►
From YouTube: 2023-02-07 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
B
C
A
A
A
All
right
so
there
is,
and
the
HTTP
semantic
inventions
group
is,
is
in
action
and
they're
kind
of
taking
15
minutes
of
the
specs
meeting
to
talk
about
some
of
the
stuff
that
they're
working
on
and
they're
making
a
lot
of
changes.
I,
don't
know
how
popular
they
are.
One
thing
that
they're
trying
to
lobby
for
is
to
revamp
the
HTTP
Dot
semantic
conventions
to
be
URL
dot.
So,
instead
of
instead
of
HTTP
scheme,
it'll
be
URL
scheme
instead
of
http.url
it'll
be
url.full,
HTTP
Target
would
be
url.path
and
url.query.
A
I
think
there
are
good
motivations
behind
this
I,
don't
know
if
everybody
fully
sees
the
eye
to
eye
that
these
changes
are
worth
it,
and
at
least
these
ones
are
coming
from
elastic
common
schema,
which
I
know
there
has
been
some
discussion
about
adopting,
at
least
in
part,
but
I,
don't
know,
I
think
there
was
a
little
bit
of
backlash
but
like.
Why
are
we
changing
these
things?
A
What
are
the
benefits,
and
are
we
just
going
to
like
piecemeal
it's
kind
of
pick
and
pull
from
elastic
common
schema
when
we
want
to,
or
are
we
going
to
take
elastic
common
schema
for
everything
in
general,
yeah
I
think
Secret's,
comment
right
here,
really
sums
things
up
and
he's
just
kind
of
saying:
I
know
that
the
conventions
are
not
stable.
However,
a
lot
of
people
have
built
stuff
that
depends
on
the
conventions
already.
A
So,
even
though
technically
people
shouldn't
complain,
if
things
are
broken
like
they
probably
won't
complain,
so
be
ready
for
it.
If,
if
we're
going
to
move
forward
with
such
breakage.
A
A
A
These
other
semantic
conventions,
PRS,
have
been
open
for
a
little
bit.
I
think
there's
been
some
minor
tweaks
and
renames,
and
the
author
of
this
was
on
the
phone
driving
in
the
rain.
So
discussion
was
not
entirely
fruitful
on
it,
but
yeah
I
think
there's
a
DB
dot
bind
that
that
values
semantic
convention.
This
person
is
adding
I
think
it
was
called
like
DB
dot
values
before,
but
people
wanted
the
bind
in
there
and
it
makes
a
little
more
sense
and
then
well
that
name
already
existed
as
a
semantic
convention.
A
She's
almost
recommending
a
fallback
in
case
you're,
not
in
a
pool
to
be
derived
from
the
connection.
A
There
is
apparently
Prometheus
has
a
unknown
type
of
metrics.
A
So
there's
a
proposal
to
add
I
think
a
are
we
see
it's
that
type
and
a
value
of
unknown
as
a
scientific
convention
for
Prometheus
metrics
that
are
scraped
with
an
unknown
type.
A
There
is
Json
stability,
maybe
finally,
almost
already
here
it
seems
like
this
is
ready
to
be
merged
after
the
1.18
spec
release.
A
So
I
don't
know
how
much
that
will
impact
anybody
here,
but
I
think
anybody
who's
relied
on
Json
or
as
a
vendor
might
care.
A
All
right
moving
on
even
further
there's
a
lot
of
talk
about
adding
links
after
span
creation
and.
A
Didn't
actually,
despite
the
large
number
of
bullet
points,
we
didn't
talk
about
this
all
that
much
because
it
would
be
something
that
would
probably
take
meetings
in
and
of
itself,
but
not
sure
if
Carlos
put
this
in,
but
somebody
put
this
in
and
it
was
just
kind
of
really
well
well
summarized.
A
There
are
a
lot
of
reasons
why
there's
kind
of
a
need
to
create
links
after
a
span
has
been
created.
I
think
like
the
biggest
problem.
A
A
Ideally,
you
would
like
to
try
your
best
to
sample
linked
as
fans
and
they
span,
that
is,
that
references
these
links
so
adding
them
after
the
fact
makes
it
harder
for
you
to
guarantee
that
you
will
actually
sample
the
span
that
it's
having
having
these
links
added
so
I,
don't
know,
I
think
there
will
be
some
suggestions
as
to
how
this
should
actually
work,
but
yeah
I
don't
intend
to
describe
all
the
details
here.
