►
From YouTube: TAG General Meeting - 2022-12-07
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
A
D
D
F
E
Yeah
the
score
folks,
their
product
manager
is
going
to
present
to
us
in
mid-January,
so
we
want
to
I,
went
ahead
and
penciled
her
in
there
or
she
penciled
herself
in
yeah.
F
F
I
can
tell
her
that
I
won't
join
December.
4Th
I
mean
the
21st
might
work,
but
we
should
really
see
whether
we
have
any
agenda
items
there,
because
otherwise
I
think
people
can
get
their
time
in
there.
I.
F
F
Yeah
I
don't
want
to
cancel
the
21st
one,
but
it
might
just
be
a
much
smaller
round
due
to
the
items
we
have
on
publishing
some
of
the
work
and
some
of
the
organizational
shop
would
carry
water
tasks.
That
might
be
very
small,
rounded
that's
meeting!
F
That's
why
I'm
leaving
it
on
on
the
calendar
right
now,
yep
I
think
or
we
could
make
a
short
time
all
right.
We
have
four
minutes
in
that
means
we
can
start
because,
obviously
the
meeting
is
recorded
for
everyone
to
watch
afterwards.
F
As
soon
as
we
can
actually
do
it,
which
we
will
discuss
right
now,
because
the
challenge
right
now
is
and
I
filed
already
an
issue
that
we
don't
have
access
to
our
Zoom
account
or
we
will
have
access
to
the
zoom
account,
but
not
to
the
email
address.
That's
linked
to
the
zoom
account.
So
I
found
a
ticket
for
this
already
as
soon
as
they
get
access
to
that
email
address
or
distribution
list.
I'll
put
the
videos
then
later
on
online.
B
B
B
B
If
you
like
what
you
hear
or
if
you
are
interested
in
getting
to
know
more
about
this
project
that
I'm
gonna
show,
you
today
feel
free
to
reach
out
to
us,
for
example,
on
the
cmcf
captain
channel
on
Slack
or
follow
us
on
on
GitHub,
either
in
the
main
Captain
repository
or
in
the
lifecycle
toolkit
Repository
all
right.
So
let's
just
jump
right
into
it.
So
today,
I
want
to
talk
about
the
captain
life
cycle
toolkit
and
how
we
are
planning
to
kind
of
enrich
the
git
Ops
approach.
B
So
for
that
maybe
let's
first
start
with
the
points
that
we
particularly
like
about
githubs
and
what
we
learned
to
and
how
we
plan
to
build
up
upon
those
points.
So,
first
of
all,
one
of
the
main
points
is
that
gitops
is
great
because
we
have
a
declarative
state.
B
Sorry
I
just
need
to
move
my
sim
window
here,
okay,
because
we
have
a
declarative
description
of
our
application
that
we
want
to
deploy
and
also
in
a
versioned
way.
Of
course.
B
For
example,
we
can
create
a
new
text
whenever
we
want
to
release
a
new
version
into
our
environment,
variate
Ops,
10
seconds
things
are
reconciled
automatically
so
as
soon
as
PR
to
the
github's
repository
has
been
merged
or
a
new
tech
is
being
created,
projects
like
flux,
for
example,
to
take
care
of
really
taking
this
configuration
and
applying
it
into
our
cluster
and
make
and
Trigger
the
application
rollout.
B
And
finally,
this
is
pretty
much
the
natural
way
to
deliver
apps
in
kubernetes
right,
okay,
so
those
are
the
the
four
biggest
points
that
we
have
identified,
which
make
githubs
really
great
but
of
course,
which,
with
each
possibility,
there's
also
maybe
a
challenge
or
two.
So
these
are
the
points
that
we
have
identified
so,
first
of
all,
when
deploying
something
we
actually
also
want
to
make
sure
that
the
environment
that
we
want
to
deploy
in
is
actually
ready.
B
So,
for
example,
if
the
required
resources
that
we
need
for
running
our
application
are
in
place
or
if
there
are
no
open
issues
within
our
environment,
because
it
wouldn't
necessarily
want
to
release
a
new
version
in
the
in
a
broken
environment,
then
we
would
like
to-
or
it's
not
always
that
easy
to
to
tell
if
all
prerequisites
are
in
place.
