►
From YouTube: End User: Salesforce - Aaron Erickson
Description
End User: Salesforce - Aaron Erickson
Join us for KubeCon + CloudNativeCon in Shanghai June 24 - 26 and San Diego November 18 - 21! Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy and all of the other CNCF-hosted projects.
A
A
One
almost
say,
thanks
to
everybody,
who's
helped
start
this
foundation
for
taking
the
time
to
participate
in
the
summit,
and
for
those
of
you
that
are
writing
or
contributing
to
the
software
for
this
foundation
and
the
other
open
source
software
that
so
many
of
us
use.
It's
it's
really
awesome
and
we
all
really
appreciate
everything
everything
you're
truly
doing
here.
A
I
thought
about
these
a
little
bit
useful
just
to
use
one
of
our
it's
the
only
this
is
not
about
a
product,
but
a
little
bit
about
a
little
bit
about
what
Salesforce
is
about
and
some
of
the
scale
it
operates
at
just
kind
of
looking
at
the
kind
of
the
things
that
happen
billions
of
times
a
day
when
when
people
are
using
Salesforce
to
do
their
to
get
their
work
done,
it's
a
truly
an
epically,
large
kind
of
scale
system
that
is
very
diverse.
Across
users.
A
Workflows
customer
calls
interactions
it's
north
of
a
trillion
actual
actual
transactions
a
year.
So
there's
just
a
tremendous
amount
of
complexity
that
goes
on
as
we're
delivering
the
services
that
our
customers
want
and
there's
a
lot
of
an
underlying
tech
that
we
then
integrate
with
and
part
of
these
are
someplace
over
here
there's
a
platform
we
actually
four
star
comes
a
big
platform
that
people
build
on
top
of
us,
so
you
know
way
before
I
joined
the
company.
I
didn't
know
all
the
stuff
that
we
did.
A
So
maybe
just
a
couple
of
interesting,
possibly
interesting
facts
about
our
transformation
journey
as
a
one
of
the
large
corporations
we
happen
to
have
we
we've
grown
a
lot
of
systems
internally.
We've
also
created
a
lot
of
systems
and
through
acquisitions,
so
that
North
of
about
twenty
CI
systems,
CIC
D
systems
we're
in
the
process
of
bringing
those
bringing
those
more
teams
onto
our
larger
systems.
A
So
as
we're
working
through
the
CI
systems,
the
the
fleets
of
machines
are
kind
of
working
through
that
are
doing
all
of
this
to
try
to
help
our
developers
get
rapid
feedback
and
accelerate
the
understanding
and
learning
that
they're
trying
to
do
to
be
able
to
make
things
happen
and
we're
very
big
users
of
Jenkins
and
thanks
to
everybody
who
who's
done.
All
the
work
there,
it's
one
of
our
largest
kind
of
integration
systems
and
all
the
plugins
and
everything
around
that
is,
is
very
motivating
and
is
very
helpful.
A
One
of
the
things
that
we
noticed
when
we
think
about
the
the
CI
CD
space
is
one
how
fragment
it
is
and
that
fragmentation
means
there's
a
ton
of
integration
that
has
to
happen
and
if
I
look
at
the
the
teams
that
are
working
on
our
organization.
A
lot
of
the
work
that
we
do
is
that
integration
and
you
know
a
company
on
our
scale.
We
can.
A
We
can
do
that,
but
when
you
think
about
smaller,
smaller
scale
organizations
that
integration
is
a
real
burden
and
one
of
the
things
I'm
hoping
that
we
can
do
as
part
of
this
foundation
is
try
to
find
ways
to
make
that
easier,
more
cost-effective,
faster.
To
be
able
to
do
that.
So
people
can
make
use
of
some
of
these
wonderful
tools,
plugins
and
different
technologies
that
we
have
so
maybe
I
could
spend
a
few
minutes
and
motivate
some
of
the
integrations
we
have
because
you're
like
okay
I've
got
a
CD
pipeline.
A
What
am
I
integrating
in?
