►
From YouTube: 2021-12-08 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).
A
B
B
B
A
Which
is
actually
the
one
in
the
open,
telemetry
google
calendar.
A
Telemetry
public
calendar
or
something
there's
a.
B
A
B
A
D
C
C
C
D
Yeah,
I
think
we
need
to
change.
I
think
we
need
to
change
it
on
the
the
the
places
all
of
the
they
change
all
the
meeting
links
with,
and
I
think.
C
C
The
amazon
usd
swan
region,
name
for
aws
was
like
super
barked
and
yeah,
and
all
of
our
stuff
is
only
hosted
there.
So.
A
Sorry,
I'm
just
typing
some
names
yeah.
We
can
hear
that
I'll
fix
that.
Oh
yeah,
that's
about
50
kilometers!
We
can
go
through
so
there
was
an
open
discussion
for,
but
that's
been
closed
for
a
question
from
an
upstream
sig.
I
think
about
updating,
instrumentation
library
to
instrumentation
scope.
That's
been
closed,
so
skip
over.
That.
B
Yeah
they,
I
actually
followed
the
discussion,
which
was
there
that
yeah
they
basically
just
wanted
to
change
the
name
from
instrumentation
library,
but
there
was
a
big
discussion.
Why
and
because
it's
all
about
logs
and
stuff
like
this,
and
sometimes
I'm
I'm
happy
that
logging
and
phps
were
easy,
because
this
was
rather
complicated
what
they
discussed
there.
A
B
So
I
just
I
only
saw
the
answer
from
the
on
javascript,
I'm
sick,
so
and
yeah.
I
don't
know
it
wouldn't
be
a
problem
for
us,
because
I
had
a
look
and
it's
basically
hidden
from
the
user
side
and
yeah
and
changing
that
with
php
sound
or
something
like
this
would
take
one
minute
max
so
wouldn't
have
been
the
problem
at
all.
It
wouldn't.
B
A
That
makes
sense,
should
we
go
through
any
open?
Pull
requests
is
the
first
thing
on
the
on
the
menu.
A
I
can
probably
talk
a
little
bit
about
the
global
logger
holder.
I
was
doing
some
work
from
that
last
night.
That's
pr495!
A
If
anyone's
following
along
this
is
where
we
add
some
logging
infrastructure
to
to
the
library,
because
there's
an
objective
in
the
in
the
spec
to
sort
of
log
log
errors
and
important
happenings,
but
not
cause
exceptions
which
could
break
user
code.
A
So
in
the
last
week's
seg
we
talked
about
having
well,
I
created
a
first
book
request
for
like
a
global
singleton
blog,
and
then
we
talked
about
using
whether
it
would
be
better
using
a
blogger,
a
web
interface.
So
I
started
down
that
road
last
night.
I've
got
something
that
sort
of
works
basically
works,
but
it's
I'm
finding
a
little
bit
tedious
to
use
a
logger
aware,
interface,
center
logo,
aware
trait,
just
because
I'm
having
to
inject
the
logger,
you
know
everywhere.
It
needs
to
go.
A
Although
I
understand
that
some
sort
of
dependency
injection
containers
could
take
care
of
that
problem,
and
probably
also
maybe
the
question
for
the
group
is
the
expected
behavior.
So
if
I
were
to
so,
if
we
go
all
the
way
up
in
at
least
in
trace,
we've
got
a
a
logger
provider.
A
If
I
were
to
set
a
logger
there
do
I
expect
it
to
propagate
that
logger.
You
know
all
the
way
down
through
any
spam
processes
and
any
exporters,
but
it's
probably
spam
processors
and
exporters,
which
are
the
most
important
bits,
but
do
it
do
we
expect
it
to?
A
A
B
Could
you
repeat
it?
What's
your
problem
or
or
a
simplification.
A
A
You
know,
say
in
a
trace,
holder
and
spam
processes
and
all
of
the
exporters,
particularly
if
you've
got
multiple
exporters
and
spam
tracers
spam
processors.
Sorry.
B
I
mean
we
will
use
the
factories
anyway,
so
I
don't
see
really
see
the
problem,
so
it
could
like.
Last
week
I
said
you
can
create
like
an
injector
or
whatever
it's
named,
and
that
could
just
take
care.
You
know
you
give
it
an
object
or,
and
the
looks
of
it
implements
the
interface
and
if
it
does,
it
just
injects
the
logos.
B
That's
basically,
that's
basically
what
a
lot
of
dependence
injection
does
as
well.
So,
for
example,
the
simple
symphony
one:
it
just
checks.
If
the
class
implements
this
interface
and
if
it
does
so,
then,
unless
otherwise
configured
it
will
inject
the
logo.