So,
for
example,
if
our
application
depends
on
some
external
service
to
make
sure
that
this
is
actually
available
before
we
can
safely
roll
out
our
application,
then
also.
B
B
Finally,
of
course,
after
everything
has
been
rolled
out,
we
would
like
to
make
sure
that
our
application
is
running
properly
all
right,
so
those
are
kind
of
the
the
challenges
that
we
would
like
to
address
with
the
lifecycle
toolkit
and
now,
let's
see
how
we
actually
plan
to
do
that
or
actually
what
we,
what
the
vision
for
this
project
is
so,
first
of
all,
one
of
our
main
goals
is
to
make
the
lifecycle
toolkit
usable
with
minimal
configuration
effort.
B
Then
we
also
want
to
make
it
easy
to
integrate
external
tools
during
our
deployment
process,
for
example,
to
trigger
certain
actions
during
the
different
phases
of
the
application
deployment,
and
maybe
one
of
the
most
important
goals
is.
We
want
to
really
provide
deep
insights
into
your
deployment
process,
so
we
get
a
really
nice
overview
of
what
happens
during
the
application,
rollouts
and
also
after
that
and
yeah.
B
Basically,
we
want
to
be
Cloud
native
and
application
aware,
and
we
are
doing
that
by
yeah
kind
of
using
the
the
cube
Builder
operator
framework
and
making
all
of
the
all
of
the
captain
life
cycle,
toolkit
entities
available
as
custom
resource
definitions
that
can
in
turn
be
applied
using
Cube,
CTL
applied
and
click
next
to
the
to
the
other
other
manifests
that
might
already
be
there
in
a
github's,
Repository
yeah.
So
I've
also
already
teased
a
little
bit.
B
Exactly
so,
as
you
can
see
here
on
the
right
side,
the
traces
will
then
kind
of
depict
how
the
the
life
cycle
of
the
of
an
application
deployment
looks
like
and
what
is
being
actually
executed
during
that
deployment,
all
right
and
as
the
next
Point,
let's
see
or
maybe
prefers
before,
I
show
you
how
you
can
actually
use
it.
B
We
can
work
with
flux,
Jenkins,
Azure,
Dev,
props
or
pretty
much
everything
that
in
the
end,
does
the
cube
CTL
apply
on
your
cluster
and
when
you
do
that,
Captain
Hooks
into
the
deployment
via
mutating
that
book
that
first
checks
for
any
incoming
part,
whether
the
related
namespace
that
you
want
to
deploy
into
is
actually
enabled
for
making
use
of
the
captain,
lifecycle,
toolkit
and
then
based
on
our
custom
resources
for
the
lifecycle
toolkit.
B
You
can
Define,
for
example,
different
pre-deployment
tasks
that
should
be
executed
before
actually
placing
the
pots
into
your
cluster,
and
you
can
do
this
by
easily
customizable
functions.
So
we
provide
custom
resources
where
you
can
either
just
put
a
node.js
code
into
a
captain
task
definition
or
refer
to
something
that
already
exists
on,
for
example,
on
a
GitHub
repository.
B
Then
we
also
enable
post
deployment
steps
where
you
can,
for
example,
yeah
as
you
can,
as
you
will
see
in
the
demo,
trigger
a
slack
notification
to
be
sent
after
your
application.
Deployment
has
been
finished
and
in
addition
to
that,
we
also,
since
I
said,
we
wanted
to
make
Prometheus
the
first
class
citizen
of
the
lifecycle
toolkit.
You
can
also
Define
evaluations
that
you
want
to
execute
by
providing
a
Prometheus
query
and
a
Target.
B
E
B
B
So
that's
how
the
lifecycle
toolkit
fits
into
the
kind
of
bigger
picture,
and
now,
let's
see
how
to
actually
enable
it
and,
as
I
said,
we
wanted
to
make
this
as
easy
as
possible.
B
B
So
that's
what
you
would
add
to,
for
example,
your
deployment
manifests.
B
Then,
if
you
have
a
multi
deployment
application,
you
would
also
need
to
provide
a
captain
app
custom
resource
which
basically
provides
a
list
of
all
the
workloads
that
you
would
expect
for
this
application
to
be
deployed.
