►
From YouTube: Ops Group Conversation Public Livestream-November 1
Description
Ops Group Conversation Public Livestream-November 1
A
A
So
that's
awesome
start
seeing
those
more
close
to
fully
staffed
I'll,
say
personally,
that's
really
awesome.
For
me.
We
don't
have
any
product
managers
in
intramural,
including
myself
and
I,
know
on
the
engineering
side,
we're
starting
to
have
fully
staffed
teams
able
to
deliver
a
higher
rate.
As
a
result,
we
had
done
some
Direction
updates,
previous,
especially
the
ops
direction,
update
previous
to
the
last
group
conversation.
A
But
since
then,
we've
also
been
putting
some
effort
into
increasing
our
monitor
vision
and
increasing
the
ambition
there
and
then
I
also
recorded
a
kind
of
walkthrough
where
we
walked
through
and
had
a
Q&A
with
the
team
of
the
ops
section
vision.
So
that's
on
the
OP
section
page
now
accomplishments
the
incident
management
going
to
Bible
we've
have
a
lot
of
improvements
in
incident
management
since
it
went
minimal,
yeah
I
think
May.
A
A
As
a
user
and
my
test
projects
of
group
level
clusters
having
the
environments
tab
in
when
you're
looking
at
a
group
level,
cluster
is
really
great.
You
can
kind
of
immediately
get
a
sense
for,
like
the
number
of
pause
deployed
and
the
different
applications
that
are
using
that
cluster.
That's
really
cool.
We
added
Jager
to
the
UI.
This
has
been
a
thing
that
I
know:
Sid
was
particularly
pointed
about.
We'd
have
like
Tracey.
Nbc
was
basically
just
linking
out
with
a
URL
to
Jaeger.
A
We've
now
embedded
that
in
the
UI
and
I
think
that
will
give
us
some
improved
feedback
into
how
people
are
using
tracing
and
maybe
what
specific
components
they
would
want
to
see
in
directly
in
the
year,
I
can't
like
natively
in
the
UI
in
the
future,
I'm
really
excited
bout
source
control,
dashboards
and
this
intersection
between
our
routes.
As
a
you
know,
source
control
management
company
that
is
applying
many
of
those
same
principles
to
operations
and
observability.
A
Just
really
cool
things
on
the
road
map
that
I'm
excited
about
dogfooding
metrics
on
get
lab
is
something
that
I
know.
Dove
has
been
heavily
engaged
with
Marin
and
some
other
members
of
the
infrastructure
team,
moving
error,
tracking
devine,
so
we're
you
integrated
with
sentry
and
we're
bringing
some
more
native
you
eyes
to
the
user
experience
for
reviewing
and
digging
into
your
errors
directly
in
the
gate,
lab
application
to
things
related
to
AWS.
A
B
C
C
There
is
an
assumption
that
the
way
that
integrating
with
eks
the
way
that
we
would
integrate
with
uks
would
be
very
similar
to
gke,
and
you
know
it
turned
out
once
we
got
into
and
it
got
started
that
there
there
were
not
as
many
similarities
as
we
thought
and
that
ended
up
being
more
complex
and
more
work,
and
so
that
is
definitely
something
that
we
are
retrospective
and
we
want
to
make
sure
that
we
assess
out
these
issues
before
we
get
too
far.
You
know
into
implementation
or
really
before
we
start
implementation,
cool.
B
Thanks
to
that
slide,
28,
it's
cool
to
see
the
community
contributions
to
monetary.
Actually,
it's
happening
earlier
than
I
thought,
like
I
thought
we
would
have
to
like
dog
food
ourselves
before
anybody
else
would
use
it.
So
that's
great
news.
Any
ideas
like
who's
adopting.
What's
the
rate,
what
are
the
use
cases
anything
in
anything,
any
context
and
like
whose
are
people
actually
using
it
and
contributing
because
of
that
or
what's
what's
happening?