A
C
A
A
It
is
yeah
yeah,
so
there's
there's
an
existing
pr
with
a
static,
a
static
order
which,
which
is
simple
and
it
works,
but.
A
An
anti-pattern
yeah
first
passing
it
around.
We
hadn't
really
talked
about
having
a
di
container.
At
least
it's.
C
C
B
Well,
the
eye
containers
are
usually
used
by
frameworks,
or
there
is
kind
of
container
specification
which
is
very
simple
which
most
of
the
di
containers
or
builders
or
whatever
they
are
implement.
So
it's
basically
a
get
set.
Something
like
this.
It's
in
this
psr
php
standard
or
repository
psr
space.
B
But
you
it's
container
infrared
well
yeah,
it's
called
container
interface,
but
usually
there
are
a
lot
of
different
implementations
and
every
framework
handles
a
little
bit
different.
So
usually
it's
dependency
injections
done
by
coming
by
frameworks,
but
it
could
be
as
well
done
by
us.
We
could
use
a
simple
lightweight
dependency
injection
package
or
something
like
this,
so
it
could
be
a
solution
as
well.
So
what
I
yeah,
what
I
was
speaking
of
was
now
about
frameworks.
C
C
Yeah
I'm
trying.
I
know
we
used
one
at
work
called
like
pimple
or
something
and
yeah
I'm
trying
to
think.
If,
oh,
it
might
be
our
like
yeah,
it
might
be
our
framework
rapper
for
it,
but
I
thought
yeah.
I
thought
it
tries
to
like
store
an
instance
in
like
a
static
variable
and
just
reference
that
all
the
time,
but
I
might
be
wrong.
B
Usually
a
lot
of
times
they
have
nowadays
and
frameworks
have
something
like
auto
injection
or
something
like
this.
So
basically,
if
you
have
a
class
and
you
type
in
an
argument
or
something
in
the
method,
usually
in
the
constructor,
with
an
interface
or
the
class,
the
class
is
better
for
the
interface
you
have
to
define
something.
Then
the
framework
will
automatically
eject
the
the
instance
from
the
content.
Dependency
injection
container.
C
B
So
and
configured
like
this,
that's
not
how
all
frameworks
work
but
symphony
and
a
few
others
do
this.
I
think,
can
do
the
same
with
the
side
legs.
You
know
the
smaller
ones,
so
that's
kind
of
normal
nowadays
and
okay,
so
usually
yeah
you
don't.
You
will
sometimes
see
people
injecting
the
whole
dependency
injection
container.
But
that's.
C
A
B
C
B
Well,
for
a
lot
of
things,
a
good
place
to
look
on
all
those
are
these
psr
specs.
C
B
They
are
quite
old
now
I
think,
and
they're
usually
implemented.
So
in
a
lot
of
a
lot
of
use
cases,
they
are
even
some
more.
B
D
B
Yep
so
this,
for
example,
one
for
cash.
So
then
so
in
the
library
when
you
have
caching,
we,
you
know
just
expect
the
cache
interface
and
then
in
case
we
need
caching
and
then
people
or
frameworks
or
whatever
can
just
use
whatever
caching
solution.
They
like
the
same
for
the
logger
and
http,
and
there
are
a
lot
of
different
stuff
which
there
are
basically
interfaces
or
standards.
B
C
A
C
A
A
D
Yeah
and
then
I'll
you
y'all
do
that
and
well
actually.
No,
I
take
that
back.
I
have
made
you
both
mate,
like
you're,
both
part
of
the
hotel
or
the
php
maintainers
group,
so
you
should
be
able
to
review,
review
and
merge
prs
now.
It
doesn't
have
anything
to
do
with
the
code
runner's
file
that
just
has
to
do
with
the
organizational
structure
of
the
the
organizational
structure
with
the
people.
So
I
think
that.
A
B
Okay,
good
one
question
regarding
the
merging
I
have
you:
do
you
use
a
squash,
merge
two.
D
B
D
D
D
B
If
that
I
had
a
check-
and
it
looks
like
the
other
rules
are
inherited
when
you're.
B
B
Yeah,
because
when
you
take
a
look
at
the
groups,
approvers
or
triages
or
however
you
pronounce
that
in
english,
it
says
it
shows
the
people
which
are
effectively
in
there
and
then
it
says
three
three
through
child
groups
or
something
like
this.
C
I
think
the
one
question
kind
of
meta
question
I
had
on
this
pr
was,
it
seemed
like
polyus
was
using
the
new
changes
I
made
to
the
zipkin
stuff,
but
I
I
didn't
think
we'd
done
like
a
release
or
anything
of
a
new
version
or
or
we
did
did
that
happen.