B
And,
finally,
if
you
want
to
make
use
of
tasks
and
evaluation,
that's
how
you
would
define
those
so,
for
example,
for
the
evaluation.
We
have
our
evaluation
definition,
custom
Resources,
with
the
supported,
Prometheus
queries
and
the
evaluation
targets,
and
for
the
captain
tasks.
We
provide
the
task
definition:
custom
resources
where,
for
example,
you
can
either
just
put
a
code
inline
into
that
or,
as
you
will
see
in
the
example
now,
you
can
also
point
to
a
source
file
on,
for
example,
a
GitHub
repository
right
and
now
I
would
say,
let's
jump
into
the
demo.
B
So
for
that
we
will
deploy
our
beloved
Potato.
Head
I
think
you
all
noticed,
but
basically
it's
a
micro
service
application
containing
of
different
body
parts
of
the
potato
and
in
our
example,
all
of
the
all
of
its
extremities,
like
the
arms
and
the
legs,
and
the
head
are
dependent
on
the
entry
service,
which
is
the
body
and
in
this
example
we
want
to
use
the
captain
life
cycle
toolkit
to
make
sure
that
the
entry
service
is
available
before
we
actually
schedule
the
parts
of
the
other
services
right
foreign.
B
This
kind
of
highlights
the
workflow
that
we
want
to
show
here,
but
yeah.
Maybe
let's
have
a
first
look
at
the
at
the
Manifest
that
we
are
going
to
deploy.
So
here
we
have
defined
our
captain
application,
which
contains
all
of
the
workloads
that
we
would
like
to
have
deployed
into
our
cluster.
So
we
have
the
left
arm,
left
leg
and
the
entry
service
that
every
other
service
will
check
before
it's
going
to
be
deployed
and
once
the
complete
application
has
been
deployed.
We
also
want
to
send
out
the
select
notification.
B
What
else
do
we
have
yeah?
So
maybe
yeah
here
is
the
task
definition
for
the
select
notification.
So,
as
I
said
before
here,
we
are
using
an
already
existing
source
code
file
that
that
lives
on
GitHub,
and
this
then
just
requires
the
select
token
to
be
available
in
in
the
secret
it's
referring
to
and
then
hopefully,
after
the
application
will
be
deployed.
We're
gonna
see,
select
notification,
informing
us
that
our
application
has
been
deployed
and
what
else
do
we
have
so?
First
we're
gonna
try
to
deploy
version.
B
One
and
I
think
I
already
split
a
little
bit
of
the
story
here,
but
this
evaluation
will
indeed
fail,
because
here
we
want
to
basically
check
for
the
number
of
available
CPUs
on
our
cluster
to
be
greater
than
4,
which
is
not
the
case
on
my
local
cluster.
But.
B
To
see
how
we
can
then
make
use
of
the
traces
that
we
export
in
order
to
figure
out
what
is
going
wrong
during
our
deployment.
Let's
try
to
deploy
it
anyway
right.
So
in
this
case,
I
will
execute
the
make
deploy
one.
So
this
is
nothing
really
fancy.
So
this
will
just
do
a
cube.
Ctl
create
namespace
and
then
apply
the
version.
B
So
it's
currently
checking
for
our
environment
to
be
ready
and
maybe,
instead
of
browsing
through
all
of
these
custom
resources
in
lens,
let's
see
how
we
can
make
use
of
our
grafana
dashboard
right.
So
let
me
just
check
if
it's
still
working
so
yeah
looks
good.
So,
as
you
can
see
here,
we
are
currently
having
one
active
deployment
on
the
left
side.
You
will
see
the
succeeded,
so
I've
tried
it
out
a
little
bit
before
I
prepared
the
demo
and
yeah
here.
B
We
will
then
later
see
the
the
average
time
between
deployments,
but
for
that
we
will
need
to
have
multiple
attempts,
but
and
yeah
here
you
see
a
timeline
of
the
of
the
active
deployment.
So
we
can
see
here
that
recently
we
started
it
and
finally,
here
on
this
link
will
come
to
the
trace
and
actually
now
we
immediately
see
see
that
we
have
a
problem
in
our
application
deployment.
