►
From YouTube: Ops 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
Everyone
is
my
honor
to
be
the
host
for
the
ops
section
group
conversation
today.
The
slides
are
in
the
dock.
I
see
some
questions
coming
in
I'll
just
cover
my
highlights
briefly.
So
I
kind
of
organize
them
as
just
general
notes,
accomplishments
and
future
plans
we're
starting
to
get
full
staffing
on
all
of
the
teams,
which
is
awesome.
The
system
group
was
kind
of
one
of
the
last
remaining
ones
so
excited
to
see
that
the
engineering
organizations
development
organization
has
started
to
organize
this
advisory
group,
which
I
think
is
really
interesting.
A
There's
a
lot
of
kind
of
complex
architectural
questions
in
the
ops
section
that
I'm
excited
that
that
group
is
kind
of
getting
together
and
at
present
it's
like
channel
I
think
they
had
their
first
meeting
today.
So
if
you
have
some
questions
about
that,
they
might
be
able
to
share,
but
I'm
excited
about
that
that
group
forming
and
then
we've
we've
consistently
said.
Dogfooding
was
really
important
in
the
ops
section
and
I've
seen
a
lot
of
great
engagement,
hat
tip
to
Devon
and
Tristan
Tristan.
A
Who
I
think
is
your
recently
done
or
is
about
to
do?
Some
like
shadowing
of
the
SRE
team?
Devon
is
or
so
Tristan
is
on
the
development
team
and
then
Devon
who's
been
spearheading.
The
dogfooding
efforts
of
using
auto
dev
ops
on
a
couple
of
small
projects
within
the
gait
lab
sra
and
infrastructure
team,
so
awesome
to
see
that
kind
of,
like
great
collaboration
between
the
two
teams,
accomplishments,
there's
kind
of
a
theme
to
both
accomplishments
and
future
plans.
A
B
A
And
I
will
ask
the
individual
product
managers
who,
on
a
call,
jump
in
I,
think
we
I'm
pulling
up
where
we
have
that
listed.
So
if
you
go
to
slide
13,
there
is
a
link
to
achieve
100%
category
maturity
and
that
link
I'll
put
here
has
the
current
planned
maturity
increments.
So
Sarah
is
not
available,
but
dovid
of
I
don't
know
if
you're
on.
If
you
would
care
to
share
the
planned
improvements
for
getting
logging
to
Bible
I,
don't
think
television,
probably
oh
okay,
so
Victor,
maybe
you
can
review
the
serverless
getting
to
viable
plan.
C
So
I
wouldn't
say
that
we
have
any
major
changes
by
now:
she's
pretty
good
news.
Basically,
we
are
currently
working
on
providing
automatic
domains
and
SSL
for
every
new
deployment
that
goes
to
ket,
and
besides
this,
we
had
this
things
to
add
which
really
just
enhances
developer
experience.
When
somebody
would
like
to
be
something
4k
native-
and
we
are
pretty
much
on
track
with
this.
A
D
Sure
so
we
previously
had
a
plan
to
use
this
open
source
project
called
cute
monkey,
and
that
was
kind
of
analogous
to
chaos,
monkey
which
is
made
popular
by
Netflix,
and
we
had
one
nut
problem,
but
one
limitation
there
that
it
only
allowed
us
to
do
one
kind
of
chaos,
experiment
which
was
killing
pods.
So
we
kind
of
hesitated
to
to
use
that,
and
very
recently
we
came
across
a
project
called
litmus.
Chaos
that
allows
us
to
do
could
have
a
wide
range
of
experiments
in
a
kubernetes
native
way.
D
So
we
feel
good
kind
of
moving
forward
with
that,
and
we
have
planned
to
do
kind
of
a
minimal
implementation
of
that
in
12:7,
which
will
allow
users
to
install
that
in
their
cluster
and
then
and
kind
of
run,
the
experiments
manually
and
then
long-term.
What
we
would
like
to
do
is
kind
of
fold
this
into
our
DevOps
and
then
allow
users
to
indicate
if
they
want
to
do
some
chaos,
engineering,
inner
project
with
a
variable
you
know
just
saying
yes,
opt
in
and
then
longer
term,
even
more
longer
term.
D
A
One
of
the
other
maturity
increments-
we
have
is
Error
tracking
and,
as
I
mentioned,
the
product
manager
for
the
health
team,
which
is
covering
or
tracking,
is
out
of
office
this
week,
but
I
can
speak
to
I've
I've
been
covering
for
some
of
the
release
post
process
this
week,
there's
a
number
of
features
related
to
error,
tracking
viability
going
in
this
release.