Yeah.
D
C
D
That,
if
he's,
if
he's
developing
against
maine,
then
he
should
be
able
to
use
the
code
that
you've
merged
in
to
perform
that
he
wants
yeah,
especially
if
you're
in
a
development.
If
you're
in
a
development
environment.
C
C
C
Oh
yeah,
I'm
glad
I'm
glad
he
tried
it
out
and
found
that.
A
D
We're
very
overdue
for
a
release,
and
that's
why
I'm
saying
we
should
just
we'll
make
sure
that
we
get
all
of
the
open
pull
request,
except
for
that
one
with
there's
one
with
a
udp
transporter
and
that
one's
just
been
sitting
around
since
I
think
like
that,
and
that
person
hasn't
been
working
on
it.
So
I
was
just
leaving
it
open,
but
we
probably
should
close
that
at
some
point.
D
C
B
C
B
Yeah
and
the
php
8
0.1
thing
or
the
test
remaining
either
so
depends
on
when
we
can
actually
get
it
to
work.
So.
B
A
Yeah,
it's
probably
a
good
good
reason
to
have
you
know
not
too
many
dependencies
as
well,
because
of
those
who
we
might
find
in
the
future.
We
do
get.
We
are
held
back.
You
know
by
other
libraries,
particularly
maybe
some
of
the
smaller
ones,
yeah.
B
I
don't,
for
example,
I
don't
think
we
really
really
need
this
dns
power
browser
library.
So
that's
you
could
solve
this.
You
know
the
the
things
that
we
needed
for
you
consult
in
different
ways.
B
A
So,
as
far
as
getting
a
release
out
the
semantic
conventions
for
resource
and
trace
attributes,
that
might
be
a
good
one
to
to
get
in,
because
that
seems
like
a
good.
A
good
change.
A
I'm
not
trying
to
say
it's,
it's
a
big
change
that
gets
it
closer
to
spec,
I
suppose,
because
all
of
the
resource
constants
are
ford.
I.
A
So
maybe
that's
the
yeah.
We
get
that
one
get
that
one
merged
and
then
we're
okay
to
to
do
a
0.4
release.
B
Okay,
I've
got
one
question
and
bought
regarding
this
because
I'm
fiddling
around
with
the
package,
the
sub,
split
and
all
the
stuff,
what's
your
process
and
really
creating
a
release,
so
I
know
in
general
how
you
create
a
decent
github.
But
what
do
you
do
specifically.
D
I
understand
I
understand
your
question,
so
you
so
what
I've
done
for
the
previous
releases
is
I've
collated
the
list
of
pull
requests
in
using
like
a
one
line,
get
a
git
log
one
line
and
seeing
all
of
the
logs
that
we
like
the
individual,
logics
that
we
have
for
all
the
pull
requests
that
are
available
and
then
I'll
open
a
new
release
with
and
the
change
log
will
be
that
list
of
one
line
git
logs.
So
if
you
all
have
a
different
process
that
you'd
like
to
use,
I'm
more
than
happy
to
do
that.
D
I
just
did
this
because
that's
what
I've
done
in
previous
lives,
so
I'm
very
open
to
feedback.
For
that
I
don't.
We
don't
have
like
a
a
very
explicit
release
process
at
this
point
it
was
more
like
oh
shoot,
somebody's
asking
for
the
package.
I
might
as
well
build
one
and
that's
how
I
did
it
before,
but
if
anybody
has
input
in
how
to
build
them
or
if
there's
anything
that
other
people
would
like
to
include
in
a
release,
I'm
more
than
I'm
all
ears
more
than
happy
to
hear.
B
I
don't
necessarily
have
any
preferences,
but
since
I'm
working
on
automating,
you
know
the
packages
and
stuff.
I
was
just
thinking
about
automating
some
steps
of
this
as
well.
So
you
know
in
case.
B
B
Yeah
I
mean
for
three
releases
you
don't
have
to,
but
I
like,
I
don't
like
repetitive
tasks.
So
if
you
have
more
releases
in
the
future,
then
I
don't
well.
Maybe
we'll
need
a
button.
What
automation-
and
there
are
quite
some
like
php
libraries
out
there,
which
handles
some
of
the
let's
say,
boilerplate
stuff
regarding
releases
or
help
with
attacks
or
whatever
or
putting
together
a
lot.
So
there
are
different
solutions.
B
A
Other
thing
about
a
release
then,
which
is
packageist
so
we're
still,
we
still
haven't
linked
up
packages
so.
B
That
it
does
automated
oh
yeah.
I
updates.
I
have
another
question
regarding
this,
but
maybe
do
you
have
any
feedback
or
regarding
the
packages.
