►
From YouTube: 2022-03-29 meeting
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
C
A
A
Okay,
so
there
are
few
agenda
items
we
need
to
discuss,
but
before
we
go
into
any
of
those
things,
I
will
just
give
a
quick
update
on
two
things.
One
is,
I
would
want
to
let
leave
in
around
15
minutes,
so
I'll
leave
early,
so
you
can
continue
without
me,
but
I'll
be
like
working
later.
So
my
main
goal
is
to
ship
the
next
rc
as
soon
as
possible.
A
We
had
some
breaking
changes
in
the
last
one,
so
this
one
really
addresses
that,
and
on
top
of
that
we
also
address
the
activity
status
for
all
three
exporters.
All
four
exporters
so
I'll
be
like
doing
a
release
once
I'm
back
later
today,
so
that
will
not
address
the
the
view
specific
issue
which
alan
and
I
have
discussed.
Last
week
there
were
like
some
discussion
in
the
spec
also
about
it,
but
like
we
consider
it's
important,
but
you
don't,
it
doesn't
have
to
block
the
entire
release
campaign.
A
So
that's
two
quick
update
on
the
next
rc
release
and
as
soon
as
I
am
done
with
the
rc
release,
I
will
go
ahead
and
re-update
the
milestone
because
it
looks
like
it
if
you're
doing
an
rc
today
and
once
we
address
the
view
thing,
we
can
do
like
another
rc
like
early
next
week,
then
I'll
give
like
at
least
two
three
days
before
we
came
on
after
the
last
rc,
because
there
might
be
like
some
folks
who
are
extending
the
sdk
but
they're
getting
their
own
exporters.
A
And
there
are,
I
mean
I
want
to
give
them
at
least
one
or
two
day
just
to
report
that
if
they
see
anything
broken
with
the
latest
one
just
being
on
the
safe
side,
there
is
no.
I
mean
we
already
have
like
enough
checks
to
make
sure
there
are
no
breaking
changes
or
any
stability
issues,
but
just
playing
a
little
bit.
So
it's
very
unlikely
that
1.2
will
land
on
the
1st
of
april.
A
It
would
be
quite
more
likely
on
8th,
that's
the
current
plan,
so
I
will
update
it
as
soon
as
I'm
done
with
the
current
release.
Yeah,
that's
all
from
the
release
site
load.
Now,
let's
look
at
the
topics
from
agenda.
B
Added
or
yeah,
these
are
all
things
I
added.
I
think
most
of
them
will
probably
be
relatively
short.
B
This
was
just
spec
pr
that
I
was
looking
at
yesterday
regarding
instrumentation
stability
and
there's
a
statement
in
there
saying,
basically
that
all
instrumentation
provided
by
all
open,
telemetry
repositories
in
order
for
them
to
be
deemed
stable,
which
I
think
for
our
purposes
would
mean
you
know
that
it
would
have
like
a
a
non-like
release,
candidate,
beta
designation
or
something
that
it
conformed
to
the
semantic
inventions
and
that
it
does
not
change
anything
in
a
breaking
way
once
deemed
stable.
D
A
So
I
think
what
we
are
asking
is
like
if
a
particular
component
is
hosted
in
open
elementary
on
repo,
which
includes
a
country
one
who
is
responsible
for
ensuring
that
they
follow
the
guidance.
Whatever
is
the
guidance
on
telemetry
stability
as
required
here?
Is
it
on
the
maintainers
of
that
language,
or
is
it
really
on
the
individual
people
who
contribute
the
in
the
packages?
A
Is
that
the
like
top
level
question,
or
do
we
need
to
like
put
like
some
process
to
ensure
that,
even
if
we
decide
to
okay
make
it
the
problem
of
the
individual
contributors?
How
do
we
ensure
that
they
are
like
following
it,
and
is
that,
like
the
high
level.
B
A
Not
much
the
only
thing
which
they
get
is
they
will
be
notified
when
there
is
a
pr,
that's
it
even
if
they
approve
it,
it
doesn't
do
the
tick
mark.
