►
From YouTube: 2022-06-01 meeting
Description
OpenTelemetry PHP SIG Meeting
A
B
D
C
B
E
Yes,
so
used
to
things
being
in
a
certain
place
or
or
even
my,
my
mom
is
still
living
in
in
our
old
place,
where
I
grew
up,
but
she
put
in
like
a
new
kitchen.
You
know
stuff
like
this,
and
everything
is
somewhere
else
now
and
I
I'm
so
used
to
the
you
know,
stuff
being
in
a
certain
place
from
my
childhood
and
it's
kind
of
crazy.
E
E
Wow,
but
did
you
figure
figure
out
if
those
are
poplars.
B
D
E
Yeah
they
can
get
really
tall
and
they
have
really
stiff
leaves
so
they're
kind
of
kind
of
a
butterfly
killer
because
they
can
cut
the
wings
yeah.
That's
the
downside
of
them,
but
they're
often
planted
where
you
know,
people
just
fast
growing
trees
and
they
can
break
easily
because
they
are.
The
wood
is
very,
it's
not
very
hard.
D
C
C
We
can
safely
say
you're
intelligent.
Well,
I
can't
be.
I
have
the
ability
to
be
all
right
well
without
further
ado,
and
let's
talk
about
trees,
more
talk
about
open,
telemetry
php
we
can
get
going
today.
So
we
have
a
couple
open
pull
requests
we
have
kashan
has
a
b3
propagator
for
b3
single
header.
I
think
who's.
C
C
E
Yeah
he
asked
me
some
questions
about
this,
how
to
deal
with
certain
things,
because
the
eb3
protocol
has
some
special
cases
where
there's
only
where
there's
a
deferred
sampling
decision
and
things
like
that,
and
so
basically
the
the
receiving
side
would
decide
whether
to
sample
or
not,
and
so,
and
he
asked
me
what,
when
we
actually
would
send
these
headers
and
for
these
special
cases,
I
just
told
him.
E
We
would
just
need
to
take
care
of
incoming
headers
like
that,
and
we
don't
send
them
because
they
are
really
special
cases
and
for
the
deferring
people
should
just
make
sure
that
their
assemblers
can
handle
this.
So
because
yeah
I
mean
you
should
have
control
well,
usually
over
the
whole
tracing
chain.
So
it's.
E
C
The
other
two
are
poor
requests
from
uk.
She
was
talking
about
updating
the
framework
integration
docs.
I
know
we
all
have
negative
feelings
about
those
documentation
that
documentation.
I
mentioned
that
she
should
rename
them
to
getting
started
rather
than
integration
guide,
to
be
a
little
bit
more
clear
and
succinct
about
what
they
actually
are
brett.
Thank
you
for
your
help.
With
that
the
w3c
test
service,
it
seems
like
she's,
circling
the
drain
on
finishing
that
one.
I
think
she
never
gave
her
some
good
insight.
She
pretty
much
spoon
fed
the
answer
to
her.
C
Thank
you,
you
have
pull
request
open.
Is
there?
Are
you
blocked
on
anything
there
or
are
you
still
working
on
it.
G
Yeah,
I'm
I'm
just
working
on
it,
but
I
thought
in
last
segmenting
we
discussed
that
timo
is
going
to
suggest
some
ideas
regarding
the
you
know:
moving
the
retry
from
the
exporter,
so
I'm
just
waiting.
I
mean
if
you
can
suggest
some
more
ideas
and
currently
I'm
looking
into
the
fiber
concept
so
till
I'm
new
to
the
php.
So
I'm
learning
it.
But
apart
from.
A
G
C
B
Okay,
I
mean
I
had
I
had
a
suggestion,
but
it
was
just
an
idea
which
was
in
the
comments
which
was
on
failure.
Instead
of
sleeping
and
retrying
straight
away
and
therefore
holding
up
sort
of
application
execution,
then
maybe
we
keep
a
like
internal
storage
of
you
know
things
to
retry,
something
like
a
batch
exporter
does.
B
Does
anyone
just
have
any
feelings
about
that?
I
just
I
know
that
when
I
saw
a
sleep
statement
in
in
an
exporter
it
it
concerned
me
yeah
the
performance
of
the
application
being
instrumented.
E
Yeah,
I
think
we
cannot
have
blocking
anywhere
and
it
depends
a
little
bit
of
course,
which
one
time
we
are
or
the
code
is
running
on
so
yeah.
