►
From YouTube: 2020-06-30 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).
B
B
B
B
We
continue
putting
at
an
open
source
and
then
once
people
are
ready
to
join,
we're,
always
happy
and
ready
to
do
these
movements
you
mentioned
about
like
making
it
formally
open,
telemetry,
so
I
think
we
shouldn't
wait
for
many
weeks,
I!
Think
right
now
it
like
you
mentioned
there
is
a
lot
of
alignment
and
consumption
suggest
get
it
done,
but
they.
B
As
we
discussed
in
the
smaller
meetings,
maybe
two
but
I
first
of
all,
I
need
to
confirm
this
officially
from
the
company,
because
this
you
know
this
is
a.
This
is
not
my
decision,
but
I
am
pretty
sure
we're
happy
to,
although
this
is
not
to
be
taken
as
an
official
commitment.
This
is
just
my
read
of
the
stance
from
people
and.
B
C
B
Name
is
yet
anytime,
but
for
us
to
do
the
work,
we
need
to
understand
that
people
are
actually
going
to
contribute
according
to
some
sort
of
roughly
agreed
roadmap,
because
it's
work
for
us
to
move
it
there
and
to
maintain
it
and
should
take
forth
back
and
forth,
and
we
want
to
understand
that
we're
doing
it
for
for
mutual
benefit.
So
once
we
have
said,
okay,
like
Fox
Fox
Blanc
would
like
to
add
this.
This
particular
interoperability,
and
that
makes
sense.
We
also
want
this.
We
have
briefly
chatted
about
the
architecture.
B
We
haven't
understood
that
this
is
happening
within
the
next
month,
whatever
able
to
say
okay,
forked
put
it
there
everything's
great,
because
we
now
do
it
for
mutual
benefit,
but
we're
not
gonna
just
played
for
the
secretly
save
the
Microsoft.
If
you
guys
have
went
through
internal
planning,
you
decided
which
road
are
you
taking?
You
would
like
to
cooperate.
You
say:
okay,
like
fermentation
is
important.
B
Angela
is
important
for
us
and
for
you
as
well,
but
for
it's
like
a
few
months
later
for
you
it's
a
few
months
earlier,
but
it's
important
for
both
of
us
and
we
have
a
joint
understanding
of
you
know
roughly
how
to
put
it
in
you're
going
to
make
the
contribution
sure.
Then
we
move
the
agent
in
Chopin,
telemetry
and
start
contributing
together,
so
definitely
for
making
it.
But
just
against
this
should
be
so.
D
C
D
B
D
Go
through
the
agenda
items
and
we'll
come
back
to
what
instrumentation
and
we
can
also
discuss
more
about
like
what
should
be
our
client
for
splitting
them
towards
the
ng.
Please
mark
your
attendance
here.
If
you
want
I
can
update
if
this
dog
should
be
like
editable
by
public,
so
everyone
should
be
able
to
put
it
if
anyone
I
think
it
should
be
like
public
I,
don't
own
this!
The
link
yeah
I,
it's
in
the
jitter
channels,
so
you're
in
digital
right.
So
it
should.
D
Go
through
the
items
in
agenda,
the
first
item
is
going
to
get
her
back
shines.
This
was
discussed
probably
a
month
and
a
half,
and
we
didn't
have
enough
bandwidth
to
tackle
this
at
that
time.
But
right
now,
Eddie
from
Microsoft
is
already
working
on
this.
We
have
already
validated
that
one
does
not
require
any
special
permissions
to
enable
this,
because
you
did
not
have
any
permissions
to
the
repo
and
he
was
able
to
like
submit
P
our
view
ourselves
and
everything.
D
So
there
is
a
discussion
already
about
like
the
benefits
of
moving
into
moving
into
github
actions,
and
there
is
an
overall
work-in-progress
guidance
from
open,
telemetry
community
that
we,
you
should
be
moving
under
cable.
You
repository
should
move
if
possible
to
get
have
action,
so
I
suggest
everyone
to
like
take
a
look
at
this
and
share
feedback.
D
It's
already
like,
like
clear,
is
already
there
where
we
are
running
both
github
actions
and
I
should
have
oops
pipeline
in
parallel
and
see
if
we
can
see
at
the
very
end
it's
already
running
and
anyone
can
see
the
results
either
from
the
PR
or
directly
by
going
into
github
actions.
Link
it's
fully
open,
so
there
is
no
need
to
navigate
to
assure
pipelines
or
anything.
It
should
be
right
here.
So
that's
one
thing,
and
there
are
few
open
questions
which
we
had
github
actions.
D
Oh
one
of
you
mentioned
about
this
code
coverage
some
time
back,
but
I
cannot
recollect
whether
what
context
it
was,
but
if
this
is
something
which
you
know
about,
please
do
help
us
to
move
to
deter
back
cents.
I
haven't
removed,
it
I
removed
a
sure
data
pipeline.
It's
still
there
like
once
this
PR
is
merged.