So
when
you
do
an
approval,
you
look
that
tick
mark
won't
be
there,
so
it
will
be.
Look
like
this
tick
mark,
so
someone
like
who
is
in
the
approval
or
maintainer
has
to
approve
it
before
it
can
get
the
actual
green
light.
A
One
good
thing
so
far
is
like
almost
all
the
components
in
the
contributor
instrumentation
libraries.
So
we
have,
we
haven't,
released
anything
as
stable,
so
so
far
like
we
are
still
in
complaints
and
as
long
as
we
don't
do
a
stable
release,
we
are
still
complaining
because
this
requirement
is
only
saying:
if
you
do
a
stable
release,
then
you
have
to
adhere
by
this
policy.
But
as
long
as
we
are
not
stable,
we
are
fine.
So.
A
I
did
see
this
pr
like
in
november.
I
the
that's
the
main
reason
why,
like
force.net
to
support
instrumentation
url
in
dotnet,
seven,
they
haven't
like
really
approved
it
yet,
but
the
latest
status
on
that,
like
supporting
instrumentation
library,
sorry,
the
instrumentation
schema
url
for
meter
and
activity
source.
A
We
said
we'll
come
back
late
march
to
april
first
week,
which
is
like
this
week,
because
dotnet
wants
to
like
know
that
this
is
a
really
really
stable
thing,
even
it's
part
of
the
stable
for
like
long
time,
but
there
were
no
like
well-defined
use
cases
on
how
how
would
someone
use
the
telemetry
schema?
So
dotnet
was
really
waiting
to
see
and
this
pr
I
send
this
pr
in
the
github.
A
This
is
what
we
intend
to
do
once
we
have
it.
So
it's!
Basically,
if
you
don't,
have
the
capability
to
store
the
schema
url
in
the
meter
and
activity
source.
There
is
no
way
you
can
really
do
a
stable,
instrumentation
library,
because
it's
kind
of
prohibited
as
per
this
one,
so
yeah.
So
I
think
the
first
step,
like
immediate
step,
which
it's
already
done
like
from
the
dotnet
time,
is
support.
A
It
in
the
activity,
source
and
meter,
but
then
like
regarding
ownership
like
who
would
control
it
or
who
would
make
sure
things
are
good
in
contrib.
That's
something
which
I
guess
we
can
come
back
like
slightly
later
as
well,
because
it's
not
a
problem
which
affects
dot
net.
So
I
expect
there
will
be
whatever
solution
you
do
would
be
like
somewhat
consistent
across
all
the
six.
B
Yeah,
I
think
that's
fair
and
also
knowing
that
you
know
it's
going
to
be
somebody
from
approvers
or
maintainers
of
this
repository
or
the
contrib
repository
yeah.
So.
A
A
Yeah,
so
that's
the
good
thing
about
it
like,
but
that's
also
a
bad
thing,
which
means
like
you
are
ultimately
like.
If
you're
an
approver
or
a
maintainer,
you
are
ultimately
responsible
for
it.
So
the
good
thing
is
like
we
cannot
just
like.
We
don't
like
really
allow
like
anyone
trying
them.
A
Yeah,
but
I
think
this
is
an
important
pr
like
for
like
one
step
towards,
I
mean
other
step
beings,
many
conventions
being
actually
stable,
but
this
is
actually
like
blocker.
Once
semantic
convention
itself
becomes
stable
and
for
dotnet
case
we
have
a
third
requirement
that
dotnet
runtime
should
support
the
schema
url,
which
is
quite
likely
to
happen
like
it's
just
that
they
are
waiting
to
see
some
genuine
use
case
of
how
we
would
use
this
before
they
are
it
into
the
runtime.
A
Okay,
yeah.
I
think
this
is
to
be
approved
by
every
language
maintainer.
So
thanks
for
remaining
again,
I
did
look
at
it
like
few
months
back,
but
now
that
tigran
was
asking
everyone
to
review
it.
We'll
do
another
look
at
it.
A
Okay,
I
think
I
should
have
time
for
this.
