►
From YouTube: 2020-11-18 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
At
kubecon.
B
A
Have
you
been
working
on
the
logging
api
in
the
sdk
yeah
framework
yeah.
B
A
Yes,
I
was
just
gonna
mention
that,
like
walmart
on
the
collector's
side,
as
if
they're
going
to
go
to
stanza
right
and
then
they
talked
about,
I
asked
a
question
last
meeting
about
kubernetes.
A
B
Yeah
and
and
that's
yeah-
I
mean
it's
one
of
the
things
about
this
project
right
is
that
it's
it's
so
lego-like
that
there
are
so
many
ways
we
can
put
this
stuff
together
that
we're
all
kind
of
looking
at
different
pieces
that
are
kind
of
higher
priorities
to
us
so,
like
my
main
focus
is
in
has
been
I've.
I'm
actually
pulled
off
on
a
lot
of
other
stuff.
I
just
care
about.
This
has
been
in
getting
the
stuff
out
of
the
the
run
times.
I'm
a
I'm
very
developer
focused.
B
Them
off
somewhere,
I
I
don't
really
care
about
how
it
gets
shoveled
over,
but
then,
like
you
know,
we've
got
multiple
ways.
We
can
do
this
right,
so
you
can
have
an
exporter
that
goes
to
kafka.
You're
gonna,
go
from.
B
B
So
all
of
these
ways
that
we're
putting
together
have
you
are
a
topology
that
makes
sense
for
somebody
and
so
like
trying
to
juggle
all
of
these
at
once
and
say
what
you
know.
What
is
the
priority
that
really
has
to
happen.
Right
now
is
is
challenging,
especially
since
you
know
for
the
wider
project.
The
focus
is
really
on
getting
traces
and
metrics,
so
we're
kind
of
last
here.
So
what
group
do
you
work
in
at
splunk?
Then?
I
actually
work
in
development
or
developer
tooling.
A
Welfare,
tooling,
okay,
so
yeah,
obviously
walmart.
The
user
is
splunk
so
to
talk
to
sean
your
head
architect
about
our
lovely
ingestion
rates
that
spike
up
eight
or
nine
terabytes
in
30
minutes,
yeah,
yeah,
yeah,
but
again
you're
getting
conference
last
year.
You
know
we
got
your
product
manager
showed
us
dsa
data
stream,
processor
data
fabric
search.
We
did
a
poc
last
year
with
esp
and
evaluated
it,
and
then
we
operate
multiple
clusters
of
you
know.
Data
fabric
search
is
kind
of
like
something
we
already
had,
but
yeah.
A
B
B
I'm
not
sure
it's
so
much
like
we've
got
one
big
product.
We
do
have
big
products
that
are
coming
out
kind
of
that
time
frame.
But
it's
it's
more
a
matter
of
the
direction
I
think
for
us.
Obviously,
I'm
you
know
can't
speak
for
the
company,
but
but
it
there's
there's
a
long
stream
of
things.
A
We've
been
working
with
josh
and
signalfx
and
tracing
I
haven't
been
known,
but
oh
josh,
blessing
yeah,
I'm
so
similar,
fx,
okay.
A
So
what
what
so
you're
working
on
the
log4j?
So
if
we
have
people
that
are
using
box
for
j
framework
at
walmart,
then
they
just
do
an
api
call
and
then
we're
open,
telemetry
compatible.
B
Well,
okay,
so
this
is
all
from
my
perspective,
okay,
but
it's
not
so
much
about
being
open,
telemetry
compatible!
It's
as
much
so
much
as
getting
insight
into
what's
going
on
in
your
system
and
right
now
you
know
logs
traces
and
metrics
are
all
very
disparate,
so
kind
of
the
whole
open
telemetry
idea
is
to
pull
them
all
into
a
way
that
they
can
be
used
together.
I
think
there's
a
lot
of
power
power
in
that
so
kind
of
what
I'm
looking
at
with
log4j.