Then
we
change
the
like
criteria
for
chicken
to
be
based
on
github
actions,
and
we
can
then
duplicate
the
or
delete
for
pipeline.
D
We
won't
be
able
to
migrate
to
github
actions
until
and
unless
we
know
how
to
how
to
use
it.
For,
like
my
get,
publish
and
you
get
publish,
believe
Mike,
you
probably
know
more
about
that,
because
we
we
need,
like
some
secret
credential,
to
push
it
and
I,
don't
really
know
how
to
host
it
in
github
action.
So
probably
we
can
leave
it
as
it
is
in
the
show
prickling,
but
for
CA
purposes
we
should
be
able
to
completely
move
to
get
your
actions
yeah.
E
I
think
that's
fine,
I
think
here
we
need
to
take
some
API
key
from
my
guest,
so
it
can
push
till
our
feed
and
say
okay
hope,
but
yeah
I've
done
a
little
bit
of
code
coverage
in
the
past.
So
that
should
be
I've
got
a
couple
of
things:
I'm
using
I,
don't
know
how
well
it's
a
the
most
popular,
but
we
can
definitely
look
at
them
again:
okay,
yeah.
D
So
it
is
walking
on
back,
he
may
reach,
or
do
you
engage
it
or
if
he
has
some
questions
on
that?
If
you
know
like
other
references,
please
do
it
yeah
yeah
thanks.
So
next
topic
is:
we've
been
getting
like
quite
large
number
of
questions
in
the
either
this
issue
and
like
through
several
channels.
What
are
the
plans
for
beta
because
we
did
not
officially
release
a
beta
when
every
other
languages
did.
We
were
waiting
for
more
changes
intro
tonight,
so
it
looks
like
the
covered
pretty
much
this
item.
D
These
still
are
like
small
items
here
and
there.
But
overall
idea
is
we
completed
this
and
we
also
agreed
to
have
an
official
version
of
like
a
spam
based
wrapper
around
the
activity
as
well.
I
think
I
should
mark
the
stress
completed
here.
This
was
decided
that
we
will
be
publishing
it
so
and
actually
moving.
D
So
we
need
to
do
this
one,
but
I
would
like
to
use
like
next
two
weeks
to
publish
an
official
beta
and
give
like
customers
an
early
version
of
what
the
final
product
would
look
like
so
I'm,
putting
like
next
two
weeks
just
to
reselect
the
original
battle
readiness
tracker,
which
we
had
from
like
four
brewery,
which
Mike
was
tracking,
but
then
we
kind
of
drifted
to
get
open,
telemetry
and
that
integration.
So
now
that
that
part
is
like
mostly
done,
I
would
resurrect
this.
D
Some
of
them
are
I
think
mostly
they
are
a
crate,
but
I'll
use
the
guidelines
from
open
elementary
beta
and
see
if
he
needs
some
changes
here
and
I
will
start
tracking
things
here.
So
I
expect
we
should
be
doing
a
beta
release
in
the
next
two
weeks
in
that
doing
more
pretty
much.
What
is
expected
of
a
beta
from
open
elementary
standards
so.
A
D
So,
for
me,
tree
for
matrix,
I
was
planning
to
release
what
or
we
have
currently
assist
the
APA
for
matrix
at
the
time
of
original
planning
is
complete.
The
SDK
is
not
complete,
but
the,
but
the
G
the
beta
guideline
says:
implement
everything
which
is
SPECT
and
since
the
metric
SDK
is
not
stacked,
we
don't
really
be
forced
to
implement
it
to
connoisseurs
meet.
You
I
believe
it
should
be
fine
with
just
the
metric
API,
because
that's
all
it
expect
that's.
My
understanding
is.
D
If
I
mean
we
still
have
the
working
SDK,
we
could
be
helpful,
Mathias
exporter,
but
I
would
consider
the
most
like
really
alpha,
not
really
beta,
because
it's
it's
like
kind
of
imagination.
You
can
argue
that
it
is
conforming
to
the
work-in-progress
spec,
but
that
since
the
spec
itself
is
not
most
I
would
say
it's
it's
like
more
like
an
alpha,
not
a
beta.
D
We
won't
be
like
violating
or
we
won't
be
any
different
from
any
other
six
like
Python
or
Java
or
any
other
language,
because
everyone
shipped
metrics
and
call
themselves
beta
without
having
an
official,
SDK
spec
with
just
the
API
spec.
So
I
think
we
can
just
follow
what
everyone
else
did
and
yeah
I'll
create
allowed
update
this
with
like
actual
tasks,
I
am
trying
to
say,
I'll
get
like
one
more
person
from
Microsoft
to
work
Raj.
D
He
he
just
started
working
on
this,
so
it
will
be
two
of
us
from
Microsoft
in
the
next
two
weeks.
So
hopefully
we
should
be
able
to
move
and
do
the
release
of
first
beta.
So
just
to
like
continue
on
the
same
context,
length
we
decided
will
be
having
an
official
Spanish
shim
as
a
light
layer
on
top
of
the
activity.