B
So
is
that
big
enough
or
I
will
make
it
a
little
bit
bigger.
So
here
we
see
our
our
trace
for
the
Potato
Head,
one
zero
one
zero
version,
and
apparently
the
pre-deployment
task
phase
was
executed
successfully.
But
we
do
have
a
problem
here
in
our
pre-deployment
evaluation
phase
and
we
can
see
that
the
three
evaluation
app
deployment
evaluation
has
been
failed
due
to
the
number
of
available
CPUs
being
not
high
enough
right
and
yeah.
Let's
try
to
fix
that
I
would
say
so
now.
B
B
all
right
and
now
the
lifecycle
controller
will
do
its
thing
again
and
yeah.
So
now
we
see
that
we
have
a
new
instance
of
our
app
version
being
our
zero
one
one,
and
actually
it
is
already
in
the
app
deployment
phase,
which
means
that
the
evaluation
has
been
successful.
And
now
we
can
see
that
we
have
a
lot
of
parts
pending
there
and.
C
B
B
Yeah
so
now
we
can
see
that
our
deployment
is
still
active
yeah
and
we
can
also
go
to
the
trace,
which
now
looks
a
little
bit
like
a
Super
Mario
level,
but
this
will
now
make
it
easily
traceable
to
see
what
has
actually
happened
during
the
rollout
of
our
application.
So
here
we
see
the
the
root
span
of
the
complete
application
and
we
can
see
that
the
pre-deployment
tasks
have
been
executed
to
pre-deploy
evaluations.
B
B
Exactly
yeah,
so
maybe
there
was
some
issue
in
My
Lens,
but
indeed
here
we
can
see
that
the
potato
head
entry
was
deployed
and
that
face
took
took
around
45
seconds
and.
B
Okay,
I
think
I
played
around
a
little
bit
too
much
with
the
pre-deployment
checks
for
the
other
services,
but
normally
they
should
have
checked
for
the
for
the
entry
service
to
be
available
and
having
reached
the
end
of
its
deployment
phase,
and
then
they
should
have
been
actually
deployed
all
right,
but
I
guess
that's
the
Rage
of
the
Democrats
that
I
summoned
here,
but
regardless
I
think
having
this
overview
is
big
benefit,
especially
when
it
comes
to
tracing
back
the
deployments
of
quite
complex
applications
and
also,
let's
check,
select
yeah
exactly
now.
B
B
All
right
and
with
that,
are
there
any
questions
so
far.
B
All
right,
if
not
I,
have
almost
reached
the
end
of
the
presentation.
So
what
we
then
also
hope
to
provide
with
the
lifecycle
toolkit
is
day
two
operations.
B
So,
for
example,
after
the
post
deployment
checks
have
been
executed,
we
want
to
provide
enough
data,
for
example,
to
to
to
the
traces
that
you
just
saw,
or
by
kubernetes
events
that
are
sent
by
the
lifecycle
controller
to
then
reach
a
kind
of
Action
level
state
so
to
provide
external
tools
enough
information
to
trigger
any
follow-up
steps.
B
When,
for
example,
one
of
the
post-deployment
checks
has
failed
in
order
to,
for
example,
trigger
a
function
or
a
workflow
that
tries
to
remediate
a
certain
problem
or
maybe
switch
the
traffic
to
a
previous
version
or,
of
course,
apply
changes
or
maybe
do
a
revert.
The
changes
that
we
made
on
our
git
github's
repo
before
in
order
to
go
back
to
a
previously
deployed
version
or
maybe
to
activate
or
deactivate
some
feature,
Flags.
E
Could
could
you
just
I
just
wanted
to
clarify
that
last
one?
Are
you?
What
are
you
hooking
into
to
to
react?
Are
you
you
know
running
health
checks
again?
Are
you
relying
on
you
know
post
deployment
because
it
sound
I?
Guess
the
post-deployment
opportunity
is
finished
at
this
point,
so
you're
hooking
into
something
else.