D
Not
only
yeah,
I
haven't
done
it,
I
haven't
done
enough
investigation
yet
and
how
to
make
this
whole.
I
know
that,
what's
his
face,
no
fuka,
the
guy
who
was
originally
working
on
this
library,
gave
me
access
to
our
like.
He
gave
me
ownership
to
our
packages
library,
but
I
have
again.
I
haven't
done
a
ton
of
digging
deep
into
that,
but
I'm
happy
to.
If
that's
what,
if
that's
my
task
for
this
week,.
B
D
Yeah,
I'm
I'm
looking
right
now
to
see
if
I
think
that
he
gave
me
access.
I
don't
think
that
I
a
lot
of
stuff.
Oh
yes,
I'm
sorry,
I
that
I
understand
your
question
now.
I
think
I
was
mis.
I
was
misinterpreting
it.
I
know
that
I
requested
packages,
access
for
open,
telemetry
and
I
haven't
heard
back.
C
D
D
You
know
that's
a
really
good
question:
I'm
not
actually
certain
there's
so
essentially
how
it
works
from
like
an
open,
telemetry
organization
perspective
is
there
are
there
are
like
open
to
limit,
so
there
are
open
telemetry
maintainers
for
each
sig,
so,
like
every
language
has
one
and
then
there
are
maintainers
for
the
spec
there's
maintainers
from
the
website,
and
then
I
think,
there's
just
like
general
admins,
and
I
I'm
assuming
that
the
general
admins
are
the
people
that
would
that
would
be
the
ones
that
are
approved
that,
but
I
think
that
they
probably
haven't
approved
it
either
because
a
they
haven't
seen
it
or
b,
because
they
don't
understand
what
packages
is
like
the
rest
of
us
php
pros.
D
B
Because
this
one
is
important
for
us,
but
another
question
regarding
the
site:
I've
seen
that
we
cannot
add
secrets
to
the
repository
right.
B
Can
we,
or
so
in
case,
for
example,
if
we
have
to
split
something
I
may
have
to
use?
If
I
don't
want
to
create
a
complicated
solution
with
like
a
proxy
repo
in
between,
then
we
have
to
have
like
a
github
secret
or
something
in
our
ci
process.
So
we
can
push
the
packages
into
the
other
repositories
and
I
don't.
C
D
Yeah
they
they
might
not
have
enabled
that
by
default.
I
can
again,
I
could
maybe
I'll
just
do
it
two
for
one
here
I'll
see
if
we
can
get
secrets
enabled
in
our
repository
and
see
if
we
can
get
packages
enabled
in
our
repository,
I
know
the
right
people,
the
the
right
place
to
ask
about
that
is
the
hotel.
A
D
D
Does
anybody
else
have
any
topics
that
they
want
to
talk
about
today?
Oh,
maybe
so
that
we're
not
playing
who
calls
tmo
do
of
you
want
to
take
ownership
of
changing
the
code
owners
file,
so
I
can
improve
it
immediately.
D
A
D
D
Had
they
had
a,
they
had
a
really
bad
thing
world.
I
think
she
said
all
of
the
like.
Their
php
and
c
group
is
linked
together
for
obvious
reasons,
and
she
said
every
single
person
on
their
team
quit,
except
for
her
and
one
other
person,
so
she
hasn't
been
able
to
contribute
that
much
because
she's
had
to
do
like
triage
all
of
their
issues
lately,
but
she
said
that
they're
hiring
more
people
now
and
she's
planning
on
coming
back
to
the
project.
So
I
would
leave
her
time.
B
C
B
D
B
Was
I
another
day
I
will
tell
you
the
story
about
it,
but
basically,
when,
when
you
guys
had
to
have
a
look
in
the
locks
I
had
to
too
because
everything
went
down
if
medicine
didn't
work.
B
D
B
Yeah,
do
you
don't
have
to
tell
me,
but
this
system
was
developed
by
outsourced
company
which
haven't
been
really
competent?
They
actually
they
created.
They
sent
first
the
request
to
the
front-end
server,
where
a
template
was
generated
which
was
later
dropped
because
then
later
they
implemented
using
the
templates
of
the
template
system
for
mailchimp,
but
the
template
was
still
generated.
It
was
just
you
know,
generated
and
then
dropped,
and
this
was
just
one
of
the
strange
things
that
the
system
does.
D
But
that
sounds
like
a
very,
very
strange
dependency
tree,
but
it's
not
as
crazy.
Well.
D
D
B
Yeah,
I
will
have
a
check
of
something
yeah
for
the
readme.
I
will
have
worked
on
this,
but
I
will
create
a
pr
just
for
the
technical
stuff
probably
tomorrow,
and
then
we
can
see
how
we
can
work
on
the
more
opinion
and
united
things.