So
the
beta
would
include
that.
D
But
this
is
in
the
plan
for
beta,
like
so,
users
will
have
the
flexibility
to
either
use
activity
directly
or
use
the
more
open,
telemetry
friendly
span
based
epa's,
and
just
to
confirm,
like
this,
like
the
sdk
like
the
open
telemetry
is
take
a
which
we
ship
will
operate
on
activity
and
there
will
be
requesting
a
converting
the
spans
from
the
wrapper
into
the
activity
directly.
So,
like
that's
one
x,
I
said
initially
respective
of
further
a
wrapper
is
used
or
not.
D
D
Calling
meter
factory
instead
of
provider,
but
it's
track
here
already
like
the
renaming,
but
that
aside,
we
should
be
good
to
call
ourselves
meter
for
matrix.
So
that
would
be
the
plan
for
next
two
weeks
not
expect
to
see
like
more
issues
like
this
being
opened
and
if
anyone
is
willing
to
contribute.
Please
like
comment
in
the
issue
or
if
you
have
the
permission
cell
for
saying
so
that
we
don't
compete
with
each
other
from
the
Sigma
time.
D
B
D
B
We
should
have
an
explicit
plan
for
how
other
instrumentation
related
to
activity.
That's
why
I'm
raising
it
stress
but
also
show
those
two
mutations
say
you
have
activity
API
and
the
the
libraries
have
chosen
should
instrument
themselves
using
that
API.
Everything
is
great.
So
from
an
SDK
scenario,
everything
is
great.
There
is
tremendous
data
is
collected
on
the
Shred,
not
a
disconnect
from
a
auto
limitation
scenario.
B
The
things
to
address
there
is
one
is
some
libraries
have
instrumented
themselves,
some
have
not,
and
we
should
have
very
general
recommendation
of
how
to
make
sure
that
they
play
well
with
each
other.
That
will
probably
be
very
easy.
The
the
the
tracer
will
just
inject
like
the
the
activity
based
additional.
B
Telemetry
tracking,
where
needs
to
be
that's
fine,
but
we
should
just
kind
of
think
for
a
second
that
it
will
converge
together.
I
think
it
will,
but
we
should
just
like
think
about
it.
The
second
thing
is
what,
if
the
Old
Testament
ation
offers
believe
that
the
information
emitted
by
the
library
should
be
modified
in
some
way,
maybe
enriched
or
changed
something
or
whatever.
So
the
scenario
should
be
like
again.
B
B
So
a
tracer
should
be
able
to
subscribe
to
all
these
activities,
and
there
is
the
nutrient
requires
registering
a
callback.
Is
this
what
we
also
expect
from
a
from
a
tracer
that
somehow
generates
some
code
that
adds
itself
to
the
list
is
like,
so
what
would
be?
What
would
be
a
recommend
pattern
for
a
in
process
agent
to
start
listening
to
all
the
activities
is
the
suffice
to
fire
diagnostic
source,
which
might
be
the
answer
or
might
not
is
the
smell?
Something
else.
B
B
They
don't
have
to
follow
it,
but
at
least
make
sure
that
the
API
that
we
design
now
and
ship
so
that
these
traces
will
happen
right.
So
this
wouldn't
hang
around.
It
should
be
like
natural.
It
should
be.
Yes,
we
have
thought
about
this
scenario
and
it's
you
know
we're
not
addressing
it
in
this
group
addressed
some
other
group,
but
we
are
designing,
for
this
should
be
supported.
D
Okay,
so
let
me
ask
you
want
to
say
like:
is
it
I'm
not
very
sure
how
the
in
process
agent
or
whatever
you
call
it
attached
to
an
already
running
process?
So
you
I
would
imagine
that
without
that,
attaching
thing
it.
It
has
to
do
like
pretty
much
what
I'm
sharing
on
the
screen
like
it
has
to
run
this
line
of
code
from
well.
B
That
that
a
tracer,
so
a
tracer
one
of
one
of
the
things
that
the
tracers
will
want
to
do
like
procedures
agents,
is
to
have
very
very
few
dependencies,
so
they
it
should
be
optional
for
them
to
have
this
decay
there
or
not,
I
think
in
any
any
good
tracer.
They
will
have
certain
booster
to
support
the
option
should
send
data
through
exporters,
but
they
should
be
able
to
work
without
taking
yeah.
D
E
Example
in
a
spiffing,
so
I
think
the
part
here
is:
we've
got
the
application,
so
a
application
that
may
not
be
instrumented
doesn't
want
to
depend
on
tracer,
but
the
agent
it's
going
to
monitor
it
externally
will
have
a
dependency
on
the
on
the
SDK
and
that's
how
it's
going
to
clap
telemetry
information
it
may
or
may
not.
It.
B
May
be
that
people
would
like
to
create
a
agent
that
attaches
to
the
to
the
process
and
does
you
know,
does
whatever
it
means,
like?