Well,
you
might
be
integrating
in
some
more
sophisticated
security
scanning
that
understands
the
provenance
of
the
code.
Where'd
it
come
from.
Do
I
know
it
didn't
come
from.
Someplace
else.
Do
I
know
it's
our
code.
Security
scanning
has
another
one
that
like
are
there
known.
Vulnerabilities
in
this.
A
Are
there
are
there
techniques
that
the
developer
is
used
that
meet
a
pattern
which
we
wouldn't
want
to
have
happen
in
in
production
code
and,
of
course,
there's
metrics
and
there's
analysis
and
and
different
items
which
we're
all
trying
to
do
work
and
analysis
around
the
source
code
as
its
evolving
and
changing.
So
these
are
all
things
that,
over
the
years
we've
integrated
in
because
the
teams
themselves
are
on
their
own
mission
and
try
to
understand
how
they
can
move
move
forward
and
get
and
go
faster
and
that's
kind
of
the
integration
at
custom
code.
A
There's
a
non-trivial
amount
of
that
in
our
world.
So
we
feel
like
we're
that
connective
tissue
between
dozens
of
these
other
systems,
including
some
of
the
custom
custom,
auto
build
type
systems
that
we've
we've
had
before
so
systems
that
predated
some.
Some
of
the
the
Jenkins
world,
so
some
of
these
are
some
of
our
main
systems
contributing
to
selenium,
and
then
one
of
the
things
that
were
we're
focusing
on
is
given
that
so
that
suite
of
tasks,
how
do
we
prioritize?
How
do
we
optimize
our
our
test
runs
against
anyone
change
into
a
system?
A
It's
the
last
thing
you'd
want
to
do
is
run
kind
of
somewhere
some
version
of
the
entire
suite
against
that.
So
how
do
you
then
choose
in
a
machine
learning
or
some
kind
of
pattern,
recognition
way
which
subsystem
the
person
is
working
on?
How
do
you
motivate
that,
so
you
can
get
faster
feedback,
not
only
cost,
but
just
feedback
is
a
real
big
deal.
A
We
call
the
software
supply
chain,
a
software
system
that
allows
us
to
quickly
reliably
and
repeatedly
build
and
manage
our
artifacts,
the
creation
of
them
and
the
deployment
onto
our
systems.
Security
is
a
huge
value
in
any
any
organization
is
very
big
value
for
us,
so
we
really
want
to
make
sure
that
we've
got
its
far
to
the
possible
a
secure,
robust
system,
something
that
works
across
the
entire
suite
of
application
architectures.
So
not
all
of
our
things
are
fully
factored.
You
know,
Heroku
12
factor,
apps.
A
Obviously,
we've
things
that
range
from
micro
services
through
to
larger
scale,
subsystems
to
some
things,
you'd
call
mod.
You
can
model
this
type
of
type
applications
and
all
of
those
application
teams
are
on
their
own
journey
and
we
need
to
support
the
different
kinds
of
applications
of
the
HEV
running
and
we
care
about
being
very
substrate
independent
for
us,
a
substrate
being
one
of
our
first
party
data
centers.
A
One
of
our
cloud
provider
infrastructure
providers
that
we
use,
so
we
want
to
make
sure
that
our
suite
of
facilities
are
running
effectively
in
an
abstracted
way,
so
we
can
make
best
use
of
those
resources
so,
as
we
think
about
kind
of
larger
scale
and
kind
of
ambitions
for
our
CIC,
this
system
CIC
these
systems.
These
are
the
kinds
of
things
we
think
about
its
ways
to
make
progress
on
that
this
year,
so
kind
of
places
to
work
together
to
make
improvements.
A
It
is
something
that's
interesting
and
being
very
noteworthy
using
the
said
earlier.
The
mission
kind
of
machine
learning
to
do
test-
optimization
is
something
that's
think,
is
a
place
where
we're
bearing
a
lot
of
fruit.
Right
now
and
upgrading
all
of
our
CDs
see
ICD
pipelines
to
leverage
the
open
source
projects.
