►
From YouTube: Chaos Engineering WG Meeting - 2018-08-28
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
A
I
think
what
I
wanted
to
discuss
quickly,
briefly,
with
with
your
focuses
your
thoughts
on
how
to
move
things
for
what
we
is.
You
know
whatever
scope
we're
trying
to
achieve
with
the
working
group
proposal,
notably
the
the
wipe
up,
is
something
that
I'm
interested
in.
But
beyond
that
so
to
you
know,
it
gives
sense
to
the
chaos
engineering
effort,
we're
trying
to
convey
I
guess
it's.
You
know
it's!
It's
quite
crystal
clear
that
it's
been
a
bit
quiet,
but
it's
also
a
very
new
subject,
except
for
those
people.
A
Who've
been
doing
that
like,
like
you,
folks,
well
Michael
anywhere,
but
at
some
time
I
think
it's
exciting,
because
it
means
we
can
actually
engage
in
setting
up
the
ground,
for
you
know
for
the
rest
of
the
industry
when
they
catch
up,
but
I
was
wondering
you
know
well,
we've
lost,
it
will
probably
be
back
yeah.
So
I
just
wants
one
during
your
thoughts
on
that,
and
you
know
where,
where
your
man's
at
on
that
subject,.
B
B
We
put
some
effort
in
the
next
couple
of
weeks
to
get
it
into
a
more
finished
state
and
then
CN
CF
can
tweet
out
and
publicize
a
little
bit
more
widely
that
we're
trying
to
sort
of
finalize
this.
Please
feel
free
to
come
to
this
meeting
and
then
we'll
sort
of
put
it
into
I
can't
remember
what
they
call
it
in
the
IETF,
but
like
a
it's
like
free
up,
close
closing
coal
or
something
yeah.
C
A
Think
that's
fair
I
think
we
need
to
that
little
push
for
the
next.
You
know
two
or
three
weeks
whatever
it
takes,
but
simply
a
very
short
deadline
for
for
bad
people
to
make
their
mind
whether
or
not
they
are
ready
to
contribute.
If
there's
anything
to
contribute,
I
wouldn't
want
the
white
paper
to
reflect
only
just
either
my
mind
or
you
know
a
few
other
people.
A
It
would
be
wrong
of.
You
know
on
that
note.
At
the
same
time,
I'm
wondering
and
all
the
commits
are
met,
I'm
really
wondering
if
I'm
actually
on
the
right
track
anyway,
so
I'm
going
to
I'm
quite
looking
forward
to
commands
feedback
I
honestly
on
my
promise,
I'll,
but
wonder
if
we've
got
the
right
notes
because
we
don't
want
to
be
normative.
A
It's
it's
very
tricky
to
be
to
be
fair
because
we
don't
we
don't
want
to
look
like
this
is
the
only
where-
and
you
know,
that's
the
standard
way
of
doing
kills
enjoying
at
some
time
come
II
to
high
level
and
just
be
basically
useless
as
a
white
paper.
So
I've
found
I'm,
not
you
know
worsen
in
writing
white
paper.
So
certainly
I
would
welcome
feedback
from
the
community
on
that
front.
D
Hey
this
is
Deepak
here
from
Capital
One,
so
I
know
like
I've,
been
trying
to
contribute.
I
I,
sent
a
message
to
Chris
and
then
got
the
details,
so
I
will
make
time
this
week.
Hopefully
you
know
before
this
September
long
weekend
here
I
will
add
my
contributions
there
and
then
probably
we
can
do
a
review
in
the
next
week.
Thank.
A
A
D
Definitely
so
you
know
at
Capital
One
we
do
practice,
yes,
engineering
seriously
like
from
the
identity,
so
various
departments
we
do
at
different
levels,
but
we
try
to
do
all
internal
presentations
about
resiliency
and
right
now,
I'm
working
with
my
communications
group
to
see
you
know
like
what
can
be
shared
so
I'll
also
come
forth
with
some
presentations
going
forward.