Just
just.
A
A
And
then
the
last
thing
that
was
really
talked
about
was
this
last
issue
we
talked
about
this
again.
The
title
is
a
little
misleading.
It
says
supporters,
specifically
removing
attributes
from
an
instrument
API
but
really
I,
think
what
the
author's
intent
here
is.
A
He
has
a
in
Time
series
created
with
some
attributes
and
it's
in
the
cumulative
temporality,
and
he
knows
that
this
time
series
will
never
produce
any
more
data
like
something
hasn't
moved
and
like
one
of
the
attributes
has
kind
of
gone
out
of
scope,
more
or
less,
but
I
think
with
a
cumulative
metric.
It'll
keep
reporting
that
old
value
and
he
wants
a
way
to
just
discontinue
this
time
series
it's
no
longer
felona.
A
Yeah,
the
last
two
issues
we
really
did
not
get
to
so
I
would
just
say
say
that
they
exist
and
and
move
on.
Any
questions
comments
concerns
on
other
stuff.
A
C
Seems
like
it
I
I
assume
people
are
using
it
and
it'll
be
a
breaking.
You
know
whatever
you
want
to
call
it,
they
can
change
against
people's
if
people
have
queries
running
against
this
stuff,
but
it's
for
the
best
to
get
HTTP
semantic
conventions
to
stable.
That's
great
and
I'm
glad
to
see
that
work.
Moving
forward.
A
Yeah,
it's
nice
to
see
it
moving
forward.
I,
don't
know
how
popular
these
name.
Changes
are
just
yet,
especially
I.
Think
the
the
schema
translations
will
probably
not
support
all
of
the
all
of
the
migrations.
So
just
curious.
This
would
be
like
something
that
would
potentially
break
things
for
Shopify
or
GitHub.
Definitely.
E
It
seems
like
wasted
bites,
okay,
so,
but
what
about
the
the
other
three
URL
Target?
Well,
they
those
are
getting
renamed.
A
A
So
they're
proposing
they're
proposing
so
okay
I
mean
one
one.
Nice
thing
I
think
you
were
recently
bitten
by
this
is
that
Target
would
become
path
and
query
which
okay,
which
would
allow
you
to
censor
out
the
query
more
easily,
but
But
ultimately
I.
Think
tigran's.
Objection
here
summed
up
a
lot
of
the
objections
that
we
were
seeing,
and
that
is
that
he
knows
the
connections
aren't
stable.
So
technically
they
should
be
changeable,
but
a
lot
of
people
have
built
stuff
that
depends
on
them
already.
So.
B
C
I
have
a
general
like
approach
to
this
stuff,
which
is
like
if
it
you
know
if
the
idea
is
to
get
to
1.0
or
then
use
the
1.0
use
the
power
of
the
the
bump
to
do
to
Signal
the
breakage.
Don't
do
a
bunch
of
things
to
get
to
one
point:
I
was
minor
version
pump
and
then
be
like
guess
what
1.0
like
chunk
them
all
with
me,
like
yep,
if
you
bump
to
1.0
like
this,
is
like
you.
Can
you
gain
a
lot
of
things?
C
But
here
you
know
we'll
list
all
the
things
and,
and
that
way
it's
like
explicit,
and
it
won't
say
you
know
nobody,
the
obviously
someone
has
bad
version
control
stuff.
You
know
in
their
in
their
in
their
you
know,
package
Json
or
their
ruby
gem
stuff,
but
like
if
they're
doing
it
properly,
that
should
at
least
correctly
signal
to
them.
But
you
know
it's
like
if
it
has
to
have
it
has
to
happen,
but
yeah,
it's
I,
think
I
can
think
offhand
of
a
couple
things.
This
will
stop
before
I.
C
A
Yeah
I
think
I
think
in
general
they
I
think
they
are
kind
of
like
looking
at
the
HTTP
semantic
conventions
and
hopefully,
with
this
1.0
in
mind
and
just
making
sure
that
this
is
last
chance.
I
think
they
were
kind
of
saying
this
is
last
chance
to
maybe
correct
some
things
that
they're
seeing
that
they
they
would
like
to
change
and
and
trying
to
make
it
happen.
So
we'll
see
how
that
goes.
A
A
Yeah
I
think
just
yeah
y'all
were
around
the
only
other
notable
thing.