A
So
I
was
really
kind
of
inspired
by
this
aspect
of
the
of
the
mission,
which
is
to
really
enable
the
software
development
teams
to
deliver
code
better
and
faster.
That's
when
I
think
about
the
hope
for
them,
for
the
foundation
is
that
we
can
build
an
organization
a
community.
Here,
that's
got
tools,
technology,
it
has
processes,
it
has
techniques
and
has
people,
because
many
of
these
things
are
well
beyond
what
a
tool
can
do.
As
we
look
through
our
developer
productivity
teams,
we
see
that
the
tools
are
really
we're.
A
A
I
think
that's
it
and,
of
course,
I'm
gonna
accept
down
for
a
second
one
of
the
things
that
what
do
we
care
about
from
a
software
perspective,
secure
and
trustworthy?
If
we
can't
trust
the
bits
that
we've
got
out
there,
then
basically
everything
else
stops.
So
we
need
to
make
sure
we
understand
where
what
is
the
provenance
of
our
code?
Where
is
it
going?
What
machines
is
it
going
on
to
how
do
I
control
that?
How
do
I
manage
that
ineffective
way?
A
A
What
you
end
up
with
is
finding
ways
to
do
that:
automation
in
a
repeatable,
structured
manner
and
then
I
think
one
of
the
things
that
kind
of
we've
kind
of
recently
gotten
a
handle
on
understanding
is
the
the
need
of
transparency
and
visibility
observability
in
your
CI
system
at
scale.
So
we
often
think
about
the
developer
centric
view,
but,
as
you
think,
about
larger
and
larger
teams
of
devs,
what
you've
got
in
is:
how
is
this
code
transforming?
Where
are
there
things
stuck
in
pipelines?
A
So
one
of
things
that
Sai
excites
us
about
as
its
kind
of
an
end
user
in
this
community,
because
we
don't
sell
our
underlying
developer
kind
of
software.
That's
we're
really
here
because
we
want
to
be
part
of
the
community
here
that
can
move
the
practice
forward
and
and
move
that
practice
forward
in
software
software
integration
and
the
practices
that
we
we
develop.
We
what
we
hope
to
be
a
member
with
you
all
to
help
build
forums
of
experience,
so
people
can
share
what
they're
doing
and
and
find
ways
to
improve
the
practices
around.
This.
A
Oops
sorry,
one
last
thing
so
I
wanted
to
announce
something.
I
was
catching
up
with
with
Michael
Windsor
earlier
today
from
Google,
and
we
were
talking
about
the
people
who
lead
or
the
developer
productivity
teams,
and
so
we'd
like
to
announce
a
meeting
of
folks
in
those
organizations
in
New
York
and
if
you're
interested
and
if
that's
a
role
that
you
have
please
come
talk
to
me
or
Michael.
Please
stand
up
Michael
and
we
want
to
talk
about
organizing
that
we
think
it'd
be
really
good
to
get
a
community
like
that
together.
A
So
we
can
find
out
more
about
what
each
other
does
our,
how
we
lead
things,
what
what
tools
we
use,
but
really,
how
do
we
manage
and
grow
that
practice
within
developer
productivity?
So
with
that
I'd
like
to
say,
say
it
say.
Thank
you.
I'd
appreciate
the
time
I'm
looking
forward
to
really
contributing
more
in
the
community.
We
have
some
time
for
questions,
I.
Think.
A
A
What's
the
goal,
yeah
so
to
repeat
the
question:
we
have
more
than
20
CI
systems
and
what's
the
goal,
the
goals
are
bringing
that
number
down
note
that
some
of
these
came
through
acquisitions.
Some
of
these
came
because
the
product
was
was
very
different
than
what
we
could
currently
have.
We're
now
increasingly
happy
with
our
main
systems,
and
we've
been
over
the
last
year
on
boarding
more
and
more
to
the
main
system,
and
we
we
get
numbers
gone
down.
It
will
continue
to
go
down.
A
What
the
question
is,
what
is
the
main
system,
and
and
what
do
we
call
that
so
the
main
system,
summa
kalla
SF,
CD,
SF,
CI,
the
Salesforce,
continuous
integration?