D
D
A
Was
thinking
aside
from
that
I
think
I
wonder
if
session,
if
there
is
a
value
in
trying
to
reach
out
either
through
examples?
Maybe
with
that
we've
done
some
examples
on
online
on
this
on
this
meeting
and
they've
been
recorded.
So
that's
great
but
I
wonder
how
far
we
could
go
is
Anna
I,
don't
know
what
would
be
the
shape,
but
some
more
success
stories
like
if
you
you,
you
know
you
for
self
I've
told
but
maybe
written,
because
I
think
the
white
paper
could
could
welcome
alongside
it
some
sub.
Yes,
some
some
success
stories.
A
A
I
think
we
we
do
need
that
kind
of
thing
like
you
know,
the
Grameen
folks
have
done
quite
good
job
of
summarizing,
kill,
Jinping
and
telling
the
story,
and-
and
we
need
more
things
like
that-
we've
tried
us
where
we
skills
IQ,
but
maybe
maybe,
as
we
could
suggest
and
I
will
suggest
you
to
the
slack
channel
whether
or
not
we
could.
You
know
create
like
a
repositories
of
a
page
somewhere
saying
you
know
just
listing
those
those
those
those
articles.
That
would
be
easy
for
people
to
to
look
through
and
one
thing
I
want.
A
Well
one
last
thing
for
me
anyway,
for
me:
I,
wonder:
either
value
in
reaching
out
to
other
working
groups
and
perhaps
having
a
discussion
with
them
like
I'm
thinking
serverless,
because
it's
hard-
and
you
know-
and
it's
fitting
in
my
opinion,
but
could
be
as
anything
else
like
storage,
obviously
or
otherwise
saying
well.
Could
we
could
we
try
to
come
out,
come
up
with
a
story
of
chaos
enduring
in
service
or
chaos?
Engineering?
A
B
A
A
I
think
it's
just
that
it
to
me
anyway.
This
is
a
personal
opinion,
obviously
to
make
your
centering
is
more
federating
kind
of
practice
and
practice
practice
on
its
own.
That
leaves
since
in
its
own
world,
so
I
lack
I
like
the
idea
to
say
what
chaos
I'm
doing
is
not
just
yeah
a
silo
that
you
practice.
You
know,
aside
from
anything
else,
it's
really
more
federating
on
that,
but
and
I'd
like
to
demonstrate
that.
So
if
the
group
feel
like
it's,
it
goes
beyond
the
the
need
of
this
working
group.
A
A
And
back
in
the
days
it
was
mostly,
you
know,
on
email,
so
you
needed
patience
to
read
long
emails
and
just
make
sense
of
what
it
is
so
I'm
used
to
it
right.
I
don't
want
to
keep
the
the
meeting
shot.
You
know
much
more
longer,
I,
don't
know,
maybe
just
just
quickly
I.
Have
you
folks
being
working
on
something
interesting
on
kills
injuring
inside
your
company
is
always
where
or
whatever
just
out
of
curiosity
for
myself,
I.
B
E
C
C
D
We
have
had
you
know
few
game
days
done
and
we
actually
have
recurring
chaos
test
running
in
our
production.
We
started
off
with
simple
scenarios
like
you
know,
ec2
stops
and
then
we've
had
a
bunch
of
learnings
with
our
tunings
and
making
sure
our
play
books
are
exercised
well
and
all
that
so
yeah
we've
been
on
the
way
and
we
use
our
internal
product,
and
you
know
my
team
has
been
contributing
to
some
new
scenarios
as
well.
You
know
creating
Network
disruptions
and
things
like
that.
D
A
Sweet
sweet
well
on
my
on
all
sides:
we've
been
so
I
work
on
be
killed,
two
kids,
if
you
were
above
it
or
Sookie
tokens
project
and
we've
been
busy
on
to
France.