Why
are
we
getting
these
awesome
contributions.
A
Sara
I,
don't
know
if
you
want
to
add
some
color
feel
free
to
after
I.
Do
I
will
say
that
it
was
it
wasn't
until
I
saw
these
slides
and
I'm,
not
honestly
sure
which
engineering
team
member
compiled
the
slides
that
I
realized
that
they
were
I
didn't
ormally.
We
did
not
expect
to
see
these
contributions.
I
have
not
dug
in,
but
I
agree.
It's.
B
Forward
to
those
I
think
that
makes
sense
too
in
the
beginning,
with
everything
in
core
until
we
got
some
usage
and
then
the
executive
functionality
is
different.
My
fourth
question:
how
do
we
see
AWS
land
our
forces
and
exists
our
existing
server,
the
solution?
So
we
we
have
a
server
solution
you
can
deploy,
you
can
deploy
to
AWS
as
well.
Now
AWS
labs
are
super
popular,
but
how
do
we
think
about
supporting
both
and
possibly
supporting,
like
I?
Think
every
big
cloud
now
has
a
variant?
B
A
One
is
like
our
K
native
based
solution
is
enabling
the
quick
ability
to
get
started
with
service
if
you're
not
doing
anything,
but
we
were
hearing
from
people
that
were
saying
well
we're
already
starting
with
K
native
and
our
so
I
were
already
starting
with
service
on
lambda,
and
we
just
want
easy
primitives
in
our
deployment
and
management
of
those
functions,
typically
natively
to
get
lab.
I.
Think
you're
also
making
me
think
through
though,
as
we
add
eks.
A
Well,
you
don't
know
if
we
will
be
able
to
install
we've,
not
I,
don't
know
if
you're
able
to
install
K
native
on
an
u
KS
cluster.
So
that
would
also
be
a
question
of
maybe
that's
our
default
and
the
reason
why
we're
getting
a
lot
of
can
I
do
it
with
lambda
was
because
we
didn't
have
any
great
native
support
for
AWS
workloads.
Well,.
D
He
cares
said
we
also
can
he
recently
got
together
instead
of
creating
a
vision
for
the
configure
stage
I'm
and
that's
what
that
discussion
of
how
we
move
forward
with
support
for
other
Hans
is
also
ongoing.
Right
now,
there's
a
lot
of
contribution
from
Nicolas,
for
example,
and
the
other
team
members
and
product
to
come
up
with
a
really
good
vision
to
match
what
we
have
for
the
monitor
and
that's
certainly
been
a
topic
of
conversation.
So
you
can
probably
expect
to
see
more
on
this
coming
up
pretty
shortly,
cool.
B
Yeah
I
looked
at
that
epoch,
I
linked
one
727
and
it
didn't
talk
at
all
about.
Like
hey,
we
already
have
a
service
solution.
This
is
how
would
the
two
will
coexist
so
I
think
that's
a
big
question
to
solve.
We
should
listen
to
customers
but
yeah.
It's
obvious
to
me,
like
added
lambda,
has
a
lot
more
usage.
So
that's
what
people
ask
for,
but
we
should
be
very
careful
of
what
direction
we
advise.
People
take
yeah.
A
A
And
I
again
the
Dov's,
not
on
the
call
that
we
were
talking
about
having
em
doing
some
product
discovery
around
tracing
to
see
I
think
there's
some
questions
about
is
is
Jager
the
appropriate
path
forward.
Do
we
need
something
a
little
bit
more
native
to
get
lab
if
we
install
it
as
a
managed
app?
How
can
we
like
the
api's
for
Jager
are
not
great
so
it
in
century
we
can
kind
of
more
easily
build
the
primitives
in,
but
in
Jager
we
can't
so
I
know.
A
B
A
Yeah
I
would
say
it's
interesting
when
you
think
about
in
the
ops
section
in
general,
we
have
both
primitives
where
we've
started
with.
