►
From YouTube: 2021-10-21 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
C
C
C
Next
week
on
sunday,
I'm
blowing
up
every
single
zoom
meeting
we
have
with
the
hotel
and
recreating
them
so
that
they're
all
tied
now
to
one
zoom
account
sort
of
the
way
a
human
would
do
it
where
you
have
individual
zoom
links
for
each
meeting,
instead
of
using
zoom
instant
meetings
for
each
sig,
so
we'll
have
no
more
overha
overlap.
However,
it
does
mean
to
edit
the
calendar.
People
will
need
to
sign
into
a
special
google
account
because
that
won't
be
linked
to
zoom,
so
you
can
make.
C
C
So
we'll
still
use
the
same
syncer
I
think,
but
I
think
automatically,
because
if
you
go
on
the
zoom
recordings
today
in
the
if
you've
admin
credentials
in
our
zoom
accounts,
it
just
says
like
zoom
personal
meeting,
zoom
personal
meeting,
zoom
personal
meeting,
the
new
one
should
be
named
because
it's
linked
to
calendar
so
another
nice
side
effect.
The
only
real
downside
is
that
to
edit
the
zoom
links,
people
have
to
sign
in
to
the
special
google
account,
even
if
their
own
account
is
edit
permissions.
C
C
C
C
C
A
A
D
I
thought
you
were
yeah
yeah.
I
think,
and
I
asked
you
if
you
wanted
to
to
go
on
there
and
you
said
yes.
A
D
If
so,
kind
of
so,
we
have
a
kind
of
you
know,
arguably
a
big
task
in
front
of
us
of
specifying
how
a
framework
for
ebtf
would
how
to
compose
one
right.
What
are
the
pieces
that
we
want
to
to
offer
as
a
framework
right?
Is
that
and
yeah,
and
so,
if
we
kind
of
how
do
we
start
tackling
that
is
kind
of
yeah?
How
do
we
break
it
down
so.
B
I
was
going
to
say
the
same
thing
I
think,
obviously
by
the
congrats
on
getting
the
announcement
out
glad
that
went
through,
but
I
think
we
were
very
rushed
before
I
think
now
we
can
kind
of
now
that
everything's
kind
of
kubecon's
past
and
everything
I
think
we
can
now
take
the
opportunity
to
kind
of
like
actually
dive
into
some
of
those
open
questions
that
we
had
from
before
figure
out.
B
D
That
sounds
like
a
great
plan.
Do
you
so
so
we
could
do
this
by
you
know,
maybe
getting
like
a
you
know
having
a
series
of
tech
talks
on
on
like
each
so
I
I
can
give
like
a
talk
on
on
the
flomo
contribution.
I
mean.
Maybe
you
have
one
on
pixie
or
so
we
can
do
it.
You
know
you
know
horizontally
or
vertically.
We
want
to
pick.
D
You
know
one
one
piece
of
supporting
technology
and
talk
about
that
at
a
time
or
one
kind
of
whatever
like
avert
a
vertical
product
right
like
the
pixie
or
or
we
want
to
take
like
a
horizontal
technology
like
like
korey
versus
loading
kernel
headers
talking
about
like
how
do
you,
how
do
you
compile
code?
How
do
you
load
it?
What
is
the
protocol
of
of
carrying
data
from
kernel
space
to
user
space
like
we
can
do
it
like
in
pieces
of
like
build,
build
kind
of
the?
B
Yeah
my
take
on
it
is
probably
is
easier
to
start
at
least
a
vertical
way,
because
I
think
we
all
probably
come
from
fairly
like
different
approaches.
Once
we
get
into
the
nitty-gritty
of
it
right,
so
it
would
be
kind
of
cool,
I
think,
to
kind
of
say:
well,
here's
how
flo
mill
does
it
here's
the
interfaces
we
have,
we
can
do
pixie
and
then
we
can
do
whatever
other
projects
are
out
there
and
then,
once
we
all
kind
of
have
a
better
idea
of
what
the
different
frameworks
look
like.
B
I
think
there
is
still
the
open
question
or
maybe,
like
I
said,
there's
still
the
open
question
of.
Maybe
the
decision
after
going
through
that
exercise,
is
it's
too
difficult
to
kind
of
try
to
hash
everything
into
one
unified
grant
framework,
and
we
just
want
to
use
the
open,
telemetry
api
and
just
kind
of
go
that
approach.
But
I
don't
think
we
can
answer
that
until
we've
taken
a
good
look
at
the
different
things
that
are
out
there.
B
We
should
also
come
up
with
a
list
of
what
other
projects
are
out
there,
like.
Obviously
we
know
about
fundamental
pixie,
but
we
should
make
a
serious
effort
to
kind
of
get
other
like
jonah.