I
think
was
otlp
Json
soon
to
be
stable.
Spam
links
after
creation
everybody's
here.
For
that
all
right
shall
we
move
on
to
our.
A
Agenda
it
looks
like
Arielle
has,
is
pleading
for
some
reviews.
E
Sure
this
looks
reasonable
to
me.
The
Sinatra
fix
by
you
know.
Our
rules
of
etiquette
here
are
two
reviewers
before
merging
person's
excited
to
get
this
merged.
I
want
to
encourage
them,
but
folks,
you
know
if
you
can
take
some
time
to
get
eyes
on
it.
That'd
be
great.
E
Basically,
what's
happening
is
they're
getting
like
a
double
span:
events
for
exceptions,
one
for
the
Sinatra
and
one
for
Iraq
and
seems
Superfluous
because
rack's
already
recording
it.
So
you
know
this:
PR
is
removing
the
Sinatra
span,
recording
the
exception
and
addition
to
the
Rax
band.
Since,
though
Sinatra
spanning
is
enriching
the
existing
rackspan
with
data
so.
E
Yeah
that
portion
of
it,
the
second
one,
is
looking
to
do
a
release
rollout.
So
this
is
bumping
the
gems
for
rails
dependencies
if
I
can
get
eyes
on
this.
So
generally
speaking,
for
version
bumps
and
for
releases,
I
think
with
releases
I'm
a
little
more
like
you
know,
release
requests
or
whatever
I'm
a
little
more
loosey-goosey.
But
my
question
for
y'all
is
for
version
bumps
and
for
release
requests.
E
E
You
know
my
preference
is,
you
know
at
some
point.
It's
like
there's
gonna,
be
an
inflection
point
where
it's
like,
not
as
many
people
are
engaged
and
I
really
like
to
move
things
along
as
as
fast
as
I
can,
but
I
also
recognize
that
having
a
second
set
of
eyes
in
these,
especially
when
it's
like
release,
coordinating
and
dependency
management,
helpful.
C
B
Think
that,
like
for
this
change,
specifically
like
you
and
I,
had
already
talked
about
it,
and
it's
been
discussed
elsewhere
and
like
like
this
I
wouldn't
say
like
you
need
a
second
approval
for
this
PR
to
merge
it
and
go
like
yeah.
Just
do
it
like
this
one,
this
one's
well
known
and
well
under
fit
I.
Guess.
B
Basically,
I
think
I'm,
saying
I
trust
you
to
use
your
best
judgment
in
like
what
is
Trivial
and
just
you
can
just
go
like
depend
about
stuff.
That's
super
super
benign,
like
you
know,
I
you've
been
around
long
enough
to
know.
What's
gonna
break
stuff
and
what's
probably
not
going
to
break
stuff
and
yeah
I,
don't
know
that's
my
feeling
on
it
is
I,
don't
want
you
to
have
to
wait
for
piddly
little
stuff.
That's
like
already
known
and
understood.
C
A
Yeah
I
feel
like
this
is
maybe
a
bad
example,
because
it
is
dropping
support
for
a
version
of
a
well-known
framework,
but
we
have
had
this
discussion
before
so
right
right
and
there
have
been
other
PRS
about
this.
This
is
just
like
the
final
version
twiddling
yeah.
B
Also
I'm
more
than
happy
to
give
your
like
reviews
and
chat
about
it
and
stuff.
Just
Spam
me
and
slack.
My
notifications
are
out
of
control
and
I'm,
not
paying
enough
attention
to
stuff
so
which
is
a
problem.
I
need
to
fix
and
I
will
fix,
but
you
know
in
the
meantime
like
please
just
yell
at
me.
If
you
want
me
to
review
stuff
and
I
haven't
looked
at
it
yet,
like
I
am
more
than
happy
to
I,
just
super
distractible
these
days,
I
guess.
B
E
So
yeah
depend
about
you
know:
I
added
dependent,
bot,
pretty
cool
thing,
but
it
is
adding
pessimistic,
locking.
E
And
I
know
in
the
past,
we've
talked
about
pessimistic
blocking
being
a
problem
right
because
we
want
to
do
stuff
like
testing
Futures
in
the
head,
but
I
don't
think
we're
doing
any
of
that
right
now
are
we
like?
We
don't
have
any
appraisals
for
like
yo
the
next
version
of
something.