B
That's
just
the
first
of
a
lot
of
like
as
an
example.
We
don't
really
think
a
lot
about.
You
know
we
think
about
log4j,
but
in
java
you
tend
to
have
a
lot
of
dependencies,
and
every
library
that
you
depend
on
has
made
different
decisions,
including
on
what
logging
library.
B
So
if
you
like
log4j
but
you're
using
spring
spring,
uses
log
back,
and
so
it
would
be
really
nice
to
have
your
application
log
entries
coming
in
at
the
same
time,
in
the
same
way
as
spring
spring
spring
entries,
and
you
can
do
that
with
some
adapters
already
through
those
projects
but
being
able
to
also
have
the
traces
and
and
resource
correlation
that
becomes
really
powerful
and
so
that
that's
kind
of
the
piece
that
I'm
looking
at,
whereas
you
know
other
parts
of
of
even
my
company
are
really
looking
at.
B
How
do
we
have
one
jar
that
you
add
instrumentation
and
it
goes
in
and
adds
all
this
stuff,
which
is
an
entirely
different
way
of
going
about
it.
So
I
really
care
about
kind
of
I've
got
an
application
that
I
wrote.
How
do
I
look
at
what's
going
on
in
the
application
as
it
comes
out,
whereas
a
lot
of
other
people
are
looking
at?
B
I
am
running
this
application.
How
do
I
see
what
is
going
on
with
that?
You
know
so
there's
different
layers
and
different
perspectives
on
this,
and
it's
really
powerful
for
all
of
us.
I
think,
but
because
because
it
is
so
many
perspectives,
we've
got
a
lot
of
conflicting
priorities,
yeah,
because
my
stuff
is
what's
important
to
me,
and
so
of
course
it
should
be
the
top
of
the
list.
A
Because
I'm
just
thinking
because
right
now
at
walmart,
we
we
have
our
own,
we
call
them
strategy,
the
cloud
logging,
libraries
right
and
they
they
call
log4j
directly
right
and
then
then
what
is
it?
What
is
the
library
in
like
4j,
I
forget
for
the
library
inside
live
for,
like
4k.
A
B
There's
a
lot
of
really
interesting
stuff
in
there
I
think
yeah
and
like,
like
I
said,
that's
where
I
care
about
it.
So
you
know
that
I
find
that
really
interesting.
You
know
being
able
to
yeah
we're
all
familiar
with
with
stack
traces
and
kind
of
you
know
when
your
program
meets
its
final
breath
like
you
get
the
stack
of
what
what
it
was
doing
in
that
last
few
moments.
B
A
A
So
your
stuff
is
in
the
git
repository
something
working
so
we're
doing
a
new
rev
of
our
our
libraries
with
the
new
our
study.
Libraries
is
the
new
log4j
and
then
the
sync
and
async.
You
should
probably
try
it
out
with
your
stuff
too.
B
Maybe
you
know
the
the
stuff
we've
got
is
really
early
days.
Okay,
so
and
in
fact
it's
not
not
all
the
way
through,
yet
what
I've
got
as
a
trial.
Basically
it's
in
the
repository,
but
it's
kind
of
walled
off
so
that
it's
not
you
know
in
the
way
of
anything
else.
B
Is
the
capability
to
pass
off
a
log
entry
object
and
have
it
go
through
and
be
passed
off
to
exporter?
We
don't
have
any
exporters
yet
and
we
don't
have
anything,
that's
actually
using
this
api.
Yet
so
I
do
have
a
log4j
adapter
and
somebody
else
wrote
a
logback
adapter,
but
right
now
those
aren't
actually
going
into
this
this
api,
so
we've
only
got
the
middle
bit
and
both
end
bits
are
not
done
yet.
But,
okay,
you
know,
as
I
was
kind
of
trying
to
push
forward.
B
It's
also
what
we
were
trying
to
to
trying
to
finalize
all
the
the
metrics
and
traces
stuff
in
the
java
sdk.
