►
From YouTube: 2022-08-24 meeting
Description
OpenTelemetry PHP SIG Meeting
B
B
B
So
recent
recent
pull
requests
are
a
point
of
mind
that
I
think
hasn't
been
looked
at
by
anybody
yet,
but
this
is
to
just
do
a
kit
split
on
some
of
their
contrib
stuff
so
that
it
can
be
independently
independently,
published
and
installed
by
a
packageist,
I
think,
with
with
no
other
maintainers
around.
That's
not
going
to
get
done
for
at
least
a
whole
week,
or
so
we've
got
refactors
and
context
amber.
A
B
B
A
B
B
It's
just
that
this
guy
just
disappeared.
I
think
we'll
just
let
that
one
hang
around
for
a
bit.
V3
propagator,
I
think,
was
pretty
close.
I
remember
reviewing
that
the
last
couple
of
days-
yeah.
Yes,
you
are
anything
you
wanted
to
speak
to
about
this
cool
request,
which
is
now
well,
it's
approved
by
me.
D
Sure
so
earlier
I
had
created
a
pr
for
v3
multi
header
propagator,
and
this
pr
was
yeah
supposed
to
be
for
implementation
of
b3
single
header
and
we
went
through
the
specs.
There
were
some
changes.
Initially
we
implemented
in
the
api,
but
as
the
two
bias
pointed
that
open
telemetry,
specs
suggest
v3
and
zipkin,
I
guess
both
propagator
should
be
implemented
as
an
extension.
D
B
B
B
B
B
B
B
Yeah
yeah,
I'm
not
sure
how
interestingly
this
will
be.
This
could
be
related
to
that
other
issue,
with
a
better
deprecation
notice,
still
trying
to
chase
up
some
details
or
a
test
case
on
there.
But
it's
probably
time
that
we
did
update
to
the
latest
predator
things,
which
I
think
is
pretty
easy
to
do.
I've
done
it
before
like
creating
a
shooter
to
do
that,
but
there
might
be.
I
can
see
a
lot
of
work
with
it
as
well.
B
A
D
D
B
B
D
A
B
B
B
B
Sdk
button
it'd
be
documentation
for
this,
wouldn't
it
and
I
I
to
be
honest-
I've
never
even
really
spent
a
lot
of
time
looking
at
this,
so
I
don't
know
how
many
of
these
things
have
been
ticked
off,
but
but
I'm
pretty
sure
that
that's
that's
what
it's
talking
about.
It's
not
talking
about
the
documentation
in
our
main
repo.
B
C
So
so
my
name
is
jeremy
gulevsky,
it's
probably
heart
name
for
you.
It's
primek
is
my
first
name
polish
name,
so
I'm
participating
in
this
in
this
meeting
for
the
first
time
and
I'm
representing
sumo
logic
company
and
we
are
interested
in
in
doing
this
php
automatic
instrumentation
and
we
would
like
to
also
participate
in
the
in
this
in
this
work.
C
C
So
maybe
everything
is
clear
now,
but
what
we
have
here
is
a
table
with
a
few
few
columns
that
contains
a
few
technologies,
and
currently
we
have
only
automatic
instrumentation
for
java
and
net,
and
also
something
that
is
that
was
added
recently
is
go.
So
this
is
working
progress
and
one
of
the
approach
for
go
instrumentation.
C
I
am
out
of
one
of
the
approach,
so
I
can
later
tell
you
more
about
it,
but
I
one
of
the
things
that
you
were
discussing
in
this
meeting
was
the
language
that
you
would
like
to
use
so
php
or
c
or
c,
plus
plus
right.
So
I
I
think
that
using
php
is
important
from
manual
instrumentation
perspective,
each
language
or
technology
that
we
have
here
in
this
table
for
manual
instrumentation
it
it's.
It
used
the
same
language.
So
for
java
we
are
using
java
for
net.
C
C
Okay,
so
and
for
automatic
instrumentation,
the
situation
is
a
little
bit
different,
so
for
for
java
there
is
still
java
that
is
used,
and
we
are
so
so
what
java
agent
is
doing?
Is
it
changes
or
re-writes
bytecode
at
runtime,
and
in
fact
the
same
is
is
done
for
for
net,
but
for
dotnet
we
have
to
use
c
and
or
c,
plus,
plus
and
profilerapi.