You
know
yeah.
B
D
D
Read
me,
but
that's
something
that
I
think
that
will
come
to
me
that
will
come
more
along
the
along
the
time
frame
of
of
when
we
do
like
our
general
availability
release
or
at
least
a
beta
release,
because
right
now
I
think
most
of
the
people
that
are
consuming
our
library
are
also
developers.
I
don't
think
we
have
a
lot
of
end
users
using
it
yet
so
I
think
it's
important,
but
not
urgent,.
B
I
think
so
in
another.
I
have
a
little
bit,
I'm
concerned
about
something,
and
this
is
those
integration
guides
and
especially
the
well.
How
can
I
say
this
nicely.
B
Okay,
this
is
very
dangerous.
The
way
how
this
integration
is
shown
there.
So
so,
first
of
all,
I
I
don't
think
personally,
it's
it's
integration
guides
it's
more
like
a
tutorial
because
it
really
starts
with
bees
and
flowers.
You
know
you
have
to
set
up
all
this
stuff,
so
I
I
as
a
senior
or
even
senior
senior
developer,
don't
need
all
this
stuff,
it's
nice
for
junior
devs,
I
understand,
but
if
I
ever
see
an
integration
guide,
I
don't
want
to
learn
how
to
set
up
symphony.
B
You
know-
or
I
probably
know
all
this
stuff.
I
just
want
to
know
how
it's
integrated,
so
that
doesn't
mean
that's
bad,
but
I
wouldn't
call
this
an
integration
guide
or
something
it's
more
like
a
tutorial
and
that's
fine.
You
know
I,
I
really
know
junior
deaths
which
for
a
woman
this
was
it
would
actually
be
good,
but
the
the
way,
the
way
this
integration
is
done.
It's
not
really
integrating
it's,
basically,
not
using
any
integration
point
of
the
framework
at
all,
but
just
hacking.
B
You
know
putting
the
the
the
tracer
and
the
global
scope
and
the
really
the
global
global
global
scope.
You
know
like
in
you
would
in
javascript,
write
just
in
document
or
something
with
a
variable
and
then
pulling
it
out
from
there
and
expecting
it's
actually
there
and
that's
really
this
is.
I
don't
want
people
to
follow
this.
You
know
what
I
mean.
What
my
problem
is.
D
Totally,
I
think
the
the
story
behind
that
is
the
person
who
wrote
those
guides
was
a
very
junior
death
and
I
think
he
was
trying
to
be
as
verbose
as
possible.
I
think
I
agree
with
you.
It's
like
it's
a
careful
balance.
I
think
that
we
could
have
a
tutorial
and
an
integration
guide
separate.
I
agree
with
you,
however
yeah
I
think
that's.
B
Just
the
the
solution
is
basically,
basically
it's
not
specific
to
any
of
the
two
frameworks,
because
it's
using
nothing
of
the
frameworks.
You
know
it's
just
basically
patching
the
front
controller,
the
index.php
file
so
which
hard
which
well
it's
at
least
questionable.
I
would
say
I
would
never
do
this
and
would
never.
B
D
D
So
it's
probably,
I
think
that
you're
right,
I
think
we
should
have
a
hello
world,
open,
telemetry
php
at
some
point,
and
then
we
should
have
an
integration
guide
and
those
should
probably
be
segmented
in
some
way,
but
I'm
not
sure
yet
how
that
is,
I'm
very
happy
to
hear
suggestions.
B
Well,
I
think
most
people
which
use
the
library
will
probably
have
some
understanding
of
how
tracing
works.
So
otherwise,
why
are
they
looking
for
a
library
which
does
this?
You
know
it's
not
like
they're
browsing
packages.
Oh,
I
have
take
a
look
at
this
library.
What
does
this
thing?
Oh,
it's
tracing.
What
is
this,
but
basically
they
probably
know,
and
they
probably
have
a
use
case.
I
want
to
use
open,
telemetry
or
tracing
in
general
and
maybe
but
that's
okay,
to
have
a
tutorial
so,
but
I
would
prefer
something
more
straight
to
the
point.
D
Sure,
and
and
playing
the
dev
playing
the
devil's
advocate
here,
just
because
a
lot
of
distributed
tracing
setup
happens
from
either
devops
people
or
people
that
don't
necessarily
have
anything
but
a
cursory
understanding
of
a
language.
So
I
agree
with
you:
it's
important
to
have
those
implementation
points.
One
of
the
beauties
and
one
of
the
downfalls
of
distributed
tracing
is
that
you
can
implement
distributed
tracing
for
a
language.
You
don't
know
a
lot
about
so
that
they
can
complete
your
system
flow.
