on
the
minification
aspect:
to
try
and
reduce
the
size
and
then
any
things
we
find
that
we
can
push
back
to
the
main
one
we
push
it
back
through
and
then
a
little
funnel
back
through
this
automated
step
is
the
idea
I
know
on
the
client
instrumentation
on
the
slack
channel.
There's
been
a
couple
of
discussions,
so
there
are
multiple
branches.