One
too.
You
can
continue
without
me
and
if
I
go
so
just
letting
you
know
that
I'll
be
leaving
it
around
420
or
slightly
before
that.
B
So
yeah,
I
added
I
added
this
to
the
agenda
as
well.
So
so
you
know,
tyler
is
actually
a
new
employee
at
new
relic
he's
on
my
team
and
we
recently
had
a
customer
or
have
had
customers
in
a
variety
of
different
languages
needing
to
adjust
their
attribute
limits
because
new
relic
at
least,
and
to
my
understanding,
some
other
vendors
as
well.
B
Have
some
restrictions
upon.
You
know
ingest
for
the
number
of
attributes
on
spans
and
like
the
attribute
lengths
and
so
on.
So
some
of
the
languages
have
that
this
piece
of
the
spec
implemented.
I
understand
that
you
know
net.
This
would,
of
course,
be
it's
an
api
concern,
so
I
guess
two
parts
to
this.
I
I
I
couldn't
remember
if
this
was
a
thing
that
had
already
been
kind
of
raised
to
the
level
of
like
hey.
You
know,
maybe
consider
this
for
net
seven
and
then
two.
B
My
the
next
thing
is:
is
there
something
that
we
can
do
in
the
near
term?
Like
maybe
deliver
like
a
processor,
maybe
in
the
contribution?
Maybe
in
that
preview
package
that.
D
B
Out
there
some
idea
like
that
to
meet
the
current
need
yeah.
So
thank
you.
A
This
was
not
considered
for
dotnet
seven,
the
list
of
items
which
was
even
in
the
dot
net
seven
planet
in
the
start
here.
So
if
we
see
like
whatever
things
which
we
proposed
got
here,
so
this
is
something
which
we
never
asked
for,
so
it
was
not
in
the
track.
So
that
means
like
it's
very
unlikely
to
be
included,
possibility
it's
like
still
april.
We
have
opportunity
till
june,
but
it's
it
needs
to
be
like
really
like
well-defined
api,
I
mean
if
it's.
A
If
the
spec
is
like
well
defined,
we
we
can
still
make
an
atom.
So
that's
a
like
you
could
cancel
the
first
part.
Second
is
there?
Are,
I
think
two
things
we
can
do
like
number?
One
is
the
processor
approach,
but
even
better
would
be
the
shim.
The
open
elementary
api
has
expansion,
so
we
should
be
able
to
enforce
that
in
darshan,
but
it
won't
really
help
anyone
like
unless
you
are
using
the
shim
itself.
A
So
maybe
like
it's
a
good
thing
for
trying
it
out
in
the
shrimp
that
will
like
really
convince
dot
and
that
this
is
like
I
mean
I'm
still
like
not
sure
like.
What's
the
best
way
to
implement
it
because,
like
people
can
like
set
tag
called
set
tag
like
now,
someone
has
to
inside
the
set
tag,
implementation
check.
How
many
are
we
added
and
based
on
that
start,
dropping
it
or
maybe
like
just
allow
any
number
of
things
to
be
added,
but
at
the
exporter
side
someone
has
to
do
so.
A
I
think
that
those
things
are
not
yet
clear,
like
what's
the
best
way.
So
without
like
a
proper
thing,
I
cannot
just
go
and
add
it
here
I
mean
I
can
of
course
ask
but
they'll
definitely
ask
like
is
what
is
the
best
way
to
do
it
or
another
option
is
like
whatnot
will
give
a
call
back
on
set
tag
just
like
activity
start.
So
then
it's
on
the
sdk
side
to
keep
track
of
how
many
tags
and
then
do
something
with
that
callback.
A
Yeah,
so
do
you
want
it
to
be
something
which
we
want
to
like
start
considering
4.7,
or
do
you
think
we
should
be
able
to
leave
with
like
any
workarounds,
whatever
beta?
I
I
just
don't
know
like
what's
that
perfect
workaround,
but
what?
What
is
it?
What
what
is
your
qr
plan
like?