Oh,
you
can
install
a
managed,
app
and
then
added
the,
but
you
can
also
bring
your
own
Prometheus
as
an
example
of
that,
and
we've
also
had
the
reverse
and
sentry
was
like
well,
first,
you
got
to
bring
your
own
and
then
the
next
one
is
manage
that
but
I
don't
know
if
there's
weird
dove
and
I
were
discussing.
Is
there
a
standard
that
we
should
follow
in
that
process?
B
E
We're
gonna
write
this
down,
but
I'll
just
say
it:
I
have
a
bug.
Report
18
slide.
7
I
can
do
this
every
section
meeting
now
Auto
DevOps
is
márquez
viable
today,
but
you
see,
the
plan
was
cute
3
by
10.
22
is
gonna,
be
sorry
it's
minimal.
Today
it's
supposed
to
be
viable,
I
assume
it's
not,
which
is
fine,
but
it
just
means
you've
got
an
inconsistent.
Fated
categories
that
get
oh
cool
I
will.
F
A
E
A
E
A
A
E
A
F
Yeah,
that's
right
so
for
Auto
DevOps.
We
just
have
a
couple
of
dependencies
on
CI
that
have
not
shipped
yet,
but
we
have
a
big
one:
that's
gonna,
take
it
to
viable,
and
that
is
basically
selectively
being
able
to
run
jobs
in
pipelines,
and
so
we're
just
waiting
for
for
that
one
and
that
will
enable
us
to
run
Auto
DevOps
only
when
certain
requirements
are
met,
and
that
is
what's
gonna
take
us
to
to
viable
there.
So
we're
hoping
that
that
will
happen.
This
discord.
A
Mark
I'll
add
this
reference.
We
have
a
little
like
how
we're
prepping
for
the
GC
issue
and
I'll.
Add
this
to
the
retrospective
that
I'll
make
sure
to
have
a
kind
of
what's
the
Delta
between
when
we
last
presented
this
and
when
we
did
in
terms
of
maturity,
plans
they'd
be
great
I'll,
actually,
love.
A
G
Yeah
hi
I
just
wanted
to
say,
I'm
really
excited
for
this
feature.
I
think
we
can
probably
be
your
first
customer,
because
this
is
something
we
didn't
know
a
lot
with
incidents.
We
already
use
issues
for
incidents
and
we
often
find
ourselves
screenshotting
with
honor
panels
into
that
so
I
just
kind
of
wanted
to
give
you
a
heads
up
when
this
is
ready.
Of
course,
it's
going
to
fit
get
like
calm
before
the
official
release,
so
you
know
we
would
love
to
try
it
out.
Awesome.
E
Say
mark
sorry,
I
didn't
put
my
full
name
there.
Yeah
I've
asked
you
this
before,
but
in
case
things
have
changed.
How
are
we
doing
on
the
last
okay
couple
cabs
ago,
we
heard
loud
and
clear
that
folks
were
using
us
for
with
communities,
but
not
using
our
communities,
integration
for
a
variety
of
reasons.
How
are
we
doing
on
closing
those
gaps
so.
F
I
can
speak
to
them,
so
I
think
that
there
are
two
main
reasons
why
those
are
not
being
used,
and
the
first
one
is
that
our
integration
has
a
very
opinionated
way
of
having
namespace
creation
and
resources
and
users
accounts,
and
things
like
that.
So
people
want
more
control,
so
we
already
took
a
step
towards
that
and
basically,
we
give
an
ability
to
add
a
cluster
and
not
have
good
lab,
manage
it.
So
you're
going
to
be
able
to
create
all
of
the
resources
yourself
internally.
F
We
do
this
with
terraform,
so
this
allows
you
to
do
this
already,
so
that
was
kind
of
the
first
step
there.