You
know,
modifies
bytecode
that
chooses
not
to
take
dependency
on
this
DK.
It
makes
sense
that
it
takes
dependencies
on
activities,
of
course,
because
it's
part
of
the
framework,
but
that
tivity
api
should
make
sense
for
such
tracer.
If
the
trace
it
shows
us
not
to
depend
on
the
SDK.
B
Otherwise
what
would
happen
and
in
reality,
when
we
may
or
may
not
wish
for
it,
but
in
reality,
what
will
happen
is
the
tracers
will
not
take
advantage
of
whatever
provision
we
have
here?
They
will.
They
will
hack
around
avoid
dependencies
on
the
SDK
and
make
make
the
dependencies
optional
and
do
something
else,
because
I'm.
D
B
B
I'm
starting
these
activities,
essentially,
the
system
knows,
if
nobody's
listening
to
that.
Tivities
activities
are
not
being
created
right,
so
the
API
didn't
need
to
invoke
should
make
sure
that
they
are
created
so
that
they
can
start
listening.
It's
this
activity,
like
doing
is
SDK
to
ensure
call
this
API.
No,
no.
D
B
Could
take
it
offline,
but
maybe
this
and
the
thing
is
I-
may
not
have
things
better
to
actually
say
how
it
would
look
like.
The
scenario
is
that
a
tracer
that
does
not
have
depends
on
diagnostic
source,
but
not
on
open
telemetry
once
sure
consume
activities.
So
what
does
the
tracer
need
to
do
to
start
getting
activities?
E
D
B
Say
say:
the
scenario
is
that
a
tracer
like
a
process
agent,
once
you
collect
data
both
from
instrumented
libraries
and
from
non-instrumental
by
adding
bytecode
instrumentation
and
open
telemetry,
API,
nougat
or
SDK
no
get
and
not
on
the
game.
So
the
tracer
doesn't
want
the
tracer
texted
dependency
on
the
activities,
but
not
on
those
limits.
D
See
I'm
still
having
some
questions,
but
let
me
see
like
please
see
if
my
understanding
is
correct.
So
if
you
have
an
agent
which
wants
to
listen
to
the
activities
emitted
by
the
libraries,
it
does
not
require
SDK.
There
is
a
built-in
mechanism
in
activities
to
do
the
listening.
However,
what
what
this
line
of
code
here
is
doing
is
rapping,
on
top
of
it
like,
when
I
say,
build
an
activity
source
if
it's
internally
calling
the
activity
or
diagnostic
source
mechanism
to
subscribe
to
it.
D
So
if
you
are
not
using
this,
then
you're
not
using
the
open,
telemetry
staccato-like
use
like
either
building
a
parallel
thing
or
using
some
other
mechanism
and
I.
Don't
really
have
any
say
in
that
part
because
I
don't
really
know
how
the
agent
is
supposed
to
work.
But
if
the
question
to
your
answer,
is
you
don't
really
need
the
SDK
reference
to
instrument?
D
B
I
think
we
should
just
sort
of
I
think
it
may
be
that
everything
is
already
possible
and
defined.
You
think
I
just
wanted
to
have
like
that.
We
have
an
exploit
I
just
wanted
to
have,
as
a
group
of
somebody
that
I
can
participate,
participate
just
that
we
have
gone
through
a
mental
exercise
that
we
have
explicitly
like
understood.
B
A
couple
of
scenarios
that
are
important
for
this
tracing
codelist
attached,
wrote
and
that
we
have
kind
of
carefully
sat
down
and
went
on
our
heads
and
understood
that
everything
is
there
and
that
nothing
is
missing
for
those
scenarios.
It
can
very
well
be
that
the
answer
will
be
yes,
everything
is
there
and
we
don't
need
to
make
any
tweaks.
I
just
would
think
I
think
it
would
be
great
if
it
wasn't
explicit
understanding
from
like
a
little
investigation,
rather
than
just
a
coincidental
outcome.
B
D
B
D
From
this
report
like
we
intend
to
provide
like
iodine
on
how
to
instrument
your
libraries
and
how
do
consume
open
telemetry
in
an
application,
but
anything
top
of
that
should
be
built
on
like
what
these
two
basic
documents
like
how
the
instrument,
a
library
and
how
to
use
open,
telemetry
like
it.
This
is
like
actual
code
example
with
how
how
do
you
instrument
a
library-
and
this
shows,
how
do
you
enable
open
telemetry
for
a
particular
exporter?
A
D
D
Think
your
point
is
like
this
should
be
like
more
like
a
document
or
like
because
the
the
point
which
I
don't
get
is
if
we
document
how
to
instrument
libraries
and
if
we
document
how
to
consume
open
elementary
data
to
a
particular
exporter
and
the
job
of
this,
like
repositories
over
like
anything.
On
top
of
that.
B
The
the
like
we
come
together
as
a
group
because
we
want
you
have
pace,
electro
and
basic
technology
for
collecting
this
data,
and
so
that's
that
the
outcome,
and
if
in
some
education
areas
that
means
that
the
actual
SDK
is
not
part
of
the
process.