Do
you
want
it
to
be
like
opened
as
an
issue
in
the
tottenham
report
to
see?
A
B
So
yeah
I
mean
I,
I
think
the
idea
would
be
that
it
was.
It
would
be
land
in
the
runtime,
ideally
with
0.97,
but
I
do
like
your
thought
of
the
shim.
I
hadn't
thought
about
that,
so
if
you
think
that
it
would
help
to
maybe
at
least
land
like
that
as
a
prayerful
concept,
first,
in
order
to
to
drive
that
conversation
with
the
runtime
team,
I'd
be
okay
with
that
approach
and.
A
But
the
reason
why
I
want
to
at
least
play
with
it
is
because
this
is
something
which
will
affect
the
like
actual
set
tag:
api
or
it
doesn't
like.
We
choose
to
do
it
like
index
folder.
Then
it
only
affects
exporter,
but
that's
the
part
where
I'm
not
sure
like.
Where
would
we
handle
it?
So
maybe
a
good
candidate
would
be
to
write
out
in
the
telemetry
span
the
wrapper
around
activity.
A
B
Yeah,
so
I've
talked
with
him
a
little
bit
and
we
we
discussed
the
idea
of
a
a
custom
processor
that
we
deliver,
but
yeah
I
mean
that
that's
less
than
ideal,
just
in
the
sense
that
it
won't
be
performing
right,
like
tags
are
going
to
get
added,
and
then
this
processor,
at
the
final
end
of
the
pipeline,
would
be
doing
all
this
work.
To
like
make
truncate
things
or
or
remove
attributes
or
whatever.
D
Alan
just
to
mention
I
maybe
I'm
mistaking
I'm,
but
quite
sometimes
without
look
at
the
collector,
but
I
think
the
collector
had
the
capability
of
truncating
fields,
but
I
don't
remember
if
they
had
kind
of
in
general
or
kind
of
you
have
to
match
specific
kind
of.
Oh.
If
I
match
I
truncate
this,
you
know
I
I
don't
remember
the
rules
for
that,
so
that
collector
may
have.
But
if
you're
not
using
the
collector,
that's
not
our
option
right.
A
A
So
yeah
and.
B
We
definitely
have
customers
that
are
not
using
the
collector
but
yeah
paulo.
I
mean
that
that's
a
that
is.
That
is
what
we
I
don't
remember,
either
what
the
exact
configuration
options
are,
but
there's
an
attribute
filter,
processor
for
the
collector
and
so
on.
That
enables
a
lot
of
that
functionality
and
so
yeah.
We
we
guide,
customers
towards
towards
solutions
like
that:
okay,
yeah.
A
So
maybe
do
you
think
that,
as
a
starting
point,
we
should
try
to
add
it
in
the
like
the
span
wrapper
and
then
take
it
from
there
or
do
we
want
to
parallely
start
a
conversation
with
the
dotnet
team
by
creating
an
issue
like
somewhere.
Here
they
come
in
door
like
create
a
new
issue
and
like
come
in
here.
A
So
we
can
see
if
there
is
any.
It
looks
like
very
unlikely,
because
it's
already,
we
had
too
many
items
and
it
got
like
spilled
to
8
plus,
but
I
mean
nothing
wrong
in
like
making
an
attempt
right
now.
I'm
trying
to
reply
like
why
we
missed
it
because
maybe
like
this
was
just
a
shoot
requirement.
Not
a
must.
So
that's
why
we
never
like
feel
very
close
attention
and
it
was
the
first
time
like
I
felt
like
someone
really
asking
for
it.
A
You
only
need
to
do
that
if
you
really
want
to
consider
it
in
the
document.
Otherwise
we
can
just
wait
for
like
some
prototyping
and
then
so
I
let
you
like
figure
out
whether
it's
important
with
that
said.
I
will
leave
right
now,
so
I
will
stop
sharing.
I
learned
if
you
want
to
continue,
please
go
ahead
anyway
and
we'll
start
calling
yeah.
Thank
you
folks.
Bye-Bye.