B
Exactly
in
what
we
do,
in
addition
to
doing
the
post
deployment
tasks,
the
lifecycle
controller-
maybe
let's
check
it
here-
we
will
also
send
kubernetes
events
got
it
with
the
with
the
lifecycle
controller
and
that
can,
for
example,
be
picked
up
by
a
monitoring
solution
to
see
if,
for
example,
one
of
the
deployment
checks
has
has
failed.
E
Not
sure
because
so
I'm
more
wondering
like
after
you
run
the
first
round,
I
guess
of
post
deployment
checks
after
the
Pod
is
I,
guess
up
and
ready.
It
sounds
like
you
still
do
like
day.
E
Two
monitoring,
and
can
it
sounds
kind
of
like
I,
want
to
ask
that
you
could
run
a
captain
task
definition
on
day
two
or
day
three
in
reaction
to
something
I
just
wanted
to
clarify
what
you're
reacting
to
so
you
said
these
events
that
would
be
I,
guess
the
you
know
they
would
report
if
your
post
deployment
check
worked
but
like
on
day
two
or
day
three
long
after
your
post
deployment
stuff
ran.
What
are
you
reacting
to
other
events.
B
Okay,
yeah
for
the
day
or
for
the
longer
time
after
currently,
we
don't
have
too
much
information
right
now,
but
we
are
working
on
figuring
out
how
to
also
provide
these
kinds
of
possibility,
but
with
the
current
state
of
the
captain
life
cycle
toolkit
after
the
deployment
has
been
finished
and
after
the
post-deployment
checks
have
been
done.
That's
right
now.
What
tells
the
lifecycle
toolkit
that
the
application
is,
in
its
let's
say,
reconciled
state
cool.
E
C
B
B
So
what
we
have
now
seen
is
a
application
aware
deployment
with
render
neutral
observability
included.
So
we
are
exporting
all
of
the
metrics
that
the
lifecycle
toolkit
Gathers
through
the
open,
Telemetry
collector,
which
makes
it
easy
for
different
Monitoring
Solutions
to
to
gather
this
data
and
the
traces
and
yeah.
As
you
have
seen
on
the
example
of
the
slack
notification,
we
aim
to
make
it
as
easy
as
possible
to
also
integrate
external
tools
and
I.
Think
this.
B
This
task,
definitions,
where
I
can
just
directly
and
put
some
code,
is
a
is
a
pretty
good
way
to
do
so.
But
of
course,
we
are
also
looking
for
feedback
on
how
to
further
or
how
to
further
improve
this,
and,
as
you
have
seen,
the
installation
and
the
configuration
is
pretty
easy.
So
we
are
relying
on
the
kubernetes
recommended
labels
and
in
addition
to
that,
if
you
have
a
multi-workload
application,
you
will
just
need
to
provide
a
simple
cat
map.
B
Okay,
yeah
and
if
you
want
to
get
in
touch,
feel
free
to
reach
out
either
on
the
CNC
app
slack
on
the
captain,
Channel
or
if
you
want,
would
like
to
contribute.
We
are,
of
course,
always
happy
about
feedback
and
are
more
than
happy
to
look
at
any
incoming
PRS
from
the
community
and
with
that
I
will
hand
it
back
to
always.
Thank
you.
F
Yeah,
thank
you
again.
A
video
will
later
on
be
available
as
well
and
but
I
would
also
recommend
that
you
maybe
post
also
a
link
in
the
app
delivery
channel.
Other
projects
are
doing
it
as
well
with
the
presentation
where
you
can
continue
the
conversation
then
as
well.
E
I
can
I
put
this
down.
I
got
a
couple
things
that
are
maybe
chop.
Wood
carry
water
kind
of
stuff.
I
could
just
go
through
them
now,
if
you
want
so
this
first
one
is
just
I
was
working
with
Robert,
and
so
it
was
just
clarifying
what
these
are.
E
If
we
should
keep
them,
if
they
are
useful,
I
actually
related
to
them
might
need
a
mailing
list.
One
of
my
questions
is
how
can
I
I've
been
trying
to
run
one-off
meetings
and
the
platforms
working
group,
and
so
I?
Don't
really
have
a
way
to
let
people
know
so.
Yeah
I
was
just
wondering
what
those
are
for
you
know.