You
have,
you
probably
said,
you're
starting
to
work
on
some
of
the
dpf
stuff.
So
it
would
be
great
to
get
your
thoughts
if
you
guys
have
any
code
like
what
you're
thinking
in
terms
of
architecture
and
other
projects
too,
so
that
we're
being
like
really
holistic.
With
this.
A
B
B
So
I
mean
do
you
have?
Do
you
have
more?
I
gotta
go
nothing!
That's
popping
like
immediately
into
my
head,
but
but
there's
definitely
a
lot
going
on.
B
Yeah,
let
me
think
about
like,
let
me
think
about
it.
If
anything
comes
up,
we
can
invite
them
as
well.
B
Kind
of
there's
so
many
projects
that
coupon
we
can
just
go
through
that
list
and
see
which
ones
we're
in
kind
of
the
observability,
plus
evpf
space
and
kind
of
invite
them
in.
B
D
And
so
what
we're,
when
let
me
kind
of
are
there
use
cases
that
we
know
that
we
want
to
tackle?
That
is
another
thing
that
I
thought
we
might
kind
of
cr.
If
we're
creating
kind
of
a
kind
of
a
list
of
of
things
we
we'd
want
to
talk
about.
Like
so
is,
is
there
other
than
npm
I've
heard
of
tracing
recently
profiling
tracing.
D
B
Profiling,
I
mean
depends
what
you
mean
by
tracing.
Is
such
an
overuse
or
overloaded
term
there's
yeah?
We
can
consider
like
apm
sort
of
stuff
as
well.
There
was
also,
I
think,
the
cpu
stats
memory
stats,
all
that
stuff,
which
you
have
a
beta
of.
B
And
yeah
profiling,
we
have
a
dynamic
what
we
call
dynamic
logging
feature:
that's
based
off
of
eppf,
again,
not
sure
these
are
just
I'm
throwing
this
stuff
out
here,
I'm
not
convinced
yet
that
that's
the
right
thing
for
it
to
fit
into
hotel,
ebpf
or
not
right,
maybe
maybe
not,
but
we
should
just
like
talk
through
these
things
in
this
meeting.
I
think
it's
you
know.
Profiler
is
another
one
which
I
think
is
more
up
in
the
air.
B
C
C
Or
at
least
logs
is
in
beta,
but
but
that's
one
like
like.
I
know
that
actually
josh
is
here.
I
know
that,
like
when
I
was
at
google,
the
google
folks
on
that
team
were
keen
to
to
bring
profiling
into
hotel,
and
I
know
it
splunk
eventually.
The
same
will
be
true,
but
yeah.
There's.
E
Yeah
yeah
and
I'm
not
allowed
to
do
due
diligence
for
it
so
anyway,
that's
that
is
coming,
so
you
should
totally
be.
If
you
want
me,
I.
B
B
I
think
we
want
to
keep
this
focused
on
not
the
details
of
what's
happening
necessarily
inside,
but
more
like
the
interfaces
of
the
different
features.
Ebpf-Based
features
that
we
have
and
how
they
plug
in
so
things
like
what
context,
what
the
output
format
looks
like
and
things
like
that
and
how
we
could
create
a
plugable
architecture,
and
we
could
start
with.
If
you
want
to
start
a
softconstant,
you
could
and
then
I
can
do
pixie
and
then
we
can
that'll
probably
take
several
weeks.
B
I
imagine,
since
we
only
have
30
minutes
and
then
we
can
go
from
there
to
some
other
projects
as
well.
D
I
can
start
us
off
next
week,
there's
like
a
a
couple
of
slides
and
you
can
discuss
the
architecture
and
the
different
components.
D
B
Same
same
with
pixie,
I
could
probably
go
on
for
hours,
but
let's
try
to
do
a
good
job
of
like
keeping
the
inter
just
sticking
on
the
interfaces
and
the
high
level
view.
B
D
I
started
we
can
start.
Yes,
I
was
going
to
give
a
talk
next
week
at
the
work
group
or
do
do.
We
want
her
to
do
this
as
a
doc.
B
I
said
doc,
that's
talk,
I
think,
like.
I
think
we
want
a
hashtag.
So
let's
don't
want
to
create
a
whole
bunch
of
work.
If
you
have
like
slides,
I
think
slides
and
stuff
are
always
helpful
like
if
not
like,
overdoing
it,
but
a
few
high-level,
slides
or
pictures
would
be
useful,
but
other
than
that,
I
don't
think
we
need
a
dock.
Yet.
D
E
Raise
have
one
one
question,
just
just
real
quick,
so
I
commented
on
the
bug
on
the
tc
discussion
around
the
contribution
and
like
further
work,
basically
go.