As
in
like
process
on
the
machine
on
the
CPU,
then
then
that's
fine,
because
I
just
have
a
feeling
from
my
understanding
of
that,
you
want
to
have
additional
dependencies
being
always
optional,
and
maybe
this
this
article
belongs
not.
B
A
Just
just
to
be
sure,
Greg
because
I
think
the
these
are
a
kind
of
definition
of
tracer
or
open
telemetry
and
actually
active
source
maps
very
very
closely.
True
the
tracer,
that
is
what
obey
telemetry
calls
the
API.
That
is
the
only
thing
that
libraries
should
be
taking
dependents
and
as
CEO
mention
the
is
this
lines
that
are
selected
here.
A
They
are
rappers
should
deal
with
activity,
but
if
you
don't
want
to
take
the
dependence
from
the
open,
telemetry
SDK,
you
are
going
to
be
able
to
do
that
just
using
achieved
source
that
maps
to
the
tracer
concept
in
open
telemetry,
the
activity
itself
that
maps
to
spam
so
I
understand
your
concern
and
we
need
to
kind
of
really
go
through
these
steps.
But
the
intention
is
really
dead.
It's
possible
I
didn't
see.
A
I,
don't
recall
it's
possible
that
there
is
some
issue
that
we
perhaps
missing
here
there,
but
that
should
be
pretty
much
the
case.
You
know
that
we
should
be
able
really
to
if
I
library,
for
instance,
wants
to
just
use
activity
source
in
the
application
also
just
wants
to
use
activity
source
that
should
be
possible.
A
You
know
that
there
you
should
be
no
case
that
death
is
blocking
and
if
that
application
swaps
I'm
there
I'm
talking
about
application
swaps,
the
day
I
use
of
set
of
consuming
the
day
activity
generated
via
achieved
source
via
the
open
telemetry
has
decayed.
It
should
be
totally
indifferent
for
for
their
libraries.
In
that
case,.
B
B
We
just
there's
at
least
see
participants
here
who
are
all
in
clinic
and
we
would
like,
if
I
think
it
would
benefit
on
a
technical
level
all
at
least
this
view
of
us.
If
we
come
up
with
and
publish
it
here,
but
maybe
it's
just
a
mental
exercise
that
we
just
publish
it
within
a
github
issue
and
close
it
later.
If
we
have
a
code
sample
of
a
coat
that
is
sensibly
being
part
of
a
tracer,
so
that
you
know
it's
not
a
did
encompasses
with
the
application.
B
B
Because
then
the
reason
why
it's
particularly
interesting
for
the
Tracy
universe
is
that
Tracy
creates
versioning
challenges
that
are
not
part
of
a
normal
application,
because
application
just
builds
it
and
ships
it
and
that's
it.
A
Tracy
needs
to
deal
with
issues
like
a
running
application
has
a
bunch
of
dependencies,
and
we
need
to
come
with
another
bunch
of
dependencies.
It
all
somehow
be
reconciled
and
made
to
work,
and
there
is
decoupling.
It
makes
things
a
lot
easier.
So,
from
the
perspective
of
tracing
architecture,
that
may
be
a
very
good
conversation
to
understand.
D
B
D
Stds
already
separated,
that
is
pure
man
didn't
open,
elementary
or
APA,
which
is
required
for
instrumentation
and
in
the
mechanism
to
export
it
large,
like
two
independent,
consulting
a
PA
and
SDK.
So
this
concepts
are
already
like
separate
in
open
elementary
itself
and
we
follow
the
same
pattern
like
there
is
no
I
mean
technically.
You
can
have
a
library
instrument
which
uses
an
exporter
from
this
repo,
but
you
are
not
forced
to
use
an
SDK
which
we
ship
from
here.
D
B
B
B
D
A
D
A
We
we
had
a
meeting
discussing
how
to
better
collaborate
and
kind
of
how
can
align
the
expectations
from
from
people
working
on
how
to
store
meditation
and
I
I.
Think
Greg,
a
read
brought
some
of
the
questions
here,
but
just
to
be
very
quick,
because
I
think
our
meeting
is
kind
of
taking
a
bit
of
time
today,
but
just
to
a
quick
update.
The
idea
is
that
the
outage
termination,
the
goal,
is
to
start
from
the
data
dog,
but
we
want
to
ensure
that
is.
A
collaboration
is
not
just
a
fork.
A
You
know
we
want
to
kind
of
keep
data
dog,
involve
it
and,
of
course,
data
dog
wants
to
see
what
is
contributing
to
to
this
out
of
cementation.
So
for
that
to
be
successful,
we
have
to
have
kind
of
independence,
m
pluggable,
just
to
use
a
term
that
perhaps
is
easy
to
understand,
independent
and
pluggable
about.