I
could
open
an
issue
to
discuss
it.
If
that's
preferable
just
wanted
to
know
how
to
I.
F
Think
yeah
I
think
the
the
short
answer
is
that
this
is
more
or
less
historically,
you
know
for
the
github's
working
group
we
created
this
before
the
github's
working
group
was
actually
established,
but
not
since,
since
then
I
was
working
on
open,
githubs
I.
Don't
see
a
reason
why
we
would
maintain
them
honestly.
F
That
this
is
more
or
less
for
collecting
initial
feedback
on
who's
interested
to
join
an
initiative
before
we
officially
propose
something
and
to
have
an
official
way
to
actually
do
it,
because
in
a
Word
document
or
like
in
a
Google
doc,
you
would
just
throw
it
in
there.
In
this
case,
irrespective
person
has
to
file
a
PR,
but
I,
don't
necessarily
see
that
we
need
to
keep
maintaining
them.
D
F
But
yeah
I
would
I
would
create
a
PR
to
remove
it.
We
can
send
it
out
on
the
official
mailing
list.
I
mean
Scots
here
as
well,
unless
the
github's
working
group
and
also
open
githubs
won't
have
any
issues
with
it.
I
mean
there's,
there
is
I,
think
very
little
value.
That
was
not
that
much
about
in
continuing
to
to
keep
but
I
think
that's
this,
because
the
second
one
was
I.
Think
for
the
platform
working
group
right.
E
Okay,
thanks
I,
don't
want
to
Rat
Hole
on
that
too
much
the
other
couple
things
I,
guess
the
the
second
one
media
I
have
a
PR
so
I'm
just
putting
that
there.
If
somebody
could
approve
the
pr,
because
we
did
change
the
Ouija
platforms
schedule.
E
So
I
guess
that's
just
general.
You
know
be
on
the
lookout
for
PRS
I'm
about
to
to
ask
you
about
a
PR
for
the
platforms
we
G
so
just
put
that
in
there
and
then
I'll
always
like
you
mentioned
I
just
want
to
work
with
you
on
publishing
the
videos
for
for
the
Ouija
meetings.
Yes,.
F
And
as
I
mentioned,
we
we
are
yeah
I
agree
with
that.
This
look
for
for
a
while,
but
now
that
we
try
to
re-log
in
we
couldn't
that's
why
the
videos
are
not
there
and
that
we
could
maybe
just
do
a
voluntary
round,
because
it
did
some
work
to
do
some
of
the
post-processing
to
be
fair
and
I
want
to
just
like
evenly
distributed
across
people
to
get
the
videos
done
and
then
upload
them.
The
upload
is
actually
the
easy
part,
but.
E
Okay,
just
let
me
know
how
I
can
help
but
yeah,
because
that'll
help
us
with
the
letting
other
people
know
what
happened
in
the
meetings.
E
Okay,
so
yeah
I'll
jump
on
now
to
the
Ouija
platforms,
update
I'm,
the
lead
for
Ouija
platforms.
That's
why
I'm
jumping
in
here
so
we're
working
on
this
platforms.
Definition,
it's
I,
think
it's
coming
along
decently!
Nicely
you
can
you
could
jump
in
there
and
see
where
we're
at
the
goal
of
it
is
to
inform
Enterprise
leaders,
Enterprise
Architects
people
considering
implementing
a
platform
or
or
trying
to
make
sure
their
platform.
You
know
meets
the
needs
that
they
want
it
to
meet.
E
That's
what
it's
being
written
for
so
kind
of
setting
definition
for
what
a
platform
you
know
means
to
us
or
means
in
distributed
computing,
defining
the
key
attributes
and
values.
Well,
the
key
values
that
we
expect
Enterprises
to
get
from
it,
the
attributes
that
would
contribute
to
those
values
and
then
the
list
of
capabilities
which
I
I
think
is
pretty
interesting.
The
10
or
15
things
that
are
should
haves
or
must-haves
in
a
platform.
E
So
yeah
we're
working
on
that.
We
hope
to
publish
something.
You
know:
I
I
set
a
goal
of
late
December,
but
that's
it's
more
likely
to
be
early,
January
I'm
sure,
don't
tell
anyone!