B
A
B
So
so,
like
I
say
the
you
know,
this
is
all
a
lot
of
legos
right.
So
you
know
just
kind
of
looking
at
we've
got
import
plugins
for
log4j
and
whatever
else
we've
got
exporters
that
are
living
in
the
same
piece.
B
So
you've
got
an
exporter
for
splunk
and
for
whatever
else
that's
in
the
same
vm,
then
you've
got
an
exporter
that
will
send
it
off
to
the
collector
via
otlp.
So
that's
another
lego,
and
then
we've
got
exporters
from
there
to
to
from
the
collector
to
wherever
you
want
to
go
splunk
or
whatever
else.
So
we've
got
several
different
layers
that
are
there
and
that's
just
going
to
kind
of
come
in
in
pieces.
B
So
I
think
it's
going
to
be
a
little
while
before
we've
got
well,
it
may
not
be
that
long
before
we've
got
a
flow,
but
before
we've
got
the
right
combination
of
plug-ins
to
do
what
you
want,
maybe
a
little
bit,
and
I
I
fully
expect
that
there's
going
to
be
a
a
fair,
better
churn
once
we
start
getting
those
full
flows
going.
B
So
you
know,
as
soon
as
we've
got
a
flow
going
now
we're
going
to
start
looking
at
performance
and
that's
probably
going
to
make
some
big
changes.
We're
going
to
be
looking
at
you
know,
does
this
work
for
zap
and
go
and
if
it
doesn't,
then
we're
gonna
have
to
make
some
more
changes.
B
B
So
definitely.
B
You
know
and
that's
I
think,
that's
why
you're
seeing
the
focus
on
stanza
and
stuff
like
that,
because
that
is
what
is
kind
of
the
higher.
B
Like
my
stuff
is
really
important
to
developers,
but
it's
going
to
be
a
little
ways
off
on
that.
It's
going
to
be
a
much
shorter,
lift
and
higher
payback
to
have
something
like
stanza
that
can
plug
into
kubernetes
and
pull
the
kubernetes
events
and
put
those
into
the
the
the
exporter
and
pull
stuff
off
of
disk.
B
B
Like
I'm
a
developer,
and
I
want
to
have
my
application
start,
putting
everything
into
the
collector
so
that
I
you
know,
because
what
I
care
about
is
keeping
a
lot
more
structure
like
that's,
where
I'm
because
right
now
we're
dumping
a
lot
of
stuff
onto
disk
and
then
we're
pulling
it
back
out
with
regular
expressions
in
a
bunch
of
different
formats,
and
it
doesn't
make
any
sense,
especially
since,
while
we're
still
in
the
runtime
we've
got
a
lot
of
structure
that
we
can
use.
B
It's
actually
like,
if
you
think
about
a
lot
of
web
servers-
log
entry-
it's
probably
better
represented
in
the
csv-
it's
extremely
structured
right,
but
because
that
you
know
we
don't
tend
to
log
in
csvs-
and
you
know
it's,
you
know:
we've
got
the
apache
common
log
format,
but
it's
it's
that
kind
of
tabular
structured
data,
there's
a
lot
of
other
stuff,
that's
less
tabular,
but
still
very
structured
that
we
end
up
just
writing
out
in
key
value
pairs
and
we've
chosen
50,
bajillion
different
ways
to
do
that.
B
So
and-
and
that's
kind
of
where
you
know
with
open
telemetry
we've
got
now.
We've
got
a
standard
for
k
value
pairs,
so
we
can
actually
go
from
zap,
which
has
one
format
to
log
back,
which
has
an
entirely
different
format,
and
we
can
kind
of
look
past
that
textual
format
and
write
it
into
a
code
that
read
it
into
a
format
that
we
can
actually
use
later
on.
B
A
Okay,
all
right
that
was
good
chatting,
so
yeah
we'll
make
a
little
lego
block.
You
know,
okay,
this
this
this
this.