D
But
alongside
that,
you
have
to
be
careful,
like
you
said.
Having
it
be
inherited
in
index.php
obviously
has
its
it's
scary
implementation
details,
so
I
think
we
just
have
to
be
as
explicit
as
we
can
about
that
and
make
sure
that
the
end
users
get
the
right
guide
that
I
need,
whether
it
be
just.
D
B
Yeah,
I
guess
for
for
the
use
cases
for
people
not
into
too
much
php.
We
are
not
quite
there
yet,
because,
there's
still,
you
know
we're
working
on
the
configuration
that
would
be
things
you
know
they
probably
don't
want
to
programmatically
configure
all
the
stuff
like
we
mostly
do
at
the
moment.
So.
D
B
Okay,
one
thing
I
have
on
my
list
are
there
in
general
plans
to
report
other
tracing
collectors?
Like
example,
do
you
guys
know
apache,
skywalking
or.
B
B
When
you
have
an
eye
on
somebody
in
a
negative
way,
you
know
you're
always
watching
what
this
person
does.
That
means
what
we
call
in
berlin
to
have
somebody
on
the
kicker,
so
keaton
is
like
watching,
so
it's
basically
yeah
like
you're,
always
suspicious
what
this
person
does.
That's
just
the
words,
but
there
are
a
few
other
collectors,
but
in.
B
Could
put
stuff
like
this
in
the
control
repository
right,
so
anything
else,
yeah
one
thing
is
there
possibly
do
you
guys
know
what
gherkin
is,
and
I
don't
mean
the
things
to
eat,
but
the
test-driven
development
language
yeah?
What.
B
Yes
or
yeah,
behavior
driven
testing.
Do
you
know
if
there's
anything
any
gurken
tests
out
there
for
open
television.
B
D
A
B
No,
but
actually
it
wouldn't
be
bad
because
then
you
can,
because
guardian
or
behavior
driven
is
better
to
test
against
specs.
You
know,
because
you
can
actually
create
tests
for
specs,
you
know
is
a
blah
blah
blah
and
that
and
that,
and
then
you
can
just
translate
it
or
create
the
tests
for
each
language
specific.
So
it
was
just
a
question
of
something
like
this
exist,
because
in
general,
that
would
be
quite
unhandy.
B
You
know
to
have
tests
for
every
language
and
then
which,
but
I
don't
wouldn't
want
to
write
all
the
tests
for
the
for
the
whole
spectrum,
but
maybe
we
could
create
some
create
some
for
us
just
for
the
general
things
you
know
testing
like
basic
outlines
of
the
library
how
to
use
them,
so
we
don't
have
to
then
I.
B
Into
it
whenever
that
time
and
then
maybe
create
a
pr,
and
if
you
don't
like
it,
then,
like
always,
you
can
put
your
thumbs
down
or
not
anything
else,
yeah
I've,
oh,
I
didn't
create
the
pull
request
yet,
but
I
have
updated
the
portal
on
photo
buffer
files-
oh
okay,
and
so
that
will
be
a
pr.
It
was
a
very
hard
work
for
me
took
days
to
do
this.
B
No,
but
what
I
what
I
saw
this
the
image
which
is
used
for
this
is
one
and
a
half
gigabytes,
large
and
yeah.
Like
already
the
last
time,
I
complained
about
big
images:
that's
why
we
have
the
alpine
images
now,
but
I've
seen
and
it's
a
little
bit
outdated.
It
hasn't
been
maintained
since
three
years
or
something-
and
maybe
I
can
just
create
a
issue
for
that.
B
So
whatever
whoever
wants
to
work
on
this
I've
seen,
there's
like
it's
called
buff
or
something
there's,
it's
like
a
tool
set
or
whatever
project
which
makes
creating
a
protocol
buffers
a
little
bit
easier
and
the
image
is
just
50
megabytes.
So
that's
not
too
much
so
we
don't
it's
not
so
super
important.
So
well
maybe
I
just
create
an
issue
for
whoever
is
interested
in
this.
D
Yeah,
I
think
we
need
to-
I
know,
at
least
from
my
perspective.
I
need
to
do
this,
but
we
need
to
keep
making
sure
that
we
keep
opening
issues
for
things
that
we
find
that
we
want
to
improve
not
only
for
documentation
stake,
but
also
some
new
contributors
can
use
them.
Now
that
you
all
have
maintainers
access,
you
should
be
able
to
add
all
sorts
of
tags
and
assign
these
to
issues.
Oh.
D
B
So
the
two
ones
I
closed,
the
one
with
symphony
you
didn't
want
to
close
for
some
reason.
So
after
one
month
I
did
it
and
the
other
one,
I
think,
is
justified
too.