I
said
that,
but
yeah
so
we've
been
having
meetings,
we
had
a
couple
meetings
the
past
few
weeks.
If
you
want
to
join,
we
have
one
at
1900
UTC
today
just
use
the
the
Ouija
platforms
channel
in
slack,
and
you
can
find
the
zoom
invite
at
the
top.
E
There
I
wanted
to
just
share
this
one
last
bit
which
I
wrote
down,
which
I'm
planning
to
chunk
it
up.
So
I
just
described
kind
of
the
outline
of
the
paper.
It's
got
like
five
sections.
You
know
what
a
platform
is.
What
the
platform
team
is
the
values,
the
attributes.
The
capabilities
I
want
to
try
to
chunk
that
up
and
put
it
in
like
five
docs,
maybe
in
a
directory
or
five
sections
of
docs
in
a
directory
in
GitHub,
so
I'm
bringing
it
up
because
I
need
PR's.
E
F
Could
you
do
me
one
favor,
especially
on
PRS?
Could
you
also
send
it
out
on
the
mailing
list.
F
F
A
F
I
most
likely
read
my
emails
more
frequently
than
those
other
bits
and
pieces
so.
E
Yeah,
that's
really
good
feedback,
and
that's
probably
that's
I
was
actually
wondering
like
how
do
I.
Let
people
know
that
we're
having
this
one-off
meeting
today
and
I've
been
doing
it
in
slack,
but
should
I
be
I
should
probably
be
sending
to
this
I'm
gonna
find
it.
E
C
E
E
E
You
know
it's
always
kind
of
a
challenge,
but
that's
the
next
step
and
then
next
level.
So
basically,
this
goes
actually
to
that
multi-tenancy
thing.
So
after
we
kind
of
Define
the
especially
the
capabilities
I
feel
like
which
are
things
like
well
API
and
portals
and
scenarios,
but
also
things
like
databases
and
cues
or
observability
systems
or
identity
systems
or
development
environments,
there's
a
few
different
ones.
E
Multi-Tenancy
is
one
of
those
that
we
identified,
which
is
relevant
to
platform
Engineers
but
probably
transparent,
to
like
app
developers.
So
we're
not
going
to
this
paper
won't
go
much
into
that,
but
it
will
mention
it
so
like
building
on
these
lists
and
Diving
a
level
deeper
into
it's
actually.
The
second
thing
here:
the
projects
related
to
the
capability
so
like.
If
we
choose
say
development
environments
is
a
capability
and
a
platform,
then
maybe
the
next
step
is
to
do
another.
E
Do
a
level
deeper
on
here's
all
the
development
environments
that
we
manage
in
cncf
the
projects
that
we
have
if
database
you
know
management
or
database
provisioning
is
another
one
that
you
know.
That
could
be
one.
So
that's
kind
of
the
ideas
like
this
paper
gives
us
a
jumping
off
point
for
going
a
level
deeper
in
each
of
those
areas.
E
Yeah
and
then
the
other
thing
I
wrote
down
there.
Another
thing
that
would
kind
of
emerged
from
the
from
our
discussions.
There
has
been
that
platforms,
there's
there's
two
real
features:
I
guess
or
or
what
a
platform
offers
to
the
users,
which
is
either
kind
of
disaggregated
capabilities
like
get
a
database
I'll
get
an
identity.
E
I'll
get
some
compute
space,
a
namespace
and
kubernetes
I'll
get
a
development
environment,
that's
one
kind
of
like
Individual
Services,
but
then
we
also
have
like
scenario:
services
like
what
people
call
a
developer
platform
is
really
combining
like
five
or
six
or
ten
of
those
capabilities
into
a
complete
template.
So
you'll
get
you
know
a
development
environment
and
a
database
and
a
kubernetes
namespace,
Anna,
Jenkins
or
tecton,
or
Argo
pipeline
or
reconciler
all
those
things
together.
E
So
that's
another
I
mean
we're
trying
to
capture
that
in
this
paper
it's
kind
of
that's
one
of
the
things
we're
trying
to
figure
out
I'm
trying
to
figure
out,
but
that
might
be
something
to
follow
up
on
also
like
what
are
the
kinds
of
platforms
that
people
build.