What
what
do
you
need
from
us
is
the
question:
do
you
need
anything
besides?
Just
like
yes,
it
looks
everything
looks
great
really
like
the
direction.
Keep
going.
C
E
Okay,
the
other
thing
is
we're
putting
together
a
state
of
open,
telemetry
newslettery
thing
with
like
roadmap,
because
hey
guess
what
we
get
asked
all
the
time,
so
I'm
gonna
be
asking
you
to
possibly
write
a
blurb
on
like
what's
happening
with
ebpf
and
where
we're
going
in
expected
timelines,
which
can
be
totally
open,
sourcy,
meaning
hey.
This
is
what
we're
investigating
over
the
next
six
months.
E
You
know
we'll
show
you
something
in
six
months
kind
of
a
thing
as
opposed
to
we're
going
to
have
a
delivery
right
deliverable,
I
should
say
yeah.
So
the
other
thing
just
so,
you
know
the
to
highlight
it.
The
we
do
expect
to
have
maintainers
be
a
little
bit
overloaded
with
say,
metrics
and
or
logs
in
the
near
future.
E
So
to
the
extent
that
you
need
their
time
and
attention,
let
us
know
how
we
can
help
make
sure
that
we
are
focusing
our
attention
very
well
and
that
each
kind
of
sub
community
is
is
operating
as
efficiently
as
possible.
So
it
sounds
like
given
the
plans.
I
saw
that
there
shouldn't
be
a
lot
of
overlap,
so
I
think
that's
awesome.
D
Right,
I
think,
if
you're,
if
you're,
asking
kind
of
what
are
the
bigger
discussions
that
we'll
need
to
have
outside
this
work
group,
I
think
both
jana
from
aws
kind
of
the
this
is.
This
has
been
kind
of
recurring.
I
think
jana
and
and
michael
from
datadog
and-
and
me
too
I
don't
know
immediately
like
I
don't.
D
I
don't
know
if
we've
talked
about
this,
but
there
is
a
question
about
how
to
represent
telemetry
right
now,
we're
outputting
logs
and
they
don't
have
a
schema
and
they're
the
the
amount
of
cpu
that
is
required
to
process
the
logs
might
be
higher
than
you'd
want
for
network
telemetry
you'd
want
to
process
a
little
bit
more
telemetry
with
less
overhead,
so
there
is
a
discussion
of.
Can
you
create,
like
maybe
a
binary
protocol,
a
schema
protocol?
D
E
Yeah,
google
actually
has
some.
We
have
some
folks
who've
been
working
on
envoy
who
added
logging
in
open
telemetry
that
does
this
and
we've
been
having
discussions
around
having
a
focused
structured
event
internally
and
trying
to
promote
that
so
more
than
happy
to
join
in
that
discussion.
I
I
totally
agree
that
that's
a
thing
that
needs
to
happen.
Yeah
yeah,
are
you
familiar.
E
Sorry,
you
can
start
with
me
and
I
will
reach
out
to
the
right
the
right
set
of
people,
I'm
not
the
correct
contact.
I
am
the
the
entry
point
if
you
will
for
things
open,
telemetry
and
google.
The
what
I
want
to
call
out,
though,
was
there's
an
otep
from
I
think
f5
on
multivariate
events
as
a
protocol
for
open
telemetry.
E
I
don't
think
when
I
look
at
that
that-
and
I
know
that
this
is
coming
from
a
set
of
network
engineers-
I'm
not
positive-
that
that's
fully
the
direction
we
need
to
go
or
actually
solves
this
particular
use
case
well,
but
I
think
it
solves
it
better
than
existing
otlp.
So
that's
worth
looking
at.
If
you
haven't
seen
it,
I
can
try
to
get
you
a
link
and
then
I
I
do
think
we
need
to
have
this
discussion.
E
So,
if
you're
having
a
discussion
as
part
of
the
ebpf
group,
I
will
make
sure
that
google
representatives
attend
they're
based
in
tel
aviv,
though
so
time
timewise.
This
might
be
awkward.
E
C
Yeah,
adding
on
to
that,
I
added
it
to
the
notes.
The
the
client
instrumentation
rum
sig
also
was
hopeful
for
some
kind
of
structured
event
type.
I
had
sort
of
told
them
to
hold
their
horses,
but
if
there's
other
parties
who
who
want
this,
then
we
should
have
them
all
come
together.
E
C
E
Right,
the
main
the
main
concern
here
is
actually
the
collector.
I
think
we
need
to.
I
underestimated
the
amount
of
work
the
collector
is
at
taking
in
signals
so
be
prepared
to
work
in
the
collectors
all
I'll
say.
But,
given
your
document,
I
know
you
are
so
it's
fine.