►
From YouTube: Diagnostics WG meeting - April 3 2019
Description
A
Okay,
so
welcome
to
the
node.js
community
Diagnostics
team
working
group
meeting
for
April
30
2019
mike
coffman,
couldn't
make
it
today
so
starting
up
the
stream
and
we'll
get
started.
So
following
the
agenda
posted
in
the
issue,
any
announcements
or
things
that
people
want
to
add
to
the
agenda
for
we
get
started.
B
Just
briefly,
I
finally
started
to
working
on
the
node.js
diagnostic
user
journeys.
What
I
promised
on
the
last
summit
my
plan
is
that
to
share
some
kind
of
draft
by
the
end
of
this
week
and
to
see
how
do
you
like
it?
What
do
you
think
about
it?
So,
basically,
this
draft
will
contain
the
structure
and
I
will
fulfill
want
to
use
a
Chinese
and,
after
that,
I
expect
some
feedback
that
even
worse,
to
continue
or
should
I
change
the
structure
as
on
something
to
share
it
by
the
end
of
this
week.
Okay,.
B
A
C
Can
just
mention
that
I
merged
the
PR
for
the
for
the
summit,
action
items
with
links
to
videos
etc,
and
if
people
want
to
to
add
their
their
slides
to
that
markdown
file,
links
to
slides
it.
It
is
a
folder
in
the
github
repo,
so
you
can
even
put
a
PDF
or
whatever
in
the
github
wrapper
with
the
actual
slides.
If
you
wanted
to
that
or
just
link
to
slice,
they
go
wherever
you
have
them.
If
you
gave
a
talk
at
the
summit
per
grade,.
A
D
D
I
know
I
I
guess,
since
everyone
is
talking,
oh,
but
they
just
did
just
say,
I'm,
probably
between
now
and
next
diagnostic
meeting
I'll
be
ready
to
have
like
perfetto
code
or
reviews
so
or
feedback,
unlike
the
initial
design,
so
I
I
think
there's
a
process
in
creating
like
a
maybe
like
a
for
under
the
no
Jess
organization,
I,
so
need
to
like
iron
out
like
the
book
files
and
stuff
on
that.
Yet
but
yeah,
hopefully
for.
A
A
B
A
D
E
B
Open
sense
is
Google
and
Google
and
the
Netflix
team,
because
we
are
planning
to
propose
some
changes
in
the
agent.
Ok,
it's
a
super
early
phase.
I
would
say
we
are
still
in
the
proof-of-concept
phase,
just
yeah,
just
a
heads
up
that
we
are
considering
to
use
that
as
an
instrumentation
for
the
stability
tracing
methods
right.
A
A
Interested
because
I'm
looking
at
it
in
terms
of
whether
it
should
be
what
we
recommend
for
the
IBM
cloud
as
well-
ok,
I,
didn't
I,
did
have
a
quick
chat
with
ally.
But
if
there's,
if
it's
already
been
held,
I
missed
the
boat.
But
if
there
was
like
a
discussion
and
how
you
get
involved
and
what's
going
on,
who's
been
interesting.
Yeah.
A
B
C
B
Yeah
in
in
a
really
nutshell,
it
uses
continues
local
storage
right
now
and
we
don't
really
need
it
because
we
have
a
full
control
over
instrumentation.
There
is
some
API
I,
don't
know,
maybe
it's
just
around
making
side
spans
which
doesn't
really
fit
our
use
case.
So
it's
true
really
just
these
basic
stuff,
okay,
interesting
things.
A
A
A
A
B
A
Think
you
know:
we've
discussed
that
at
the
diagnostic
summit,
I
think
we
have
a
way
forward.
In
terms
of
you
know,
we
need
to
get
more
things
torrefied
and
sounded
like
yang
was
gonna
work
with
I
forget
his
name,
but
does
the
guy
from
Microsoft
I'm
thinking,
maybe
that
one
we
can
take
off
the
agenda
for
now
and
you
tell
us
there's
something
that
we
need
to
discuss
the
meeting
itself
is
that
make
sense
to
people.
E
C
E
C
A
A
Think
I
was
still
half
connected
because
I
didn't
reboot
my
whole
machine
I
just
logged
in
with
something
else.
Okay
right,
so
we
were
talking
about
Freight
t27,
which
we
took
off
the
agenda.
So
next
thing
was
the
best
practices
guide.
I've
been
away,
I
was
bein
vacation
last
week,
so
I
haven't
had
any
time
to
work
on
that.
Peter
already
mentioned
he's
starting
to
work
on
some
of
the
use
cases
that
were
like
that.
Yes,.
B
A
E
A
A
The
next
one
is
157,
which
is
expectation
about
tier
support
and
I.
Think
we've
agreed
we're
gonna
drive
that,
based
on
the
best
practices,
so
you
know
we
need
to
get
one
of
the
best
practices
in
place
and
then
once
we
have
that
in
place,
we'll
look
at
the
tiers
to
make
sure
that
we're
properly
supporting
the
tools
they
call
it
so
otherwise,
nothing
more.
On
that
front,
a
stable
API
facing
cooks,
I,
don't
know
if
anybody
has
any
any
newer
report
on
that.
C
Not
specifically
on
basic
hooks,
it
was
yeah,
but
I.
Don't
think
that
we
actually
updated
the
issue
after
the
summit,
which
might
be
aquaria,
because
we
did
we
a
semi
conclusion
on
the
summit
about
not
trying
to
necessarily
reach
the
stable
API
facing
hooks
right
now,
but
instead
look
at
the
server
storage
as
a
first-class
citizen.
That
would
then
May.
C
C
C
A
A
A
A
One
thing
there
is
I
think,
as
we
see
some
initial
PRS
we'll
need
to
think
about.
You
know
the
what
we
the
approach
we
took
on
the
Java
side
at
IBM
was
like.
We
added
lots
and
lots
of
trace
points,
because
you
never
really
sure
what
you're
gonna
need
but
they're
off
by
default,
and
then
we
enable
them.
The
other
approaches
like
to
try
and
be
more
selective.
But
it's
then
again
a
pretty
good
idea
of
what
you
think
you
need
right,
so
I
think
I.
You
know
and
I've
been
talking
to
James.
A
B
A
D
A
A
A
There's
basically
different
categories,
so
you
could
in
turn
you
could
enable
specific
ones
or
specific
categories
and
I
think
they
each
also
had
like
a
level
associated
which
one
which
was
kind
of
like
a
cost.
Okay,
so
you
know
if
they
were
super
cheap.
There
were
probably
level
one
if
they
were
more
expensive
level
two
and
then
like
five,
it's
like
okay
and
this
one's
super
expensive.
So
you
could
also
enable
I,
remember
correctly
based
on
how
much
overhead
you
were
adding
by
turning
them
on.