So,
if
not,
you
can
still
reopen
it.
So,
a
lot
of
fun
anything
else,
no
substitute
yeah,
I'm
going
to
work
on
the
sub
split.
I
will
get
back
probably
next
year
next
week
next
year
as
well.
Well,
next
year
is
not
that
far
away.
So
oh
what
one
thing
I
have
do
we
want
to
maybe
create
a
pull
request.
D
C
D
B
Opinion
me
neither
so
I
can
see
on
one
hand,
you
know
if
somebody
creates
a
full
request.
The
first
request.
They
sometimes
don't
know
what
to
write
there,
so
they
could
be.
You
know
helpful,
so
they
know
what
they
can
write
on
the
other
side,
yeah
we'll.
B
We
can
maybe
word
it
like
this-
that
it
doesn't
have
to
be
used
like
this,
so
they
can
just
write
more
or
less
whatever
they
want,
but
you
know
just
have
some
like
in
the
the
issue
template
or
something
or
the
issue,
the
ones
for
enhancement.
It's
not
called
enhancement.
Help
me
out
feature
requests
features
in
boxing
yeah,
no
yeah,
there's
one.
So
you
can
still
override
everything.
It's
not
like
this
is
that
it's
a
form,
but
because
then
you
know
there's
this.
B
Well,
it
says
there
are
checks.
You
know
this,
I
don't
know
where
to
find
it
now
the
community
tab
and
then
oh
there's,
another
thing
regarding:
did
anybody
use
this
discussion
feature
yet
on
github
in
any
project.
D
We
haven't
used
it.
I
think
most
of
the
discussions
that
we've
done
have
been.
In
slack
I
mean,
if
you
want
to
use
the
discussions,
feature
that's
more
than
fine.
I
know
there
are
some
other
things
that
really
like
using
that.
I
think
it's
just.
It
really
is
just
more
about
where
the
community
wants
to
have
that
conversation,
at
least
to
me.
Well,.
B
I
think
for
some
things
it's
maybe
helpful
because
you
can
create
polls
or
something
like
this.
You
know
to
have
a
general
idea
what
people
like-
or
I
haven't
actually
actively
used
that
so
I
can
have
a
look
so
maybe
for
some
things
it
will
be
or
you
can
pin
things
there.
So
that's
just
really
I'm
reading
now
from
the
tab
there,
but
I
could
imagine,
for
some
things
might
be
useful
because
not
everybody
is
maybe
participating
in
slack
or
a
low
entry
entry
level
if
you
want
to
have
some
feedback
from
people.
B
C
B
C
B
D
I
also
just
go.
I
also
just
saw
my
php
cs
fixer
issue.
Php
cs
fixer
does
not
yet
support
php,
eight
one.
So
we'll
have
to
wait
on
that.
B
Well,
okay,
the
other
one
there's
the
simple
fixer,
which
actually
does
the
same
we
couldn't
have.
I
can
have
a
look
of
into
if
we
can
use
this
or
if
it's
basically
works.
In
the
same
way,
though,
so
there
wouldn't
be
much
to
to
be
changed
yeah
I
will
have
a
look
at
five
times
that
would
be
solutions.
Otherwise,
basically
it's
yeah
the
things.
B
Okay,
it's
not
so
super
pressing,
but
I
will
just
put
them
all
down,
so
I
want
to
look
into
this
anything
else.
No,
no!
No!
No!
No!
The
only
thing
is,
I'm
tinkering
a
little
bit
around
with
some
auto
instrumentation
stuff
and.
B
D
B
While
you
can
read
it
another
time,
it's
basically
there's
this
issue
that
we
don't
haven't
yet
implemented,
retry
logic
for
the
exporters
or
auto
exporters,
but
there
are
some
problems
with
that.
I
like
also
maybe
at
some
point.
We
need
something
like
a
demon
or
some
process
which
is
runs
outside
of
the
of
a
web
server
or
the
classical
web
execution
contacts.
A
D
Yeah,
so
the
the
idea
behind
an
agent,
I
think,
is
going
kind
of
going
away.
I
think
there
are
some
other.
There
are
some
other
sigs
that
have
agents,
but
I
think
that
in
general
we're
pushing
more
towards
using
the
collector
as
the
main
focal
point
of
of
trace
aggregation.
D
Yeah,
correct
yeah,
so
you,
the
idea
now
is
you
use?
This
is
what
we're
doing
at
mailchimp
to
mailchimp
intuit
we're
using
the
open
telemetry
collector
on
each
of
our
hosts
as
a
side
card,
but
you
can
also
definitely
do
it
in
kubernetes
as
a
side
car
or
you
can
run
it
as
its
own
standalone
service
or
whatever.