A
Let's
say
what
open
telemetry
defines:
has
the
SDK,
the
one
that
is
actually
exporting
the
data
and
generating
the
data
and
also
internally,
the
goal
is
to
use
activity
source,
so
it
doesn't
have
any
other
dependence,
besides
their
runtime
and
a
NuGet
package,
that's
provided
for
by
Microsoft
for
any
application
library.
In
this
case,
and
in
general,
we
don't
want
to
make
these
things
pluggable
for
context
propagation
true.
So
besides
reporting,
WC
and
d3,
we
we
want
to
be
able
to
allow
others
to
plug
different
forms
of
context,
propagation
and
I.
A
Think
I
think
this
captures
kind
of
the
summer
of
the
meeting.
We
often
talk
a
long
time
discussing
these
things,
but
the
idea
is
to
capture
these
in
a
in
a
short
document
and
kind
of
do
the
follow
ups
from
there.
You
know
so
play
an
experiment
with
making
the
the
sdk
trace
a
pluggable
and
kind
of
perhaps
do
a
small
pick
up,
one
of
the
instrumentation
that
already
exists
on
data
dog
and
transform
that
to
the
new
achieved
source,
to
show
the
viability
and
how
that
would
work.
A
You
know
that
that's
something
that
crosses
my
mind
and
III
will
be
on
vacation
until
next
week,
so
I'm
not
going
to
be
doing
that.
I'm
gonna
try
to
put
the
document
up
to
date,
so
people
can
keep
having
conversation
on
that
part
Kay
and
anyone
that
wants
to
join
can
participate
and
contribute
there.
So.
D
I
have
a
basic
question
on
the
auto
instrumentation
order,
because
my
understanding
of
water
instrumentation
is
auto
instrumentation
from
they
put
Microsoft
dis
with
I.
Sure,
like
you,
have
an
application
running
in
a
sure.
You
did
not
explicitly
choose
to
do
any
monitoring,
but
like
we
do
a
cordless
attach
where
we
inject
something
into
your
running
application
and
starts
emitting
telemetry.
But
this
is
somewhat
different
from
using
pi
code
instrumentation
to
like,
because
the
way
microsoft's
photo
cordless
a
test
works,
is
it
won't
collect
telemetry
from,
for
example,
MongoDB.
D
Fair
application
is
using
MongoDB
and
there
is
the
MongoDB
library
itself
is
not
instrumented
with
diagnostic
source,
so
we
won't
collect
it.
But
if
your
application
is
using
HTTP,
client
or
secret
client,
then
we
will
collect
telemetry
from
it.
So
my
question
is:
is
the
goal
of
photo
instrumentation
just
to
like
attach
to
those
applications
which
uses
libraries
which
are
instrumented
with
activity
or
open
elementary
APA,
or
is
a
goal
to
get
instrumentation
to
those
libraries
which
do
not
have
any
sort
of
instrumentation
like
MongoDB
or
something?
Oh,
okay,.
B
Engine
of
Microsoft
is
a
low-level
technology
that
makes
it
that
makes
it
easier
to
write.
Profilers
profiler
is
irrelevant
because
profiler
is
the
dotnet
API
that
allows
you
to
actually
attach
to
the
process
and
do
these
things
that
are
required
and
the
data
doc
technology
is
more
high-level.
It's
actually
a
very
special
type
of
profiler
that,
rather
than
collecting
traditional
profiling
data,
modifies
an
existing
application
to
make
it
emit
telemetry
and
talking.
At
the
same
time,
it
also
collects
limit
reso.
B
The
there
is
baked
in
knowledge,
like
called
in
Cole
about
libraries
and
their
versions,
and
the
the
tracer
explicitly
knows
whether
a
given
library
and
its
given
version
is
already
instrumented
and
if
it,
if
it
is,
and
it
simply
listens
to
the
activities
that
are
emitted
by
it.
Okay,
that's
not
as
tremendous,
then
it
on
the
fly.
A
running
application
depending
on
scenarios,
have
made
a
query
and
application
restart,
but
more
it
modifies
the
application
to
actually
start
a
meeting.
D
B
Each
library
has,
of
course
many
API
is
because
you
know
if
you
say
there
might
be
10
public
API
is
that
you
could
potentially
invoke
and
all
that
you
write
or
read
some
processes
for
us
when
we
say
one
integration
means
take
MongoDB
and
it
has
like
and
sumit
all
the
10
API
that
are
affected.
Okay,.
D
D
A
E
B
B
A
B
A
B
I
think
in
the
future,
hopefully
and
I,
think
in
terms
of
future
that
transition
just
questions
the
future
integrations
for
both
of
us
driven
by
customers.
So
when
one,
when
customers
ask
for
an
additional
integration,
then
we
build
it
and
right
now
we
are
balancing
internally
new
integrations
with
improvements
to
the
actual
agent
to
make
it
faster.
D
So
let
me
ask
you
one
more
thing
like
if
someone
asks
this
question
to
general
community
like
under
support,
for
instance
MongoDB.
The
answer
would
be
one
if
MongoDB
like
be
able
to
monitor
it
by
just
using
the
open.