B
On
the
agenda
that
maybe
some
of
you
might
have
some
thoughts
on
this.
B
Pr
I'd
kind
of
forgotten
to
get
back
to
you,
but
basically
grpc
net
client
was
recently
backboarded
to.
I
can
actually
share
my
screen.
B
Yeah
grpc
net
client
has
recently
back
part
of
the
net
standard
2o,
which
enables
it
to
work
in
windows
environments.
So
that's
cool.
However,
I've
never
personally
tested
it
in
in
that
kind
of
a
context,
and
while
I
think
it's
probably
pretty
safe
too,
basically
all
this
guys
pr
is
doing
at
this
point
in
time
is
just
introducing
the
net
standard.
2O.
B
Well,
our
tests
currently
run
against
I
think.net5.net,
six
and
and
net
core
at
3.1,
and
all
of
those
flavors
of
the
tests
are
going
to
go
and
grab
the
net
standard.
2
1
binary
of
the
actual
library
that
this
is
instrumenting.
B
B
B
C
C
Sorry,
I
said
again
yeah
we're
going
to
work.
This
is
like
so
like
you're,
saying,
net
standard
2.0
support
is
now
available,
so
like
net
framework
applications
can
also
use
the
grpc
net
client
instrumentation
yeah
I
mean
I
would
say
I
think
we
should
be
adding
unit
tests
for
framework
apps
as
well.
Just.
B
Literally
right
before
this
meeting,
I
put
out
a
this
guy
notes
that
all
the
tests
that
we
currently
have
against
this
library
are
spin
up
an
asp.net
core
server
that
actually
stands
up.
A
legit
grpc
endpoint
that
the
the
client
communicates
with,
in
part
that's
to
test,
not
just
the
grpc
instrumentation,
but
also
that
it
invokes
the
underlying
http
client
instrumentation.
B
D
B
Like
in
the
context.net
framework,
but
I
think
that
this
might
be
sufficient,
I
mean
at
the
end
of
the
day
we
can
probably
continue
this
conversation
in
the
inside
of
this
pr.
Once
I
get
it
cleaned
up
a
little
bit
more
because
there's
quite
a
bit
of
forklifted
code
that
I
think
I
can
trim
down
that.
I
took
in
from
the
actual
grpc
library
tests.
B
But
basically
the
test
now
is
just
a
a
simple
test
and
it
uses
all
that
forklifted
code
to
basically
create
a
mock,
http
client
that
communicates
as
though
it
was
like
an
http
2
client
in
the
ways
that
the
the
grpc
client
expects
and
this
test
actually
passes.
At
least
in
the
context
of
you,
know:
net
diamond
five
and
done
six.
C
Yeah
I
mean-
I
think
this
so
you're
saying
we
could
have
just
this
mocking
thing
be
available
for
the
net
framework
tests
as
well
right
yeah.
I
think
that
should
be
good
enough,
at
least
for
start
to
start
up.
B
Yeah
and
I'll
test
that
out
next,
but
I
guess
ultimately,
the
question
will
be
like
the
other
test
that
that
test,
like
the
underlying
actual
like
http
client
instrumentation,
actually
gets
invoked
too.
That
is
a
that's.
B
I
mean
it's
anything's
possible,
but
that
that's
a
that's
another
leap
in
complexity
of
like
making
this
work
in
the
context
of
net
framework,
in
the
sense
that,
like
you,
can't
use
like
that
asp.net
core
server
very
easily,
you
gotta,
like
maybe,
do
an
integration
test
and
like
spin
up
spin
up
a
server
on
the
attacker,
or
something
like
that.
So
I
haven't
come
that
far.
C
B
B
Does
a
grpc
call
and
this
whole
suite
of
tests
spins
up
a
grpc
server
that
actually
you
know
that
listens
on
some
port,
that's
constructed
like
when
the
tests
are
initialized
and
then
ultimately
the
we
validate
that
both
grpc
and
http
client
spans
are
generated.