The
second
step
there
is
that
for
people
that
do
want
to
take
advantage
of
the
integration
and
they
already
have
an
existing
practice
and
they
want
to
use
their
own
namespaces,
we
want
to
give
them
the
ability
to
map
their
environments
to
existing
namespaces
and
not
the
ones
get
lab
is
going
to
to
give
them.
F
That's
something
that
we're
working
on
12:5
and
then
finally,
I
think
that
the
big
thing
is
that
the
integration
currently
requires
cluster
admin
privilege
and
that's
something
that
a
lot
of
people
are
hesitant
about,
given
an
integration
externally.
So
what
we
want
to
do
is
we
want
to
have
that,
be
more
flexible
and
allow
users
to
shave.
They
want
this
to
be
a
cluster
admin
or
scope
to
a
single
namespace.
E
Because
if
I
remember
correctly,
one
of
the
big
things
was
that
they
had
like
teams
of
people
that
allocated
namespaces
you'd
request
the
namespace,
and
they
would
give
you
a
quoted.
Namespace
and
all
your
stuff
had
to
be
contained
that
within
within
that
namespace.
And
if
you
wanted
to
install
you,
know,
Prometheus
or
something
we
install
it
in
a
different
place.
It
doesn't
work
so
as
part
of
that
will
people
be
able
to
install
prometheus
in
their
own
namespace
and
have
multiple
prometheus
is
running
on
media
Evernote.
The
plural
yeah.
F
So
that's
what
we're
envisioning
is
that
that's
kind
of
the
that's
kind
of
the
compromise
is
that
you
can
have
this
more
tight
security
that
will
do
it
in
a
single
namespace,
and
that
means,
if
you
want
multiple
people,
to
make
use
of
something
like
like
Prometheus.
It
will
have
to
be
like
in
multiple
namespaces
yeah.
E
Well,
intermedius
normally
runs
really
well
that
way
anyway,
I
mean
in
fact
that's
the
best
practice
system.
You
could
everybody
should
have
their
own
Prometheus
and
then
maybe
aggregate
or
whatever,
but
that
totally
is
fine
for
Prometheus,
but
I
haven't
looked
at
all
the
rest
of
the
manage
that
to
see
whether
that
works
like
he
said,
serverless
doesn't
ork
a
native
specifically
doesn't,
but
but
as
long
as
a
lot
do
than
people
thinking
a
lot
done,
yeah.
F
And
then
the
the
final
thing
that
you
can
brought
up
is
the
self-service
thing.
So
we
have
a
lot
of
big
large
customers
that
when
were
doing
self-service
portals
that
are
homemade.
So
let's
say
they
they
log
in
and
they
say:
hey
and
I
need
a
namespace
with
his
resources
and
then
they
get
back.
You
know
credentials.
So
this
is
something
that
they
can
do
in
get
lab
right.
If
an
operator
provides
a
cluster
at
the
instance
level,
then
teams
kind
of
are
implicitly
doing
that
today,
but
it's
very
homogeneous.
So
they
want
something.
F
E
I'm
glad
you're
gonna
say
it
was
actually
gonna,
be
a
follow-on
question
because
it
seemed
to
me
like
everybody's,
doing
this
homegrown
thing
and
like
why?
Don't
we
just
provide
them
a
nice
way
that
they
can
get
a
quoted,
isolated,
namespace
with
all
their
policies,
anything
else,
and
maybe
they
still
have
to
manually,
approve
it
or
maybe
not
whatever
it
is,
but
why
don't
we
just
make
that
all
easier
for
everybody,
yeah.
A
F
Yeah
it's
funny,
but
one
of
our
customers
built
a
Jupiter
based
app
that
allows
users
to
self-service
in
kubernetes,
saying
no
capacity.
You
know
namespace
what
it
looks
like
and
things
like
that.
So
yeah
we've
seen
homegrown
things
all
over
the
place
and
that
clearly
indicates
an
opportunity
of
something
that
we
could
do
well,
since
we
have
all
of
these
knowledge
and
all
of
this
context
of
their
apps
awesome.