►
From YouTube: Configure Group Conversation (Public Livestream)
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
Okay,
I
have
1500
UTC,
so
let's
go
ahead
and
get
started.
Welcome
to
the
configure
group
conversation,
I
am
Daniel.
I
am
the
product
manager
for
the
the
configure
group
which
currently
entails
the
configure
group
at
and
the
circles
circle
I'm
so
dealing
with
all
of
their
our
DevOps
coordinators
and
service
needs.
Let's
go
ahead
and
take
their
questions.
A
So
I
think
you
said:
that's
a
good
question
so
city's
asking
about
the
Jupiter
one
book,
so
one
can
deploy
Jupiter
in
your
cluster
and
then
you
can
create
executable
one
books.
So
we've
had
conversations
both
with
internal
groups
and
the
two
producers,
but
it's
mostly
the
use
cases
that
we've
met
so
far
is
that
it's
being
used
for
data
science
teams
just
as
a
regular
notebook
for
Jupiter,
we
have
not
come
across
users
for
executable
run
books.
A
We
think
it's
mainly
due
to
two
reasons
so
number
one:
it's
the
the
bar
of
entry,
like
all
the
stuff
that
you
need
to
write
even
with
something
as
simple
as
the
route
group,
it's
library
is
high,
so
we've
identified
there
that
having
more
examples
on
how
to
do
common
tasks
will
help
us,
that's
number
one
and
then
number
two
and
the
the
provisioning
of
access
for
large
teams.
So
you
know
people
still
have
a
hard
time
figuring
that
out
in
the
Jupiter
world.
How
would
we
provision
this
for
an
ops
team?
A
So
we
have
a
lot
of
change
with
our
infrastructure
team
in
how
the
notebooks
get
executed
and
all
the
different
systems
that
that
we
use.
So
with
that
being
said,
the
two
systems
that
we
have
that
are
kind
of
easy
to
write
notebooks
for
is
kubernetes
and
also
a
double
yes.
So
for
the
kubernetes
related
things,
I
think
that
we
could
start
there
and
I
think
that
we
could
use
it
for
our
kubernetes
related
notebooks
I.
A
Don't
think
that
we
have
any
yet,
but
when
we
do
as
we
move
more
and
more
to
kubernetes,
we
could
start
there,
but
other
than
that,
we
could
just
manually
write
all
of
the
queries
and
all
of
the
things
that
we
do
systems
wise
without
having
to
make
use
of
the
rubix
library.
As
you
know,
that's
optional,
that
that
is
kind
of
a
bit
of
work.
But
that's
also
another
good
sort
of
point
will.
B
Help
me
a
bit
here
because
we're
and
most
of
our
infrastructure
most
of
gitlab
comm-
it's
not
kubernetes
yet
and
we're
switching.
But
it's
gonna
take
some
time,
but
I
assume
these
REM
books.
Although
they
have
to
be
deployed,
they
run
on
kubernetes.
They
don't
necessarily
have
to
operate
on
kubernetes
infrastructure.
You
could
I,
don't
know
be
hosting
stuff
on
raspberry
PI's.
C
A
That
is
the
case,
and
what
I
meant
by
that
is
that
the
libraries
that
we
have
right
now
that
help
users
write
those
run
books
or
right
now,
specifically
for
kubernetes
and
for
AWS.
So
as
we
move
more
of
our
infrastructure
into
kubernetes,
we
could
make
use
of
what
we
have
today
to
write
those
farm
books,
but.
B
We
don't
we
don't
need
any
of
our
existing
infrastructure
in
kubernetes
right
if
we're
going
to
use
these
from
books.
That
has
to
be
separate
infrastructure
anyway,
probably
even
on
a
different
cloud
provider,
because
if
you're
gonna
do
debugging
like
you
don't
want
if
Google
goes
down
like
it
went
down
last
week,
you,
the
last
thing
you
want
going
down.
Is
your
monitoring
stack
and
your
red
book
stack
because
that's
how
you
can
make
changes?
So
in
my
mind,
although
it's
all
in
kit
lab,
we
should
have
it
deployed
on
separate
infrastructure.
B
A
B
We
don't
want
like
so
so
kid
lab
should
be
able
to
do
self
monitoring
it.
Lab
should
be
able
to
deploy,
run
books,
but
I,
don't
I,
don't
think
we
want
our
monitoring
and
our
run
books
running
on
the
same
infrastructure.
Like
suppose,
we
have
an
outage
of
some
sort,
a
load
balancer
VPN
that
goes
down.
We
get
flagged
for
something
like
all
of
these
failure
modes.
B
Like
you,
the
last
thing
you
want
going
down
is
your
one
string
stack
like
big
reason:
Google
got
in
trouble
last
week
was
because
they
were
using
the
same
network
to
do
diagnostics,
so
they
had
a
really
really
tough
time
recovering
from
that
so
I
wanna
make.