B
E
Okay,
by
the
way,
I
configuring
the
penabot
to
to
not
do
this.
E
E
Yeah
I
mean
in
a
lot
of
cases,
it's
like
a
lot
of
production
use.
Cases
I
have
is
using
the
lock
file,
lonely,
for
example,
whereas
these
are
all
kind
of
like
you
know.
There
are
no
lock
files,
so
this
is
trying
to
do
like.
There's
got
to
be
some
sort
of
like
rule
about
pessimistic
enough
optimistic,
versioning
defendancies,
and
why
not
I
just
didn't
look
at
it,
but
if
you
got
a
minute
hey
worth
I'll
take
it.
C
E
Yeah
so
there's
a
couple
of
folks:
who've
been
participating
in
collaborating,
but
they
are
it's
like
dinner
time
basically
commute
time
for
them.
For
this
meeting.
E
I
know
Nicholas,
for
example,
like
we
haven't
seen
him
in
a
little
while,
but
he
would
like
call
us
on
the
phone.
He
was
communal
because
it
was
like
he
was
trying
to
participate
in
the
meeting.
So
I
don't
know
if
it's,
if
we
don't
have
any
West
coasters
here,
do
we
say
Rob?
Are
you
on
the
west
coast
I
know
we
have
a
mountain
time,
I'm.
E
A
West
coaster-
oh
okay,
so
you're
on
Pacific
I'm,
not
Pacific,
put
you
at
a
9
30.
B
B
Would
we
want
to
change
the
day
or
something
I
I
would
be
curious?
Are
these
the
folks
that
have
been
particular
are
they?
Are
they
asking
for
us
to
change
the
time.
E
No
I've
just
asked
informally
had
a
conversation
with
people
I'd
ask
them
like
hey.
Well,
you
know
if
you're
available
it'd
be
great
to
have.
You
come
to
this
meeting
to
discuss
this
particular
thing
when
they
were
advocating
for
like
a
PR
or
something.
E
Yeah
options
options
alternating
weeks,
I,
don't
know
I'm
just
spitballing.
E
No
because
we
have
like
a
one
week
overlap
because
they're,
you
know
most
of
the
time
at
least
you
when
they
switch
to
central
European
summer
time.
We
move
one
hour
closer
together
and
then
about
two
weeks
later,
most
of
the
US
will
switch
and
then
we're
back
to
the
same
time.
You
know
time
difference.
A
Yeah
I
guess,
if
Poland
should
not
make
sense.
A
8
AM
is
probably
the
earliest
you're
going
to
get
out
of
me
and
not
on
Tuesday,
but.
E
A
All
right
so
pull
pull
incoming
yeah.
Anything
we
can
do
to
accommodate
people
that
want
to
show
up
I
think
you
should
try.
A
C
C
A
All
right
do
we
need
to
release
instrumentation
on.
E
I'd
like
to
try
to
see
if
we're
about
getting
that
Sinatra
fix
merged
and
then
I
will
really
select
a
release
process
now
because
of
Ruby
three
two,
not
a
you,
know
being
very
strict
about
how
it
works
with
dependency
declarations
right.
A
So
this
PR
just
to
revisit
it
and
summarize
we're
just
letting
this
exception.
Bubble
Up
and
the
rack,
the
Iraq
instrumentation,
will
actually
rescue
and
record
the
error
event,
and
then
we
only
get
it
recorded
once
not
twice.
Is
that
hey.
C
Yeah
sorry
I'm
just
catching
up
on
the
combo
and
the
right
in
the
your
okay
right,
okay,
cool
that
guy
was
on
an
outdated
clone
that
you
can
disregard.
Someone
was
trying
to
double
check
it.
It
was
like
it's
not
working
and
they
hadn't
bumped.
The
version.
A
Oh
and
we
also
got
Robert
so
all
right.
This
looks
like
it's
ready
to
go,
but
actually,
unless
Eric.
C
I'll,
just
I,
if
I
need
to,
if
my
boss
tells
me
I
need
to
like
get
judged
in
my
performance
reveal
a
number
of
stars
I'll.
Just
after
the
fact,
you
know
get
everything.
Oh
after
the
fact.
Just
you
can
approve
things
after
effects.
E
I
guess
I
just
got
one
more
topic,
because
this
is
my
is
my
television
show
I,
guess
posted
it
in
the
chat
and
zoom
instead
of
on
the
agenda,
but
I'm
gonna.