It
is
based
upon
a
Jenkins
underlying
engine
with
a
lot
of
customization,
plugins
and
automation,
and
things
around
that
that's
kind
of
at
a
high
level.
A
So
we
have,
we
have
20
system,
just
wait
the
question
about
20
systems
and
why?
What
do
they
need
and
is?
Is
it
really
are
they
different
enough
there
that
we'd
have
a
goal
to
have
get
it
down
to
one?
It's
not
a
goal
to
get
it
down
to
one:
it's
a
goal:
to
get
it
down
to
a
reasonable,
covering
set
where
they're
able
to
get
their
needs
met
in
a
in
a
pretty
optimal
way,
and
so
we're
really
not
trying
to
force
fit
people
into
things.
A
We
have
a
great
belief
in
and
kind
of,
freedom
and
choice
and
optimization
each
one
of
these
is
a
significant
business
that
it's
supporting.
We
want
to
make
sure
we
support
that.
The
growth
in
health
of
that
business,
and
so
part
of
what
we've
been
doing
is
understanding
what
those
needs
are
and
figuring
out
where
we
can
start
to
accommodate
them
and
in
that
kind
of
way,
it's
very
important
for
us
that
we
don't
declare
a
lot
of
things
to
people
and
say
you
must
use
this.
A
A
We
are
not
entirely
kubernetes
I
wish
we
were
most
likely,
but
no,
we
have
a
complex
set
of
workloads
that
range
from
kind
of
all
the
substrates
you'd
expect
from
bare
metal
to
the
m2
containers.
So
in
that
space
we've
got
to
figure
out.
How
do
we
manage
those
artifacts?
How
do
we
create
and
deploy
those
against
a
wide
set
of
data
centers,
and
some
of
that
software
is
very
new
kind
of
service-oriented.
Some
of
that
software
is
more
database
oriented.
A
A
A
A
B
A
So
how
do
we
have
we
built
community
within
kind
of
the
area
about
productivity?
We've
got
some
a
lot
of
internal
boards,
an
internal
messaging
that
we
have
a
few
of
the
teams
still
have
groups
that
are
kind
of
focused
around
their
specific
needs,
so
we
team
up
significantly
with
them
around
partnering
and
occasionally
we
hire
occasional
people
go.
Oh
you've
got
a
would
like
to
join
your
team,
and
so
that
happens
non-trivial,
and
so
we
try
to
make
sure
we
support
all
those
teams
with
the
best
applications
in
tech.
We
can.
A
Okay,
the
question
was
about
that,
come
with
a
comparison
between
Jenkins,
X
and
spinnaker,
and
can
do
we
see
those
both
being
able
to
work
well
for
us
right
now.
We
are
just
using
Jenkins
and
we're
using
it
in
a
CI
manner.
We
have
not
yet
gotten
into
Jenkins
X
I.
Think
we've
had
it
in
the
lab
that
we
haven't
yet
deployed
that
we're.
A
A
We're
too
excited
that
for
a
community
to
form
around
these
practices
around
the
techniques
as
much
as
around
the
tools,
I'm
very
happy
that
the
tools
now
have
a
place
where
they
can
comfortably
live,
because
the
journey
that
people
in
CNC
ICD
are
on
is
very
different,
I
think
than
the
journey.
If
you
look
at
the
broader
rest
of
the
conference,
the
CNC
F
I
mean
these
are
typically.
These
are.
These
are
more
leading-edge
more
greenfield,
more
very
forward-leaning.
A
I
think
we
have
to
we
in
this
community
I
hope
we
can
foster
a
support
across
the
entire
space
of
kind
of
software
development
deployment,
even
for
people
that
are
relatively
waterfowl.
They
want
to
most
of
those.
Many
of
those
companies
want
to
get
on
a
CI.
So
how
do
we
help
that
process?
How
do
we
help
people
get
better
at
CI,
because
once
they're
better
at
CI,
they
will
start
to
get
better
and
better
at
deployment
and
release.