B
B
So
this
is
the
part
that
would
be
like
this
test
would
in
particular,
would
be
a
little
bit
trickier
to
do
in
the
context.net
framework,
because
it
requires
a
legit,
http
client
versus
that
mocked
out.
One
that
I
was
showing
you
a
second
ago.
C
B
Yeah
I'll
get
that
cleaned
up
and
and
open
it
up
and
see
what
see?
What
folks
think
I
mainly
wanted
to
kind
of
get
this
moving,
because
this
is
actually
the
second
individual
who
has
attempted
to
introduce
the
net
standard
2o
target
the
the
first,
the
first
person
just
kind
of
went
m.I.a
after.
B
I
want
to
hopefully
use
this
as
an
opportunity
to
show
this
person
some
some
attention,
yeah
and
so
yeah.
Thanks.
B
D
D
It's
no
big
deal,
but
the
the
thing
is
right
now
it
doesn't
work
out
of
the
box
for
the
faults
for
http
protobuf,
so,
if
possible,
like
kind
of
get
the
sig
to
take
a
a
a
look
at
this
vr.
You
know.
D
So
basically
there
was
in
the
derived
types
for
metrics
and
trace.
There
was
kind
of
similar
code
doing
that
check.
I
kind
of
moved
to
the
base.
I
didn't
the
log
is
following
a
different
pattern.
I
didn't
touch
the
log
for
this
time,
but
basically
this.
Basically,
this
review
is
moving
a
bunch
of
code
to
the
base
class.
B
B
D
Yeah,
so
my
understanding
is
because
the
options
actually
apply
to
all
of
them.
They
don't
have
the
path,
so,
for
instance,
not
only
because
of
the
signal
can
be
different,
but
also
because
the
default
is
grpc.
For
now
we
have
this
kind
of
extra
handling
for
the
http
protobuf
and
what
I
try
to
do
is
kind
of
ensure
that
works
out
of
the
box
and
in
the
process.
I
end
up
kind
of
reorganize,
a
bit
the
cool.
B
D
Yeah,
so
we
don't,
of
course
we
don't
need
to
do
it
right
now
in
the
meeting,
but
just
I'd
like
if
the
sig
has
some
cycles
to
give
a
little
bit
of
priority
to
this
one.
B
Yep,
I
can
definitely
review
it
more
closely.
D
Yeah
yeah
one
thing,
but
that
I
I
actually
started
testing
this
on
the
first
interaction.
I
was
testing
on
the
integration
because
I
I
kind
of
wanted
to
have
a
test
out
of
the
box
kind
of
okay.
They
just
set
the
http
protobuf
as
the
protocol,
but
what
I
noticed
is
that
the
test,
actually
the
integration
test,
uses
the
docker
compose.
So
I
can't
do
the
integration
test
with
the
defaults
because
the
default
is
localhost.
D
D
Perhaps
I
I
can
do
in
a
separate
pr,
but
in
the
same
file
instead
of
I'll
keep
the
tests
as
they
are,
but
also
we'll
have
the
test
just
a
single
test,
running
from
the
the
not
from
a
docker
but
directly,
and
then
I
expose
the
same
collector,
that's
already
using
the
test,
I
suppose
the
parts
in
the
local
host.
D
So
then,
I
can
test
the
defaults
both
for
jrpc
and
http
protobuf.
B
D
Yeah
yeah
so,
but
the
thing
is,
if
we
run
the
hotel
collector
like
that,
but
run
the
test
local,
I
can
expose
the
port
right,
but
but
the
way
that
the
test
is
configured
both
are
running
in
a
darker
containers.
D
So
then,
in
front
of
a
docker
composer,
I
think,
but
anyway
they
they
see
hotel
collector.
Whatever
is
the
alias
that
was
giving
to
that
one,
not
the
local
host.
That
should
be
the
default
so
right
now
I
don't
have
a
test
that
actually
sends
data
with
the
defaults.
That's
the
the
part
that
I'm
kind
of
not
very
satisfied,
but
I
wasn't
sure
because
the
current
configuration
really
doesn't
allow
for
that.
You
know.