That
dotnet
gives
gives
us
so
and
also
it
worked
in
the
same
way,
so
the
instrumentation
is
done
at
runtime
and
for
go.
C
One
of
them
is
doing
also
some
changes
at
runtime
and
it
uses
c
and
c
plus
plus,
and
it
requires
support
from
operating
system
in
this
case,
because
it
uses
a
mechanism
called
ebbf
and
for
now
this
mechanism
works
only
on
at
linux.
So
probably
it
will
be
available
available
soon
on
windows
as
well,
but
I
don't
know
about
any
other
platforms
and
the
other
approach
that
I
am
out
of
for
go.
Is
I'm
calling
it's
automatic
manu
manual
automatic
instrumentation?
B
C
C
So
so
this
is
a
very
simple
example
of
of
go
application.
It's
calculates
fibonacci
number,
so
we
have
a
main
function
and
in
in
other
file
we
have
an
implementation
of
of
two
functions.
In
fact,
there
is
fibonacci,
helper
and
fibonacci,
and
this
approach
works
in
in
the
way
that
we
have
to
tell
where
will
be
our
starting
point
of
instrumentation,
so
we
we
have
to
add
one
additional
call
into
the
our
source
course.
C
So
so
one
one
more
thing
this
this
will
in
start
instrumentation
in
mind
main
function,
and
then
it
will
search
for
all
the
function,
definitions
that
we
have
in
the
project
and
also
all
functions
calls
from
this
place.
So
here
we
have
a
feed,
bonucci
helper
and
from
fibonacci
helper
it
calls
fibonacci.
So
it
should
instrument
all
this
all
these
functions
that
are
visible
right
here.
C
And
as
a
as
a
last
parameter,
I'm
telling
that
it
should
take
all
the
packages
that
are
in
this
application.
So
that's
the
dots
that
you
see
here
and
now
currently
it
it.
It
is
doing
a
few
passes
to
to
first
find
this
root
call
root.
Call
that
contains
the
function
that
I
showed
you
this
out
entry
point
and
then
it
searches
for
all
function,
definitions
and
all
functions
calls
and,
as
you
can
see,
that's
how
a
result
looks
like.
C
So
if
in
each
in
each
function
we
create
span
and
so
on.
So
that's
that
tool
works
on
the
source
code
at
compile
time,
and
it
has
some
advantages
over
this
binary
instrumentation
because
it
doesn't
require
support
from
operating
system,
it's
portable
and
also
it.
It
doesn't
require
some
elevated
privileges,
for
instance,
so
yeah.
Basically,
that's
what
I
wanted
to
say
and
for
php.
In
my
opinion,
we
should
for
automatic
instrumentation.
C
B
C
B
C
B
If
you
haven't
already,
is
to
jump
onto
our
slack
channel,
which
is
where
all
the
sort
of
planning
takes
place
where
we
talk
to
each
other
outside
of
meetings
yeah,
maybe
just
yeah
jump
on
there
and
just
even
just
start
the
conversation
with
me
and
then
we'll
try
and
hook
in
okay,
yeah
we'll
get
a
group
going
because,
as
I
said
there,
there
are
a
few
a
few
different
ideas
floating
around
people
are
thinking
about
the
same
problem
that
you're
thinking
about.
So
let's
try
and
get
some
get.
B
Yeah
I
did,
I
did
raise
that,
and
it's
particularly
pertinent
right
now,
because
our
other
maintainers
can't
all
yes
yep
sure
talk
to
me
on
on
slack
and
we'll
see
what
needs
to
happen
there.
It
might
actually
need
some
guidance
from
from
bob
when
he
comes
back
from
holiday
anyway,
because
I'm
not
sure
how
to
how
to
make
that
happen,
but
appreciate
your
interest
and
let's
see
what
we
can
do.
A
Also
great
one
thing
that
I
would
like
to
suggest
is
this
auto
instrumentation
thing:
since
many
people
are
interested,
we
can
streamline
the
efforts
somehow
so
that
this
discussion
can
go
at
a
little
faster
based
on,
because
I
think
many
are
maybe
interested
in
the
getting
to
the
solution
so
yeah.
Somehow,
if
these
things
can
be
collaborated
more
and
so
we
can
go
to
the
results
faster.