Add
it
here
to
the
agenda.
Officially
now.
E
Okay,
so
for
as
notification,
migration
work,
I
guess
so
we
had
talked
about.
We
were
kind
of
like
up
in
the
air
because
you
know
I,
guess
I'll.
We
don't
have
enough
information
to
make
decisions
about
like
how
we
want
to
implement
or-
or
you
know
like
what
conventions
we
want
to
adopt
Etc.
What
do
we
want
to
name
the
spans?
E
What
do
we
want
to
do
about
backward
compatibility?
So
I
put
together
a
spreadsheet
there,
whatever
cool
sheet-
and
this
is
my
hope
at-
like
us-
cataloging
things
like
going
through
looking
at
the
existing
notifications
that
are
documented
on
the
in
the
rails
guides
and
then
coming
up
with
a
mapping
guide.
So
if
you
look
at
the
spreadsheet.
E
E
So
daddy
would
be
like
you
know,
for
each
notification
that
we
want
to
record,
or
we
want
to
instrument
that
we
would
have
a
different
tab
on
here
and
then
fill
out
details
in
there.
Everyone's
help
would
be
greatly
appreciated,
but
for
the
most
part,
I
was
like
I
started
like
brainstorming,
okay.
Well,
what
could
a
particular
key
map
to
for
semantic
conventions
right
or
what
would
we
consider
like
internal
conventions
to
our?
E
You
know,
framework
specific
things,
and
this
will
hopefully
we
can
lay
out.
We
can
probably
take
some
inspiration
or
see
whatever
it
is
that
other
vendor
specific
things
have
done.
E
You
know
tough,
avoid
Reinventing
the
wheel,
but
I
kind
of
feel,
like
we've
got
a
map
to
use
to
you
know:
Hotel,
seven,
comma,
obviously
right
so
simple
example.
Here
brings
up.
I
won't,
say
simple,
because
this
also
has
some
Nuance
I
should
have
I
shouldn't
have
said
that
what
seems
simple
is
like
hey,
look.
E
Here's
the
SQL
key
and
we've
talked
about
this
before,
but
you
have
a
d
that
would
map
to
DV
statement,
for
example,
but
there's
nothing
in
the
specification
that
says
how
you
should
do
sanitization
whether
you
should
do
sensitization
at
all.
This
same
attribute
would
be
recorded
Again
by
the
database
driver
instrumentation.
E
So
it's
part
like
there
might
be
some
potential
thought
process
around
around
kind
of
like
Hey.
How
do
we
want
to
do
this?
Do
we
want
to
enrich
a
database
driver?
Do
we
want
to
turn
off
specific
attributes
by
default
when
we
know
that
there's
an
automatic
instrumentation
that
has
those
attributes,
do
we
you
know,
do
we
wanted
to
make
it
for
both
I,
don't
know?
That's.
E
To
get
that's
where
it
starts
to
the
conversation
starts
to
get
tricky
there.
I
don't
want
to
resolve
that
right
now,
but
you
know
that
that's
kind
of
one
example
of
a
weird
instance
the
the
next
one
being
like
Oh
well,
this
is
cool,
because
this
is
this.
B
Yeah
I
think
that
would
actually
let
us
keep
pretty
close
to
backwards
compatibility
and
still
feel
right
actually
like
I,
like
the
fact
that
we
are
saying,
like
user
load
or
user
save
for
those
spans,
because
it's
not
just
a
database
operation.
It's
a
higher
level
thing
I
didn't
realize
that
and
it's
wild,
because
it's
right
there
staring
me
in
the
face
and
I
really
should
have
like.
That
was
my
main
concern
about
active
record.
B
Changing
would
be
that
we
would
lose
that
nice
band
naming,
but
you
know,
maybe
we
won't
so
yeah
I,
think
that
would
actually
make
it
a
lot
easier
as
far
as
like
the
SQL
stuff,
like
I,
think
we
should
probably
add
it
as
a
span
attribute
for
now
and
then
decide
how
we
want
to
suppress
instrumentation
separately,
like
we
kind
of
have
the
same
problem
with
some
of
the
HTTP
based
information
instrumentations,
and
they
had
some
ideas
and
strategies
that
we
can
use.
B
But
you
know
I,
don't
know
if
we
have
anything
yeah
I
feel
like.