First
of
all,
we
have
quite
a
few
contributions
now
on
some
drivers
like
on
AWS,
most
specifically
in
communities.
So
we've
been
busy.
You
know
just
you
know,
tribe,
trying
and
and
handling
all
the
peels
and
stuff
like
that.
A
So
it's
been
quite
awesome
to
see
people
the
other
body
has
driver,
for
instance,
is
mostly
community
driven
it's
just
because
I
don't
use
it
the
videos,
my
sense
is
there's
just
I,
wouldn't
I,
wouldn't
presume
what
to
test
or
what
to
fail
or
whatever.
So
it's
so
awesome
to
see
the
community
come
in
with
ideas
about
what's
the
right
thing
to
actually
go
and
look
for
and
and
and
try
to
break
or
you
know
or
just
push
over.
A
So
it's
been
quite
awesome
on
that
front
and
on
the
other
front,
I've
spent
quite
a
lot
of
my
time
for
the
past
two
or
three
months
working
on
the
kill
serve,
which
is
project
we're
going
to
put
out
in
September.
It's
like
the
DUI
on
top
of
chaos
toolkit.
So
it's
an
open
source
project
where
it's
a
fun.
You
know
web
front-end
for
funneling
your
you
know,
collaboration
on
top
of
chaos
toolkit.
So
you
can.
A
You
know,
create
organization
and
stuff
like
that
and
and
just
deal
with
all
your
efforts
of
chaos
and
rank
through
the
chaos
it
get
that
way.
So
it's
really
a
collaborative
portal.
On
top
of
the
girls
to
kids,
we
requite
it's,
you
know,
excited
about.
You
know
putting
it
out
there.
It's
hopefully
we'll
pick
it
up
and
I'm
sure
I'd
be
happy.
If
you
have
some
feedback
on
it
at
some
point
would
be
brilliant
and
that's
about
it
right.
A
That's
good!
That's
good!
To
hear
we've
been
busy
all
right!
I!
Really!
Don't
want
you!
You
know,
I
know
it's
probably
early
or
whatever
it's
time
it
is
so
I'd
rather
like
leave.
You
I'm
going
to
enjoy
a
confident.
You
know,
Bo
you
with
more
details.
I,
don't
know
if
you
have
anything
else
to
say
to
the
hey.
D
I
just
had
one
question:
you
were
talking
about
the
Kaos
toolkit
right.
Yes,
so
listen.
What
is
the
vision
for
that?
Is
that
going
to
be,
like
you
know,
like
part
of
the
landscape?
Where
are
we
expecting
everyone
like
who's,
just
starting
to
you
know,
start
with
the
chaos
toolkit
and
contributor?
What
is
the
vision
of
that?
So.
A
A
If
you
know
hopefully
after
the
white
paper
is
done,
the
group
will
remain
and
continue
and
and
figure
out
if
there
is
a
project
that
can
be
set
as
as
a
as
a
proper
sincere
project
for
Cassandra
I
have
no
idea,
we
would
have
called
you
know
to
push
it
forward,
push
it,
but
in
a
negative
way,
but
something
we
are
proud
enough
and
we
think
it
deserves.
You
know
a
discussion
on
that.
So
coastal
kids
is
just
an
open
source
project.
People
can
start
there
or
start
anywhere
else.
A
Tiffany
I
see
okay,
make
sense
right
actually
quickly.
Michael
I,
don't
know
if
I've
asked
or
if
we've
asked
that
question
before,
and
and
so
it
for
that.
But
I
can't
recall
exactly
what
status
of
the
various
tools
you've
created
in
you
know
internally
in
terms
of
being
used
as
well,
they
seem
to
be
very
specific
to
LinkedIn.
You
know
basically
way
of
working.
So
much.