Telemetry
is
decayed
number
two.
If
MongoDB
does
not
choose
to
instrument
with
diagnostic
force,
then,
if
the
data
dog
is
obvious
is
like
is
capable
of
instrumenting
MongoDB,
then
you
would
still
be
able
to
monitor
the
particular
library.
Assuming
you
are
okay
to
use
the
auto
instrumentation
agent
/
director
is
that
the
right
treatment.
D
We've
been
getting
this
question
like
in
several
channels.
Like
do
you
plan
to
support
this
librarian
I've
been
always
giving
the
answer
that
if
they
choose
to
instrument
with
Diagnostics
or
then
yes,
but
other
than
write,
this,
like
this
repo,
officially
be
only
how
instrumentation
for
like
four
of
them,
like
ICD,
big
line,
secret
line
and
reduce
reduce,
is
using
some
special
APA.
But
the
answer
is
like
more
to
that.
If,
with
order
instrumentation,
we
should
be
able
to
support.
B
Also
your
question:
if,
if
the
library
is
not
yet
supported
and
the
the
preferred
way
to
start
supporting
interest,
of
course,
to
add
water
to
rotations
to
the
library,
but
there
is
a
possible
scenario
where
somebody
some
third
party
is
very
interested
in
having
support
for
that
library.
But
it's
not
the
actual
person
who
controls
the
library
and
the
library,
for
whichever
reason
is
not
ready
to
add
that
support.
Then
the
third
party
can
simply
authors
at
such
an
integration
and
contributed
to
the
tracer
rather
than
choo-choo
the
library
directly
and
then
assistance.
A
Okay,
just
one
and
one
thing
that
make
verify
if
I'm
more
residual
is
that
the
auto
store
meditation
for
data
dog.
You
can
target
by
version,
so
you
instrument
and
MongoDB,
but
you
say
the
range,
your
version
of
MongoDB.
That
is
you
outer
instrument.
You
know,
and
let's
say
from
version
of
X:
they
start
to
themselves.
A
E
C
B
I
would
appreciate
if
we
can
wait
a
month
starting
a
new
meeting.
If
you
guys
are
okay
with
this,
we
already
have
a
channel
so
yeah.
The
reason
is
like
I'm
driving
this,
but,
as
you
all
know,
I
am
new
to
this
effort
and
I
am
asking
stupid
questions
about
open,
telemetry,
I,
don't
yet
know
the
agents
right
now.
It's
just
like.
B
C
B
E
B
So
we
would
like
to
do
this,
not
for
the
sake
of
doing
it,
but
because
we
have
found
a
partner
with
whom
we
have
a
mutually
beneficial
roadmap.
So,
for
example,
anybody
here
says:
okay,
we
would
like
to
make
contributions,
should
I
show
this
code
base
and
they
todo
closely
this
contribution
that
says
yes,
we
would
benefit
from
these
contributions,
not
just
benefiting
you.
B
So
there
is
potential,
it
would
be
a
case
I,
don't
believe
it
is,
but
it
could
be
a
case
where
people
say
we
would
like
to
add
certain
capability
to
the
agent
and
I'm
sorry
to
the
trace.
I
am
interesting
and
they
to
talk,
looks
at
the
test.
That's
fine,
but
it's
zero
benefit
to
us.
Please
feel
free
to
do
it,
but
you
can
work
it
and
do
it.
It
has
absolutely
here
been
to
us.
B
A
Integrations
we
we
have,
we
don't
have
right
now
any
anything
red
to
contribute.
We
have
some
customer
asking
for
specific
libraries,
but
we
didn't
move
to
that
stage
yet
because,
ideally,
we
would
like
to
invest
on
something
that
can
be
shared
and,
for
you
know,
we
have
more
focusing
on
the
infrastructure
of
hope
and
telemetry.
So
far,
so.
B
B
That
is
an
example
of
something
that
benefits
everybody,
including
ourselves,
and
then
we
would
say:
okay,
let's,
let's
move,
if
the
contribution
is,
we
have
entered
the
data
exported
to
the
Splunk
backend
then,
for
example,
its
contribution
that
is
really
Same,
Same
Same
for
Microsoft,
for
example,
I
know
that
it's
rotation
and
engine
is
very
important
to
Microsoft
to
us.
We
would
do
the
verb,
but
not
soon,
because
I
think
it
benefits
we're
happy
to
have
this
as
a
basis
for
the
tracer,
but
we
have.
We
are
more
important
things
that
are
high
yeah.
B
Make
some
particular
performance
improvement.
There
was
some
particular
stability
improvement
that
it's
not
about
the
engine.
It's
about
some
I
know.
Microsoft
had
some
some
concerns
about
the
sort
of
native
part
of
the
code,
and
some
of
it
is
essentially
at
the
very
top
of
our.
We
will
do
it
very
soon
anyway,
so,
but
some
of
it
we
might
not,
and
if
Microsoft
says
for
us
it's
important,
we
will
actually
contributing
or
microsoft,
says
the
stuff
that
is
already
on
the
top
of
your
roadmap.