We
should
make
sure
that
we
do
use
it
on
different
things
like,
for
example,
in
Basecamp,
all
their
infrastructure
nodes
aren't
in
confluent
because
they
ship
a
SAS
product
and
if
their
sounds
fingers
down.
The
last
thing
they
want
being
down
are
their
notes.
B
A
Libraries
for
sure
that's
a
good
starting
point
yeah,
so
I
think
that
it's
just
a
matter
of
identifying
and
a
good
starting
point,
even
if
we
write
it
and
doing
it
there.
But,
yes,
it
is
viable
for
our
team
to
use
that
internally,
especially
now
that
we
have
the
git
integration
set
to
ship
in
twelve,
oh
you're
going
to
be
able
to
version
control
all
of
those
things
from
from
Jupiter,
so
it
is
viable
and
yeah.
It
is
something
that
we'll
follow
up
on
as
part
of
our
internal
customer
being
at
the
m-40
thanks.
It.
B
C
A
A
Maybe
we
should
chat
further
about
that
and
see
if
we
can
come
up
with
a
proof
of
concept
that
we
can
use
before
Prometheus,
which
I
think
it's
a
very
common
use
case
that
you,
if
you
have
a
downtime
scenario,
and
you
execute
a
couple
of
queries
or
a
couple
of
changes
or
your
infrastructure,
and
then
you
want
to
query
achieve
that
change,
something
so
yeah.
It's
interesting!
It's
something
a
whole
up
on
cool
thanks.
A
B
There's
a
follow-up,
I'm
gonna,
taking
some
notes
at
Eric,
because
I've
seen
this
hesitancy
before
so
maybe
we
should
document
like
we
want
to
use
the
same
product
but
different
instances.
We
have
the
gitlab
become
service
to
other
people,
but
then
we
have
get
local
monitoring,
which
should
be
separate,
get
lot
back
home,
run
books
which
should
be
separate
com
CD,
which
probably
should
be
separate.
B
And
if
we,
if
we
have
this
confusion
and
we
get
things
like,
oh,
we
have
to
use
kubernetes
in
order
to
use
the
remm
books
or
we
can't
use
gila
monitoring.
Because
what,
if
get
lamp,
is
down-
and
that's
because
that's
something
I
heard
from
the
infrastructure
team
and
that's
that's
preventing
us
from
using
our
own
products,
which
we
should
and
I
think,
for
example,
for
monitoring
its
held
us
back
that
we're
not
using
our
own
product.
B
C
There's
what
another
on
the
Prometheus
API?
We
also
need
to
make
sure
we
have
the
correct
sort
of
network
access
between
the
cuban
aires
cluster
and
prometheus,
which
which
isn't
reveal.
But
you
know
it's
not
just
a
matter
of
getting
the
right
library
and
pulling
the
values.
We
also
need
to
be
able
to
access
the
prometheus
values
from
the
you
know
where
the
jupiter
kernels
are
being
executed.
B
A
So
let
me
share
my
screen
now:
walk
through
the
docks
here
and
I
will
tell
you
what
it's
all
about
all
right.
It's
up
here,
let's
go
to
people
that
run
books,
and
here
I
see
the
docks
is
the
second
link,
and
so
basically
what
run
books
are
all
about
are
a
set
of
procedures
that
generally
an
operator
will
conduct
at
a
given
time.
So
they
are
very
popular.
A
Let's
say
in
in
in
cases
of
downtime,
so,
for
example,
it's
system,
X,
it's
currently
down
an
operator,
would
bring
up
that
one
book
and
follow
through
a
series
of
steps,
historically
that
those
are
kind
of
text
only.
So
if
you
saw
here
when
I
executed,
this
Google
search,
the
run
books
of
repo
is
kind
of
the
first
thing
that
came
up
our
own.
A
Basically,
and
here
you
can
kind
of
see
that
these
are
on
called
one
books
and
they
have
kind
of
a
set
of
procedure
of
what
to
do,
and
it
tells
you
go
to
a
certain
slack
channel.
Do
this
do
that,
so
the
the
value
of
having
executable
run
books
is
that
you
visit
your
run
book
once
and
you
don't
have
to
leave
your
run
book
to
follow
all
of
those
steps.
A
So,
instead
of
tweet
this
and
you
having
to
go
to
tweet
deck
and
send
out
a
tweet,
you
would
configure
that
access
within
your
run
book
and
simply
click
a
button.
Maybe
enter
some
text
with
your
your
run
book
and
do
that
then
you
move
on
to
the
next
step.
You
know:
go
to
slack
Channel
and
go
to
this
channel
and
do
this,
so
you
would
do
the
same
thing
with
you
within
your
run
book.
A
So
I
was
chatting
about
the
end
of
the
high
level
of
entry
there
at
the
bar
of
entries
honey
because
you
kind
of
have
to
can
write
all
the
code
above
old
systems
that
are
interacting
with
one
another.