I
think
it
at
this
point.
A
You
know
we're
so
close
to
the
end
of
the
quarter
and
the
team
feels
like
they're
on
track
for
these
increments,
that
there
aren't
a
whole
bunch
of
unknowns.
I
will
share
in
the
spirit
of
your
question,
David
we'd,
originally
I,
think
at
the
start
of
this
quarter
had
tracing
getting
to
viable
in
this
release,
and
we
in
the
process
of
doing
some
validation
track
exercises.
Product
discovery
realized
that
there's
a
lot
more
effort
and
question
marks
around
our
embedding
of
jäger.
A
We
recently
kind
of
improved
our
embedding
of
the
Yaeger
UI,
but
the
api's
aren't
great
for
building
the
level
of
interaction
that
we're
building
and
error
tracking
for
tracing.
And
so
that
was
actually
a
case
where
we
pushed
out
at
maturity
increment
because
we're
trying
to
find
to
get
more
validation
of
the
problem.
B
Yeah
Amy.
Thank
you
for
the
insight,
the
kind
of
the
goal,
the
question
and
everybody
answer.
It
is
no
for
at
least
secure
and
defend
we're
struggling
with
the
what
you
build
yourself
versus
what
you
leverage
open
source
for,
and
you
know
we
as
we
be
endoscope
some
of
the
larger
projects.
We
start
realizing,
there's
gaps
and
other
parts
of
the
portfolio,
that's
causing
Astin,
to
work
with
somebody
else
to
run
something
out.
B
A
B
The
next
question
is
also
mind,
so
some
of
parts
of
the
monitor
Direction
align
nicely
with
other
sections
and
I'll
use.
Myself
is
the
example
here
you
know
defend
initially
want
to
have
a
lot
of
central
logging
for
security
events
and
then
I
outside
of
these
types
of
calls.
I
talked
to
Kenny
about
that
in
the
past,
about
how
defend
could
do
that.
The
question
I
have
is
for
the
monitor
team.
What
you
see
is
the
best
way
to
collaborate.
A
Think
of
things
like
when
we
spot
a
vulnerability,
insecure
or
similar,
and
then
we
have
the
new
a
similar
concept,
we're
thinking
about
in
terms
of
errors
or
alerts
in
the
monitor
world
that
maybe
we
can
think
of
a
smart
small,
primitive
that
would
be
useful
across
all
of
those
use.
Cases
is
something
that
there's
a
little
bit
ephemeral,
not
quite
as
robust
as
an
issue,
but
typically
it
comes
at
high
volume,
like
you
see,
with
vulnerabilities
and
alerts.
A
C
So
these
both
in
the
dogs,
clearly
lambda,
is
the
market
leader
here
this
huge
market.
She
actually
and
definitely
have
to
take
this
into
account
actually
for
this
actively
interviewing
land
users
in
order
to
understand
the
whole
surveillance
ecosystem
in
general.
What
what's
important
for
several
s?
How
can
we
creating
much
better?
Okay,
it
experience.
C
E
Would
you
might
turn
those
the
templates
and
stuff
we
just
would
be
all
spending
good?
We
can
also
make
sure
Amazon
those
at
P
awesome
but
and
long
term.
Do
we
feel
like
there
shifts
at
all
I
mean
I
know:
we've
made
a
big
bet
on
Keaney
it
for
obvious
reasons,
but
do
we
see
that
shifting
is
just
it?
We
don't
have
the
system's
insight
that
obviously
we
can
with
K
NATO,
so
I
get
them
being
quite
different.
C
Yes,
they
are
quite
different
by
nature.
Basically,
on
the
other
hand,
are
a
few
projects
right
now
in
the
open-source
community
which
try
to
create
runners
that
are
identical
to
a
tiberias,
lambda
runners.
In
this
way,
it
might
be
easier
for
someone
who
would
like
to
move
away
from
number
to
have
it
done,
but
actually
the
real
problematic
around
services.
That
is
not
just
with
service
functionary
that
you
provide,
but
the
whole
ecosystem
and.
A
I
think
one
way
that
Victor
and
I
have
talked
about
framing
the
distinction.
Is
you
know
in
our
configure
group,
we
talked
about
enabling
the
utilization
of
different
types
of
infrastructure,
typically
kubernetes,
and
that's
how
we've
fallen
to
rely
on
key
native,
but
server,
listen
specifically.