That
would
actually
be
something
we
could
tackle
after
we
cleaned
up
some
of
the
like
common
database,
stuff
that
we
do
across
various
instrumentations,
like
any
suppression
mechanism,
we
might
create
we'd,
probably
want
to
create
for
all
of
them
as
a
part
of
that
project,
but
yeah,
sorry,
that
was
my
ramblings.
E
It's
not
helpful
that
the
trying
to
decipher
what
the
binds
and
like
the
Typecast,
so
those
are
unuseful-
those
are
useless.
Cash
Cash
might
be
interesting
right.
E
Yeah
but
but
there
isn't
a
semantic
convention
that
maps
to
it,
so
these
would
be
cases
where
it's
like
well.
Do
we
come
up
with
our
own
conventions
for
the
rails,
specific
Library,
stuff.
B
I
think
we
do
I
mean
we
do
that
for
active
job,
as
it
is
I
forget
what
the
status
is
of
that
proposal
to
allow
languages
to
have
like
language
specific
conventions
for
things
that
are
just
not
applicable
broadly,
but
it
feels
like
it
would
fall
in
that
category
there
until
they
come
up
with
semantic
conventions
for
orms,
specifically
yeah
I.
Don't
know,
I'd
be
I'd,
be
fine
with,
like
nesting,
that
under
like
activerecord.statement.cashed
or
something
but.
A
Yeah
I
think
Eric
Eric
points
out
that
that's
useful,
because
those
will
take
essentially
no
time
and
if
hopefully,
yeah
and
depending
on
how
you
are
tracking
the
stuff
in
your
basic
back
end
like
you
might,
if
you're,
just
looking
at
like
an
average
time
for
for
your,
you
know,
for
your
selects
then,
like
these
caches
are
going
to
dramatically
alter
that
time.
So
I
think
knowing
to
look
pretty
cached
and
uncached
would
be
an
important
thing.
A
A
So
looking
at
this
I
want
to
know
what
happens
if
you
have
select
posts
from
Posts
where
ID
equal
blah?
If
you
end
up
with
a
question,
mark
ID
equals
question
mark
and
then
the
question
mark
is
filled
in
with
binds,
because
I
seem
to
recall
this
being
the
situation
for
most
active
record
queries,
in
which
case
I
feel
like
you're
SQL,
might
already
be
obfuscated
in
that
case
and
you're
actual
values
might
be
in
in
the
binds.
E
A
Yeah
that
that
may
be
true
driver
might
be
a
variable,
but
something.
D
A
Think
to
be
worth
some
Discovery
well
well,
while
looking
at
this.
A
A
It
would
be
nice
to
see
what
the
recommendation
is
for
for
database
span
names
in
the
spec
I
haven't
been
in
there
in
a
little
while,
but
yeah
they're
they're
there
could
be
a
mapping
of
name
I
guess.
Ultimately,
if.
E
A
B
B
E
A
E
You
know
consider
that
Java
also
has
like
a
a
grammar
Alexa
parser
for
sanitization
right,
so
they
parse
the
entire
tree
and
sanitize
the
SQL
and
I'm
trying
to
use
that
as
a
model
for
standard
SQL,
parsing,
there's
like
a
flex
Java
thing.
So
what
you're
probably
looking
at
here
is
more
of
an
extension,
more
extensious
fast,
specifically
for
storage
procedures.
E
You
know
I,
guess
the
question
so
so
there's
like
the
question
that
we're
trying
to
answer
by
kind
of
you
know
thumbing
through
this
code
here
is:
do
they
do
they
have
representations
for
hibernate
specific
things
like
the
hibernate
like
like,
let's
say
you
have
a
repository
class,
do
they
have
or
for
the
you
know,
for
the
hibernate
query
language?
Do
they
have
some
sort
of
representation
of
internal
spans
to
the
orm
like
I?
E
E
You
know
like
trying
to
I
kind
of
feel
like
we.
We
want
to
probably
make
a
choice
of
saying
like
do.
We
want
fine
grain
or
RM
instrumentation
with
details
about
what
the
orm
is
doing
and
are
we
enriching
SQL
database
or
database
bands
with
things
that
the
orm
is
doing
or
do
we
treat
the
orm
as
its
own
internal
instrumentation
and
leave
the
database
drivers
to
be
their
own?