D
B
D
I
think
that
if
you
can't
reach
the
collector,
you
probably
you
most
likely
wouldn't
be
able
to
do
anything
with
the
traces
anyway,
the
the
collector,
usually
and
again,
this
is
like
a
very
generalized
case,
I'm
sure
there's
other
cases,
but
I
know
this
is
the
generalized
use
case.
Is
you
have
your
individual
hosts
that
are.
D
C
D
D
D
The
collector
is
generally
accepted
as
the
commonplace
methodology
to
take
distributed
traces
and
push
them
somewhere
like,
for
example,
we
at
mailchimp
are
taking
our
choices
from
our
monolith
and
pushing
them
to
google
cloud
trace
using
the
collector
and
the
collector
is
really
nice,
because
it
gives
you
a
lot
of
flexibility
and
how
you
deliver
those
distributed.
Traces
to
one
or
more
locations.
B
Okay,
then
we
actually
don't
have
to
think
about.
You
know
a
demon
or
something
like
this.
D
I
think
the
the
idea
behind
it
damon
is
that
was
really
big
in
open
census
and
open
tracing,
but
I
think
yes,
I
think.
C
D
D
And
it
was
bad,
it
was
just
really
we
I
dealt
with
that
a
lot
at
my
company
and
it's
just
yeah
and
the
so
to
give
you
more
historical
context.
The
maintainers
of
the
collector
also
were
the
maintainers
of
the
open
census
and
open
tracing
collectors,
and
they
took
a
lot
of
things
that
they
learned
from
those
first
two
collectors
and
did
a
lot
better
with
the
current
collector
and
another
example.
D
I
know
that
we
had
a
lot
of
problems
with
drop
spans
in
open
census
from
the
collector
perspective
and
I've
instrumented,
the
telemetry
collector,
not
yet
in
production,
but
it
takes
all
of
our
logs
and
aggregates
and
pushes
them
out,
and
I
haven't
seen
nearly
as
many,
if
any
errors.
Yet
from
that,
don't
quote
me
on
that,
but
it's
I
think
it's
a
lot
more
stable.
I
think
that
the
open,
telemetry
team
has
spent
a
lot
more
time
with
this
collector.
For
that
reason,.
A
Yeah,
I
think
my
personal
opinion
there
is
like
we.
We
could
make
things
a
lot
easier
for
the
php.
I
think
in
drop
control
and
suggest
that
you
know
use
one.
You
know
one
or
two
really
well
defined
protocols
to
just
get
things
to
the
collector
and
do
it
once
and
do
it
well.
So
if
you
want
to
go
to
new
relic,
then
you
reality
should
invest
their
time
in
having
a
really
good
integration
into
the
collector.
D
B
Well,
there
yeah,
we
should
still
keep
other
parts
open
because,
like
you
said
you
never
know
what
I've
worked
for
many
different
companies
and
they're
many
different
constraints
who
just
want
some.
Some
things
are
just
opinionated.
You
know
some
people
just
want
to
do
not
want
to
work
with
some
solution,
because
it's
written
in
that
or
that
language
or
they
have
you
know
they
have
to
use
something
from
that
vendor
or
whatever
there's
so
many
use
cases
and
different,
sometimes
over
complicated
things
which
happen
on
the
problem.
B
So
to
be
open,
my
I
mean
it's
called
open,
color
and
telemetry,
so
I
should
be
open,
but
yeah.
We
have
to
keep
keep
in
mind
that
not
everything
respect
with
best
practice
in
every
company.
I
can
guarantee
you
that.
B
Yeah
and
people
do
the
strangest
things
that
I've
seen
really.
I
could
you
know,
go
for
hours
and
talk
about
strange
codes
or
setups
or
architectures
I've
seen
so
it's
sometimes
it's
scary.
Sometimes
it's
funny.
D
B
Really
it
is,
if
you,
you
see
some
of
the
code
which
actually
runs
it
and
yeah,
but
yeah
I've
seen
this
my
last
company
with
the
agent
a
lot
was
really
really
really
bad
spaghetti
quote,
but
for
some
reason
there
were
nearly
no
box
in
the
whole
system
and
it
was
quite
quite
complicated
because,
but
just
because
the
product
team
was
so
good,
you
know
when
there
was
a
bot
they
pushed
for
to
be
fixed.
B
A
D
B
B
D
D
We'll
see
you
all
on
the
internet
and
if
you
haven't
I'm
going
to
ask
about
the
packages
and
the
secrets-
and
I
will
keep
up
with
all
you
on
anyway.
Okay.
B
Everybody
have
a
good
evening
day
morning
for
mr
mcbride.