It's
also
super
important
to
us.
B
Let
us
take
as
a
subset
of
it
and
we
will
contribute
and
you
can
contribute
to
analysis
or
something
like
that
in
any
kind
of
video.
That
is
actually
very
charts
that
also,
then
we
say:
ok,
we
move
it.
You
see
what
I
mean
if
Microsoft
essentially
hangs
out
on
ways
until
we
move
it
at
only
once
we
move
it,
then
when
they
are
ready
to
deliver
them,
I
would
suggest:
let's
not
play
politics
like
we're
a
small
company
compared
to
Microsoft.
Well,
we
want
to
do
it.
C
B
B
C
C
So
what
from
Microsoft
perspective,
I'm
Finn
in
is
should
I
also
should
be
also
like
fork
data
to
look
at
this
and
try
to
use
it,
and
only
then
they'll
start
thinking
about
a
pyramid,
or
you
actually
good
to
go
with,
like
shared
ops
team
between
that
a
dog
and
its
plan.
Taking
this
apps
team
up
same
repo
and
from
V
sub
same
leave
of
Microsoft
will
be
able
to
create
a
Safran
and
start
laying
the
list
and
getting
more
and
more
interest
and
eventually
contributing
back
right.
C
B
B
Just
for
the
to
understand,
you
imagine
that's
plank
for
whatever
reason
and
then
one,
but
just
understand
your
your
position.
It
says:
oh,
no,
we
would
we
don't
want
to
participate.
We
want
to
participate
only
three
months.
Then,
why
like
Microsoft
and
beta
dough,
can
still
create
the
open,
telemetry,
a
branch
and
sorry?
C
B
C
C
B
We
can
talk
again,
that's
what
I'm
saying
if
Microsoft
were
to
contribute,
but
you're
not
saying
was
Ward
like
if
you
want,
if
you
want
to
contribute
with
stuff
that
as
Microsoft
specific
that's
what
I'm
trying
to
understand
that
is
not
of
value
to
anybody
else.
That's
also
totally
okay,
and
we
can
also
discuss
this
and
we
can
create
a
just
essentially
pool
without
adding
an
optical
imaging
process
just
for
easy
merging.
This
can
be
any
then
it's
like
no
process,
no
official,
open,
telemetry,
just
engineering
level
is
emerging
that
we
can.
B
A
Think
I
can
give
a
good
example.
I
think
it's
good
for
our
companies
to
try
to
really
be
together.
Besides
sharing
the
resources,
it's
things
like
first,
a
lot
of
customers.
They
don't
want
to
kind
of
be
tired,
the
true
anything.
Even
if
it's
open
source,
they
don't
want
anything
that
specifically
to
any
vendor.
B
Alright,
let's
talk
next
week
and
Michael
justice
is
a
summary
for
you,
like.
Let's,
let's
talk
anytime,
that
works
for
you,
maybe
about
what
are
the
specific
changes
you
want
to
start
with,
because
there
is
lots
of
benefits
to
have
a
sure
thing,
but
I
understand
that
no
means
reviewing
more
it
support
if
you
more
pull
requests
blah
blah
blah
discussing
touching
code
code
that
affects
everybody,
blah
blah
blah
its
work
and
even
a
common
area
upstream
Michael.
B
It's
also
work
it
because
now
we
are
talking
about
discussing
code
reviews
and
what
not
everybody
wants
it
but
like
we,
we
about
going
to
propose
work
to
deaf
directors
and
they
will
ask
how
this
work
is
justified
and
a
potential
rosy
world
where
everybody
collaborates
is
not
going
to
do
it
not
for
small
company.
It
will
be
like
these
are
the
features
that
our
partners
they.
B
Of
course
they
are
not
signing
any
contracts
but
they're,
very
serious,
and
they
have
verbally
committed
on,
like
some
kind
of
level
that
they're
going
to
contribute
this
thing
within
this
time
frame,
and
this
is
why
we
are
scheduling
additional
work
and
that's
what
was
going
to
fly
right
and
that's
what
I'm
trying
to
establish
like
what?
What
are
these
things
that
everybody
contributes,
so
that
we
can
justify
the
overhead
of
like
additional
code
reviews,
additional
synchronization,
okay,
I.
C
A
B
B
C
C
B
Exactly
the
sort
of
thing
that
I'm
really
that's,
why
I'm
keeping
you
that's?
Why
I
talk
to
you
guys
like?
Let's
have
a
listen?
Let's,
let's
have
a
list,
because
that's
exactly
what
I
want
to
do
now.
I
think
we
kind
of
landed
on
this
and
lost
our
conversation.
You
said
that
you're
still
like
considering
other
options,
looks
like
maybe
your
maybe
you
haven't
decided
if
you're
trending
traverse
they
like
working
together
on
this.
That's
great,
then
definitely
this
roadmap.
That's
exactly
what!