So
the
beauty
of
our
run
books
is
that
we
it
with
the
Rubik
library
here
and
it's
an
open
source
library
that
has
it
makes
it
easy
to
write
commands
specifically
specifically
for
AWS
and
kubernetes.
A
So
things
that
would
be,
let's
say:
maybe
a
hundred
lines
of
code
become,
maybe
you
know
five
to
ten
lines
of
code
and
it
has
calls
for
all
the
most
common
things
like
plotting,
metrics
from
clad
watch
and
things
like
that,
and
it's
very
easy
to
get
started
with
good
lab.
The
only
thing
that
you
need
to
do
is
install
Jupiter
from
your
cluster.
You
can
even
use
like
an
it
that
io
DNS,
based
on
your
ingress,
DNS
and
login
to
your
notebook,
and
you
can
login
with
it
lab.
A
A
B
A
And
right
now,
I'm
spending
a
bit
of
time
trying
to
find
candidates
for
our
service
T.
That
has
been
tough
and
show
with
some
help
from
owl.
We
sort
recruiting
team
and
I
now
have
some
tools
that
I
can
use
to
make
it
easier
for
me
to
find
candidates
and
then
basically
work
out
list,
and
so
that's
that's
I
would
say
up
there
with
my
top
three.
A
Additionally,
pants
is
also
I
in
my
mind
and
kind
of
the
strategy
that
we
will
use
towards
pass
and
what
even
past
will
mean
for
us.
So
we
know
that
we
want
to
use
resources
in
an
efficient
way
and
we
want
it
to
be
secure,
and
now
we
are
leaning
more
towards
separating
those
those
concerns
right.
So
we
want
to
give
security
by
default,
but
that
doesn't
mean
that
we
have
to
restrict
how
you
use
resources.
A
So
those
conversations
kind
of
have
evolved
to
having
that
that
separation
of
concerns
and
I
think
that
yeah,
we
may
be
focused
too
much
on
get
left
calm
at
the
beginning
and
I.
Think
that,
because
most
of
our
user
base
is
self
managed,
we
should
maybe
solve
the
use
case
for
the
75%
of
users
first
and
provide
them
with
a
nice
application
development
platform
or
pass.
And
then
maybe
we
move
on
to
get
lot
calm
and
providing
a
good
strategy
there
and.
A
A
Second,
one
is
around
customer
discovery
sessions
and
that's
basically
having
not
only
talking
about
a
feature
or
a
single
use
case,
but
more
of
a
deeper
conversation
with
our
customers
and
I.
Think
that
that's
that's
great,
because
that's
something
that
we
have
to
be
intentional
about
the
use
case-
and
you
know,
I-
have
a
problem.
Conversations
come
up
very
easily,
but
those
deeper
conversations
yet
that
do
take
some
some
more
work
and
then
the
last
one
is
just
growing
the
use
of
people
at
so
those
are
no
in
everything.
A
E
So
this
just
has
to
do
with
something
that
you
know
when
I
saw
you
bring
up
the
documentation.
I
saw
just
a
lot
of
Oh
auntie's,
have
a
thread
Intel
engineer
on
the
security
team
and
I
guess,
taking
it
back
step.
I
could
see
this
as
something
that
it.
Maybe
the
security
team,
in
addition
to
the
infrastructure
team,
would
also
be
like
a
natural
kind
of
user
of
this,
but
we,
you
know
there
is
that
that
setup
step
so
I
think
you
know
I
just
to
reiterate,
like
examples
of
getting
started.
E
E
How
easy
is
it
to
get
going
with
another
with
an
alternate
kernel?
Do
we
have
any
like
additional
support
for
that,
or
is
it
once
it's
running
like
you're
on
your
own,
managing
that
and
then
the
follow-on
is
was
what
do
we
see
customer
using
because
when
you
say
data
science
I
think
python?
Is
that
true
of
what's
being
used.
A
Yeah,
that's
a
great
question,
so
I
think
that
Jupiter
is
notably
like
a
data
tool
or
I
like
data
science
teams
and
data
teams
use
the
this
is
the
sort
of
tool
that
they
used
to
like
share
data
and
basically
share
approaches
to
data
and
things
like
that.
So
right
now
it's
when
you
say
Colonel,
do
you
mean
the
OS
that
we're
deploying
Jupiter.
E
A
Yeah,
so
the
the
only
one
that
we
have
right
now
is
in
did
the
Python
one,
but
anything
that
supporter
by
Jupiter.
You
could
install
on
your
Jupiter
instance
that
it's
deployed
to
get
laugh,
so
it
doesn't
have
any
limitation.
This
is
like
a
full-blown
to
prepare
notebook
and
Jupiter
helps
server
with
all
the
features
that
come
with
it.
So,
yes,
you
could
put
any
any
kernel
on
it
that
look
like.
A
E
I'll
I'll
post
a
link
to
the
security
room
book.
It
doesn't
come
up
in
Google
search
because
it's
it's
internal.
It's
it's
probably
a
good
thing
that
it
doesn't.