If
it's
some
event
loop
runtime,
then
it's
way
over,
I
mean
you
wouldn't
need
sleep.
You
shouldn't
sleep
there
anyway,
but
there
are
different
ways
to
defer
things
than
an
apache
or
caddy
or
nginx
server,
or
something
like
that
yeah.
This
is
also
a
little
bit
problematic,
like
I
think,
added
a
long
text
of
war
on
about
the
problems
of
retries
and
single
threats.
E
So
yeah,
I
don't
know
and
that's
the
reason
why
I
said
it's
problematic,
because
there's
actually
no
good
way
to
handle
retries
and
singer
execution
right
other
than
sleeping
and
that's
yeah.
That's
blocking,
and
it's
deferring
the
the
well
finishing
the
response.
B
I
mean
if,
if
retries
were
only
a
function
of
say
the
batch
processor,
then
we
could
maybe
use
you
know
the
existing
infrastructure
of
the
batch
processor
to
you
know,
add
an
extra
field
or
you
know,
I
suppose
not
has
been
said,
because
it
would
not
be
there.
But
you
know
number
of
times
left
to
retry
before
I
give
up
on
this
or
something
and
then
as
part
of
the
regular
cycle
of
batch
processing,
just
iterate
over
everything
and
try
them
again.
E
I'm
not
quite
sure,
but
the
grpc
can
is
this
actually
synchronous
or
asynchronous,
or
can
they
be
asynchronous
the
grp
sequels,
because
that
would
if
they
are
only
if
they
would
be
asynchronous,
they
wouldn't
there's
a
better.
There
would
probably
a
way
to
better
handle
that
than
if
it's
just
blocking,
and
we
have
to
wait
on
a
response.
E
I
don't
I'm
not
sure
if
they
call
the
grpc
calls
if
they're
synchronous,
so
if
they
are
blocking
or
if
they're
asynchronous.
So
if
you
could
do
work
with
kind
of
a
promise
or
something
like
that,
so
while
the
responses
had
some
other
sense,
some
other
posts
could
be
executed.
G
In
grpc,
basically
they
are
a
seeker
on
us,
but
there
is
a
weight
call.
If
we,
if
we
use
that
call,
then
we
make
it.
You
know
we
wait
for
the
response,
so
yeah.
G
A
E
Yeah,
but
so
anyway,
the
most
important
thing
in
this
part
is
that
it's
configurable
and
there
are
no
basically
no
retries
plenty
fault
because
and
then,
if
one
people
or
users
really
want
to
have
a
re
retry
on
that
part
and
can
configure
it.
But
they
should
be
aware
that
maybe
the
process
is
longer
running
and
that
it
can
be
a
problem.
So
if
the
max,
if
it
exceeds
mexican
time,
then
we
will
create
a
fatal
error.
B
Might
be
worth
documenting
as
well
that
the
open
telemetry
collector
is
really
the
ideal
solution
here
and
then
you,
you
know
you
locate
it
next
to
your
next
to
your
code
as
a
sidecar
or
something
and
it
itself
acts
as
a
as
a
buffer
to
you
know,
hold
on
to
and
then
eventually
send
and
retry
traces.
C
E
And
for
the
other
stuff
I
mean
that's
not
part
of
the
specification,
I
think,
but
once
I
get
to
reworking
all
the
exporter
stuff,
then
we
will
use
asynchronous
chords
for
the
http
ones
as
well,
and
then
it
would
be
easier
to
implement
something
I
mean
for
jager
or
whatever
we
have
there.
So
that's
actually
on
my
list.
C
All
right,
I
think
that
covers
all
the
base.
Repo
things
there
are.
There
are
two
things
that
are
ready
to
rumble
or
kind
of
ready
to
rumble
in
the
make
target
for
the
package
split.
You
know
I
made
one
small
comment
on
that
pr,
but
besides
that
it
looks
fine
for
me,
I'm
happy
to
I'm
happy
to,
or
you
can
merge
it
at,
will,
there's
a
validation
for
the
package.
Composer
files
actually
didn't
even
realize
that
pr
had
been
opened.
Is
that
ready
for
review
team?
C
C
C
C
Yeah
just
remove
it
and
then
merge
it
up
here.
That's
fine!
The
validation
for
the
package.
Composer
files
is
that
ready
for
review
number
865
in
the
contribu
repo.
I
I
didn't
even
see
this
one
get
open.
So
I'm
sorry
that
I
missed
that
and
I
will
review
it
after
this
meeting.