You
know
be
the
things
that
do
the
client
spans
I.
Guess
right.
That's
the
question
that
we're
trying
to
answer
here.
A
A
D
B
A
Case
I
guess
maybe
we
can
dig
through
this
and
try
to
answer
some
questions,
probably
not
in
real
time.
A
At
the
end
of
the
day,
I
think
there's
there's
going
to
be
some
prior
art
that
we
can
look
at
in
other
repos.
So.
A
Java
would
be
one
I
think
python
must
have
some
orms
that
we
can
take
a
look
at
as
well
and
just
try
to
get
an
idea
of
what's
already
happening
and
then
work
our
way
backwards
from
there
see
if
there's
like
any
spec
to
kind
of
justify,
what's
happening
in
in
the
various
languages,
and
if
there
is
great,
if
not-
and
we
have
open
questions,
I
think
we
can
start
to
raise
some
issues
at
the
spec
level
about
how
exactly
to
model
Telemetry
for
an
orm
versus
like
an
actual
database
client.
A
A
C
I
mean
I
I
left
a
minor
comment
on
the
issue
that
you
know
I'm
not
super
concerned
about,
because,
like
Andrew
pointed
out
like
the
name
is
there,
it
says,
love,
cardinality
and
yeah
is
the
the.
This
is
the
only
real
blocker
to
kind
of
moving
everything
to
you
know
ASM.
If
I
understand
correctly,
it's
like.
E
There's
one
one
problem
with
the
ACT
there's
a
few
problems
with
the
existing
instrumentation
one.
Is
you
can't
customize
a
span
name
easily,
all
right
and
I
think
yeah
I
think
that's
it
really.
E
C
Saying
the
ASN
instrumentation,
it's
already
customized
right
and
we.
C
You
you
can't
just
like
Supply
override
a
a
public
method
that
provided
to
spending
or
something
yeah.
It's
good
I
think
this
is
great
to
move
forward,
though
so
I'm
just
not
able
to
contribute
actively,
but
you're
gonna
be
a
big
win
and
lower
the
maintenance
version.
You
know
it's
a
you
know
over
time
tremendously,
so
yeah,
good
stuff,
man,
sorry
I'm,
sorry,
I'm,
not
more
helpful.
A
So
would
an
ideal
end
state
for
this
spreadsheet
contain
tabs
for
more
than
just
active
record
like
the
action
controller
business?
Do
we
do?
We
want
active
support
notifications
for.
A
Or
those
those
events
as
well.
E
Oh
yeah
I
think
we
want
to
migrate
everything
and
then
start
identifying
things
that
we
haven't
done
already
so
Andrew
pointed
out
like
we
don't
have
one
for
Action
mailer
or
action,
cable
or
any
of
those
like.
We
certainly
want
to
expand
on
our
offerings.
There.
A
E
A
Like
it
would
be
nice
to
identify
at
least
like
at
a
high
level,
what
what
notifications
we're
interested
in
and
then
kind
of
start
to
Noodle
on
a
mapping
like
you,
are
and
then
figure
out
how
that
stuff
dovetails
with
existing
semantic
conventions
and
and
what
holes
there
are
for
semantic
conventions
as
well
and
go
from
there.
C
E
E
D
B
That's
a
good
reference.
I
gotta
drop
I've
got
another
meeting,
but
it
was
good
talking
to
y'all
again
yep
bye.
C
No
yeah,
Ariel
I
know
my
responses
have
been
awful
but
again,
like
you,
know,
I'm
not
working
on
Civic
right
now,
but
if
you
do
need,
you
know
feel
free
to
ping
me
your
DM.
If
you're
like
stuck
on
a
you
know,
you
need
cursory
reviews
or
you
do.
500
depend
about
PR
to
get
or
whatever
things
I'm
happy
to
open
my
mind
away.
But
that's
like
the
extent
of
my
trying
to
contribute
the
moment
of
which
yeah.
A
Yeah
I
I
I
can
extend
that
as
well.
If
you
just
have
like
a
a
benign
version,
bump
or
kind
of
release,
PR
that
you
need
some
eyes
on
and
nobody's
looking
at,
it
feel
feel
free
to
reach
out
I
I
check
in
on
cncs
lack
only
periodically
and.
A
It
had
already
been
done
and
maybe
I
actually
catch
it
in
a
reasonable
time
and
try
to
do
something.
But.