B
A
B
The
application
level
stuff
I,
don't
really
see
being
open
sourced
in
saying
that
some
of
the
components
in
the
transport
transport
framework,
which
is
called
wrestle
II
those
changes
that
were
made
for
the
chaos
engineering
purposes.
They
are
open
source,
mm-hmm,
I,
know
Tami
tweeted
them
out
a
few
months
ago.
B
So
some
of
that's
there
but
I'm
not
sure
how
much
of
the
magic
is
there
that
you
actually
need
so
I,
don't
think
the
whole
thing
is
going
to
be
open
sourced,
but
maybe
there
are
the
host
level
stuff
and
there's
not
like
we're
doing
anything.
Super
fancy
like
if
I
had
some
time
I
could
probably
go
and
contribute
my
own
modules
on
the
on
salt
but
yeah
that.
A
E
It
kind
of
kicked
my
tires
and
I
just
wanted
to
wrap
up.
We
have
been
working
on
powerful
Co,
making
it
a
little
bit
easier
to
use.
We
actually
got
a
very
gifted
intern
who
ended
up
working
on
that.
So
there's
a
few
features
that
are
currently
in
review
on
the
repo,
and
there
are
all
many
around
actually
making
it
easier
for
the
learning
curve
to
stars
to
have
something
set
up
quickly
on
your
kubernetes
cluster,
so
that
you
can
start
saying
what
happens
in
a
kind
of
semi-automatic
mode.
E
So
there
is
this
semi-automatic
mode
that
read
some
metrics
like
CPU
ROM
and
tries
to
kind
of
guess
what
might
be
interesting
to
kill
and
as
kills
it
for
you,
there
is
a
web
interface
that
makes
it
easier
for
you
to
actually
see
what's
going
on
there.
E
If
you
don't
want
to
stare
at
logs
in
the
command
line-
and
there
are
some
other
other
snow
filters
that
also
make
it
easier
to
kill
the
different
pods
I
and
an
interactive
mount
that
you
can
deploy
that
if
you
don't
have
like
command
line
access
to
your
cluster
or
you
don't
want
to
give
to
someone
else-
and
there
is
also
UI
kind
of
interactive
mode.
That
shows
what
you
can
kill
and
you
can
kind
of
click
and
select
and
filter
and
see
what
happens
there.
So
yeah.
C
E
B
E
B
A
That's
interesting
because
it
seems
to
me
that
most
a
lot
of
companies
are
building
their
own.
You
know
tools
to
do
that.
Now
there
are
various.
You
know
ways
you
know
reasons
probably
because
I
started
quite
a
while
ago,
and
there
was
nothing
else
or
maybe
because
I
have
specific
needs
at
no
other
tools
can
provide,
but
I
I
wonder
how
we're
going
to
settle
that,
not
honestly
I'm
not
talking
from
a
business
perspective.
You
know
even
from
a
standard
way
of
doing
things,
I
wonder
if
you're
centering
can
actually
standardize.
E
A
E
Also,
the
scene
safe
approach,
I
think,
is
quite
nice
in
the
fact
that
they
don't
try
to
choose
one
technology
or
one
particular
over
others.
They
kind
of
serve
the
different
ones
that
are
potentially
competing
and
serve
slightly
different
use
cases.
So
I
think
that
you
know
realistically
we'll
just
have
to
kind
of
see
how
this
all
evolves
and
where
it
goes
and
probably
merge
some
things
later
on.
Yes,.
A
A
Yeah
I
will
so
see
the
fact
that
people
are
there.
Are
you
have
people
like
you,
you
folks
doing
doing
it
already,
and
then
there
are
many
people
who
are
not
even
that
stage
of
resiliency
in
their
mind
or
in
their
designs.
There's
such
a
stretch
all
right,
it's
exciting
times
all
right.
Well,
if
that's
all
from
any,
you
know
any
of
you
I'm
happy
to
just
stop
the
call
and
get
it
shot,
ready
all
right.
Go
fight,
fill
up
all
right
thanks.