E
I
think
it
is
it's
the
same
as
for
for
the
main
repository.
C
C
E
So,
but
regarding
the
validation,
well,
that's
basically
like
a
sanity
check
or
a
lint
linting,
the
composer
files.
What
I've
been
working
until
now,
I'm
nearly
finished
is
on
actually
installing
all
the
packages.
So
if
we
go
into
open
telemetry
php,
we
have,
I
think,
six
packages
right
now
there
and
then
I'm
working
on
this
dev
tools,
thing
which
we
post
on
the
open,
telemetry
php
organization
to
basically
resolve
all
of
the
composer
factory
packages.
E
We
have
for
the
http
on
clients
and
stuff
like
this,
so
it
took
a
little
bit
longer,
but
it's
working
now,
I'm
just
just
in
the
process
of
finishing
some
stuff
up
and
then
we
actually
yeah.
We
can
test
that
every
package
is
installable,
so
that's
basically
it's
taking
you
know
resolving
all
the
packages
and
then
creating
local
repositories
by
linking
and
so
on.
It's
a
little
bit
a
bit
more
tricky
than
just
validating
the
files,
but
it
works
now
and
then
yeah.
E
E
Yeah
we
can
see-
maybe
I
put
it
in
another
repository.
It
could
be
useful
for
other
projects
as
well,
but
for
now
we
have
it
and
we
can
make
sure
that
all
the
time
every
package
acts
well
actually
installable
and
with
all
the
dependencies
we
have.
E
And
that's
why
how
our
phones
this
one?
You
know
error,
which
I
did
a
pull
request.
I
think
two
days
ago.
E
Yeah,
you
emerged.
It
was
basically
just
a
left
over
some
packages
that
I
oh
yeah
in
the
package
and
left
the
old,
the
old
name
in
there,
but
it
of
course
doesn't
exist
and
why
it
did
never
exist,
so
it
wasn't
installable.
It
was,
I
think,
the
contrib
package,
so
but
everything
else
now
works
and
installable
and
yeah.
Once
that
is
finished,
I
will
create
a
pull
request
to
integrate
this
call
into
the
ci,
and
then
I
can
actually
close
this
ticket
for
the
packages.
E
Yeah
and
I
will-
or
we
can
add
some
more
stuff-
you
know
some
utility
things
to
this
dev
tools.
So
if
you,
if
somebody
has
any
idea
which
we
can
add
some
whatever
you
can,
could
edit
there
it's
just
a
composer,
some
composer
commands.
So
I
compose
a
console
application
and
if
you
want
to
know
how
it
works,
then
you
can
just
have
it.
E
E
E
Adding
that
to
the
meeting
notes-
oh
yeah
yeah,
one
things
to
add
this
because
brad
and
bob
and
me
had
some
or
I
asked
some
conversations,
so
this
is
actually
using
some.
E
If
you
go
into
the
workflows
in
this
patch,
it's
basically
you
using
a
reference
of
a
workflows
which
I
put
in
this
repository
as
well,
which
we
can
add
which
actually
could
reuse
for
the
main
packages
as
well.
E
So
we
might
do
that
and
also
I
created,
I
don't
know
well
if
we
want
to
use
it.
I
created
like
a
a
meta
package
for
all
the
development
dependencies
we
have
so
meta
package
is
basically
just
a
composer
file
which
just
references
other
packages
and
doesn't
have
any
anything
else,
but
maybe
we
want
to
use
it
as
well
for
the
for
the
main
repository.
So
we
don't.
E
E
That's
perfect!
Yeah,
because
very
often
I
found
or
we
changed
something
in
this
regard
or
in
the
main
repository,
and
then
we
forget
about
the
contract
repository.
So
when
we
have
some
some
single
point
of
truth
so
to
speak,
then
we
could
just
change
everything
there
and
then
we
will
be
changed.
E
Yeah
and
that's
just
some-
I
want
to
make
some
effort
that
we
don't
have.
We
have
left
redundancy
and
less
you
know
stuff.
We
have
to
change
if
we
don't
have
to
change
everything.
C
To
do
an
author,
an
authentication
service,
I
believe
timo
is
volunteering
to
do
that
and
it'll
be
really
neat
to
be
able
to
use
that
project
in
order
to
be
able
to
see
all
the
different
languages
working
together
in
one
little
shop.
That
is
docker
compose
installable.
So
if
you
want
to
check
that
out
I'll
post,
a
link
in
the
chat
for
github.com,
open,
telemetry,
open
telemetry
community.