Service
lambda
is
more
what
we
would
call
a
an
application
type,
and
there
are
ways
that
we
can
improve
the
use
case
for
that
application
type
that
feel
unrelated
to
the
typical
work
that
the
configure
stage
does
but
is
still
important
to
driving.
A
You
know
the
adoption
of
gitlab
and
the
usefulness
of
gitlab
for
lambda
use
cases
so
brandon.
I
don't
know
if
that
term
resonates
with
you,
but
we
we
talk
about
application
types
as
things
like
web
apps
or
static
sites
or
other
things
and
services
with
lambda
at
least,
is
more
about
enabling
that
application
type
than
some
other
infrastructure
primitives
that
we
might
do
with
things
like
Candida,
where
we
are
installing
it
for
you
and
getting
it
set
up.
Yep.
E
Yeah
no
makes
perfect
sense
in
terms
of
how
we
think
of
it,
so
I
think
I,
think
framing
it
from
the
outside
will
be
really
useful
for
people
coming
at
it,
because
there
will
be
people
that
want
ubiquitous
lambda,
n,
AWS
and
lamda
capabilities
elsewhere.
There's
you
go
to
people
and
want
K
native
in
both
and
then
there's
a
very
large
number
that
literally
come
at
it
as
service
equals
lambda
and
in
some
ways
like
they
made
that
choice.
So
we
don't
really
need
to
confuse
them
and
just
affect
great
here's.
How
we
help
you.
A
F
So
I
just
recently
had
a
discussion
with
a
customer
who
is
we're
sort
of
complaining
a
bit
that
we,
we
don't
necessarily
have
any
official
best
practices
when
deploying
infrastructure
as
code
with
our
CI
CD
pipelines,
specifically
with
regards
to
lambda
and
and
they're
actually
using
Sam
to
deploy
to
lambda.
So
they
were
just
trying
to
figure
out
if
there's
anything
like
best
practice,
he
or
any
any
examples
that
we
have,
that
might
be
relevant
to
them.
C
Okay,
yeah,
nothing
to
that.
We
don't
have
best
practices.
It
was
just
so
before.
One
reason
for
this
is
that
lambdas
are
actually
an
application
type
and
that's
really
the
core
focus
opposed,
but
we
added
some
dogs
to
lambda
deployments
a
few
months
ago
by
Nicolas
big.
Actually,
he
even
wrote
this
already
and
I
think
he
right
now
linked
it
as
well.
So
this
is
just
very
busy
ICD
template
using
so
nothing
special
there
if
you're
asking
in
general,
but
is
this
code
not
just
London
and
some
that
bit
a
different
topic?
C
A
Simon
Oh
all
cheer
I
think
it
was
a
product
group
conversation
where
some
other
team
members
asked
about
like.
Is
there
an
equivalent
forget
lab
flow
for
infrastructure
as
code
and
I?
Think
part
of
Victor's
validation
effort
that
he's
described
is
understanding
if
there
is
a
common
pattern
and
how
we
better
support
it
in
whether
or
not
we
should
recommend
it.
I
think
the
industry
as
a
whole
is
asking
questions
about.
A
What's
the
appropriate
CI
workflow
with
infrastructure
areas
code
and
it
would
behoove
us
to
have
to
be
opinionated
there,
but
I
don't
think
we
have
the
information
to
suggested
like
definitive
way.
We've
I
think
you
talked
to
ten
different
customers.
You
hear
at
least
three
different
best
practices
for
managing
infrastructures,
grandma.
E
Victor
I
will
get
you
a
link,
but
the
wagga
talk
that
we
gave
it
reinvent
was
with
get
lab
and
terraform,
and
the
questions
at
the
end
are
literally
probably
ten
or
twelve
straight
questions
around
how
to
use
terraform
and
get
labs.
So
I'll
get
you
that
link.
So
you
can
just
listen.
The
video
but
you'll
get
a
bunch.
Interesting
use
cases
right
there
that.
G
Go
up
Dan,
Gordon
here:
okay,
thanks,
so
I
think
this
might
be
a
more
general
engineering
question.
But
just
following
off
of
the
first
question
where
we
said
we're
looking
at
a
technology,
it
didn't
have
what
we
wanted
in
it,
and
so
we
moved
on
we're
looking
for
another
technology
or
product.
How
do
we
make
that
determination
of
looking
for
moving
on
looking
for
the
new
tool
to
integrate
in
versus,
adding
to
it?