B
C
F
So
bob
this
is
adele.
Can
you
hear
me
yep?
We
can
hear
you
yeah
yeah,
so
I
don't
have
any
specific
topic,
probably
next
time
onwards,
if
I'm
active.
Definitely
I
will
raise
some
of
the
point
just
want
to
introduce
myself,
I'm
from
cisco
and
the
sanket
produced
kishan
and
amber
who
who
normally
contribute
in
the
php
community.
They
are
from
my
team
and
they
report
to
me.
So
as
a
part
of
this
collaboration.
Definitely
I
would
like
to
be
more
active
on
the
php
community.
Apart
from
php.
F
My
team
is
also
contributing
on
the
python
hotel
community
and
on
the
paper,
server
c,
plus
plus
sdk
community,
as
well
just
wanted
to
give
you
the
context
and
also
this
hotel
demo,
which
you
are
talking.
Definitely
some
contribution
is
going
on
c,
plus
plus
sdk
site
from
my
team,
but
if
there
is
anything
specific
on
the
php
you
can
and
other
people
can
definitely
be
available
to
contribute.
C
Awesome
aj,
thank
you
for
introducing
yourself
and
we're
very
happy
to
have
your
team
members
working
out
with
us.
Thank
you.
E
For
the
demo
that
should
be
rather
straightforward,
so
it
already
made
a
plan
with
it
so
from
the
requirements
it's
easily
to
be
set
up
with
api
platform.
So
a
lot
of
stuff
wouldn't
even
need
a
code
or
coding
and
small
and
generating
the
api
and
stuff
like
this
shouldn't
be
much,
but
if
there's
any
other
need,
then
of
course
I
will
reach
out.
E
C
C
E
E
Yeah,
okay,
maybe
this
week
brett
we
find
a
a
date
to.
We
can
have
a
date
regarding
the
configuration,
but
at
the
moment
I'm
still
a
little
bit
busy.
So
I
will
just
I
will
reach
out
to
you
but
yeah.
This
is
something
we
should
have
for
beyonce.
A
E
G
E
Yeah,
basically,
there's
a
command
for
me,
that's
from
december
the
8th.
So
that's
basically
a
lot
of
what
I
was
talking
about.
E
It
puts
in
better
words
what
I
was
trying
to
explain
a
few
minutes
ago.
So
there's
actually
you
know
some
links
to
how
the
execution
and
php
works
and
why
it
can
or
cannot
be
a
problem
and
stuff
like
this,
so
with
a
shutdown
and
process
shutdown.
Now
that
works
yeah.
That
basically
explains
it
better
than
I
did
here.
I
think
okay.
E
Yeah
then
also
the
the
event-based
runtimes
and
a
link
there
like
swoon
swap
there
and
some
more
now,
so
you
can
have
an
idea
how
it
can
be
different
and
different.
H
I'm
currently
working
on
one
of
the
ticket,
and
that
is
issue
632,
so
I'll
be
creating
the
pr
today
only
in
next
star,
but
I've
raised
and
I've
put
a
comment
on
the
ticket
and
I
basically
have
a
problem
with
recreating
the
issue
so
that
you
can
have
a
look
and
reply
as
soon
as
possible
so
that
I
will
be
able
to
close
it
because
mostly
I've
done
the
development
part
testing.
I'm
doing
so.
For
that
I
needed
some
help.
I
put
the
comment
you
can
check.
H
E
Scenario
performance:
well,
it's
basically
when
there's
an
exception,
that's
just
I'm
formatting
the
the
normal
x
stack
trace,
for
my
exception,
to
be
a
little
bit
more
well,
I
didn't
even
write
it
a
little
bit
more
telling
or
I'm
not
quite
sure
what
the
initial
purpose
of
it
was,
but
it
basically
just
using
you-
can
just
create
a
create
an
exception
and
use
the
stack
trace,
which
is
there
and
as
to
test
how
it
works.
So
it's
basically
you
know
reformating.
These
stack
trace
to
be
more.
H
A
E
This
I
basically
just
created
the
ticket.
Sometimes
I
see
things
which
should
not
be
somewhere
or
could
be
refactored
somewhere
else.
You
know,
then,
because
yeah,
that's
not
really
the
the
span
or
the
I'm
looking
forward.
Give
me
a
second,
I
know
okay.
Well,
the
concern
of
this
ban
is
not
to
format
a
step
trace,
basically
like
this,
which
should
be
just
separated.