What
we
need?
Maybe
it's
API
level
or
whatever
and
then
and
then
contributing
that
back.
A
Examples
in
the
past
we've
we
look
at
size
and
makeup
of
the
community
right.
So
if
the
community
is
typically
solely
driven
by
a
singular,
singular
corporate
entity
that
doesn't
have
a
lot
of
direction,
that's
a
completely
different
story
than
something
that
you
know.
It's
like
a
CN
CF
project
that
has
a
bunch
of
corporately,
diverse
contributors,
but
Daniel
I,
don't
know
if
you
can
share
a
specific
example
from
that
evaluation
of
coop
monkey.
But
I
don't
know.
D
A
E
E
D
G
It
is
a
great
one.
Thank
you.
No
thanks
for
discussing
it.
So
the
next
one
is
is
around
just
kind
of
the
multi-cloud.
What
are
our
plans
going
forward
to
be
competitive
as
a
multi
cloud
provider,
because
that's
that's
one
of
the
things
that
we're
going
out
with
that's
unique
about
us:
we're
differentiated,
we're
not
tied
to
Amazon,
so
you
don't
use
code,
pipeline,
etc,
but
tools
like
github
actions,
measure
DevOps
and
a
lot
of
the
competition
have
plugins.
A
It's
really
another
interesting
question.
Thank
you.
Dan
I
would
maybe
start
with
I'm
willing
to
accept
challenges
on
this
one,
but
the
choice
of
a
tool
like
terraform
is
an
answer
to
that
right.
If
we
support
terraform
terraform
is
by
its
nature,
has
those
integrations
and
plugins
available,
so
more
native
support
for
terraform
in
our
C
I
would
put
enable
that
multi
cloud
without
us
having
to
build
those
primitives
specifically
for
each
cloud
provider
in
our
like
CI
mo
whatever
DSL
I.
G
The
question
I
would
ask
following
that,
then,
is
like:
how
do
we
differentiate
that
right?
If
multi-cloud
is
going
to
be
a
differentiator
for
us,
and
maybe
that's
the
thing
to
be
challenged,
then
then
doing
terraform
is
good,
but
we
have
to
do
something
more
to
differentiate
our
use
of
terraform
yeah.
A
And
I
think
you
know
one
of
the
things
we
think
about
in
Victor's
talk
about
this
earlier
when
we
think
about
infrastructure
as
code.
It's
not
just
like.
Do
we
support
it,
but
how
do
we
make
that
workflow
easier
for
the
practitioners
who
are
who
are
doing
that
who
are
using
a
restrictions
code,
and
that
involves
not
just
in
configure
stage,
features
that
make
the
workflow
easier
and
kind
of
more
empower
the
person
defining
that
infrastructure,
but
also
making
connections
between
other
stages?
D
I
would
add
to
that
that
integration
with
tools
like
cross
fling
is
something
that,
like
nobody
out,
there
is
doing
that
today
right
and
that
allows
you
to
kind
of
provision
resources
in
any
cloud
in
a
kubernetes
native
way.
That's
one
of
the
ways
that
we
differentiate
I
think
that
cluster
creation
is
another,
as
you
pointed
out,
and
we
have
plans
to
do
that
with
other
providers
in
the
future,
but
I
think
that
you
know
doing
things
that
are
cloud
specific.
D
This
is
slippery
slope
right
if
we
are
start
to
do
like
plug-ins
for
one
cloud
and
having
to
maintain
them
or
open-sourcing
to
the
community
and
I
think
that
that
kind
of
I
don't
know
it
grows
beyond
a
manageable,
valuable
and
maintainable
stage
very
quickly,
as
we've
seen
in
the
market
prior,
so
I
think
that
we
we
want
to
add
value
there
in
more
generic
ways.
I
I
wouldn't
say
that
we
want
to
do
anything
beyond
cluster
creation.
That
is
specific
to
a
single
cloud.
A
Okay,
Brian
thank
we're
out
of
time
which
I'm
loathed,
because
Lucas
I
would
love
to
answer
your
question.
I'm
gonna
copy
it
into
the
ops
section
channel
and
in
you,
and
we
can
continue
the
discussion
there.
It's
a
great
question.
I
guess!
Thank
you
all
reminder.
My
Corden
quote:
Easter
Egg
was
that
today
is
International,
Human
Rights
Day,
there's
some
links
for
how
you
can
celebrate
on
slide.
30
check
them
out
thanks.
Everybody
have
a
great
one.