I
mean
we've
identified
app
Dev
ml
Ops,
but
there
might
be
others
too.
E
I
I
talk
about
cops
apps,
you
know,
I
just
want
to
run
Mongo's
Helm
chart
in
a
cluster
or
something
or
you
know,
information
worker
I
want
to
run
remote,
desktops
or
something
like
that
or
office
or
something
yeah.
So
that's
another
area
to
go
into
next
yeah,
so
I'll
stop
there.
That's
I
just
wanted
to
share
yeah.
F
G
E
So
a
couple
things
I'm
thinking
of
one
is
that
the
folks
from
Dapper
offered
to
help
when
we
met
in
kubecon.
They
said
they
had
a
couple:
people
that
we
could
interview.
I
mean
I've,
been
sharing
the
guitar
GitHub
template,
where
we
have
kind
of
a
question
that
people
could
answer
about
it.
You
know.
Clearly,
that's
not
working.
E
People
aren't
just
gonna
go
do
it,
but
we
can
use
that
as
the
foundation
for
the
question
so
like
at
least
when
I,
when
I
talk
to
the
Dapper
folks
I'm
like
here's
the
questions
we're
going
to
ask
here's
what
we're
going
to
talk
about
now.
Let
me
share
that.
So
you
know
what
I'm
talking
about
yeah
here's,
this
I'll
put
it
in
the
chat
for
now.
E
There's
the
there's
kind
of
the
rough
outline
of
what
we
want
to
talk
to
people
about,
but
then
yeah.
So
that's
one
option,
maybe
I'll
be
able
to
get
a
little
through
my
work,
Red
Hat,
you
know,
I
do
I
definitely
do
get
to
hear.
Sometimes
we
get
teams
talking.
E
We
just
had
a
couple
companies,
the
past
few
weeks
talking
about
their
platforms
and
and
what
their
experiences
are
so
I'm
getting
that
way,
but
I
can't
share
it
so
broadly
and
then
oh
yeah
I
wanted
to
see
about
the
user
group
research
that
cncf
has
and
I
was
wondering
if
they
are
gonna,
if
they
would
be
at
all
helpful
in
that
anyone
have
thoughts
on
that.
Yeah.
F
They
are
helpful,
I
mean
they're,
very
nice
people,
it's
just
that
everybody
has
the
same
idea,
so
that
should
be
helpful
to
everybody
out.
There
I
think
that's
kind
of
the
challenge
with
them,
but
I
would
also
like
reach
out
to
to
to
projects
that
you
consider,
because
a
lot
of
projects
have
success
stories,
use
cases
because
they
also
to
some
extent
need
and
want
them,
and
maybe
those
people
are
more
willing.
F
So
if
you
have
already
done
a
success
story
with
backstage
I
would
assume
that
the
same
people
would
be
willing
to
also
talk
to
you
and
that
might
also
make
a
difference.
Actually
talking
like
having
like
an
actual
interview,
if
possible,
I
know,
there's
some
some
work
there,
some
additional
work,
it's
like
just
filling
it
out,
I,
just
briefly
looked
at
it.
I
would
not
necessarily
know
what
to
write
to
where
and
I
think
it's
easier
to
maybe
walk
somebody
through
those.
C
H
Maybe
we
need
to
think
about
those
areas
I
and
fill
it
up
for
my
company
and
you
can
set
it
inside
his
developer,
a
bunch
more,
but
we
need
most
of
them
doesn't
know
when
those
terms,
so,
at
least
in
that
area.
E
F
I
think
you
might
have
to
run
the
first
couple
of
them
live
and
I
think
that
it's
going
to
be
the
guidance
for
people
will
actually
increase.
I
think
it
would,
for
example,
be
easier,
maybe
even
for
people
to
highlight
projects
that
they're
using
like,
if
you
take
the
cncf
landscape
and
also
beyond
our
CTC
ncf
landscape
projects
that
they
might
be
using
and
might
also
give
some
good
input
and
just
as
a
spoiler
I
might
have
a
an
interview
partner
for
you
and
be
some
data
on
on
projects.
F
I'm
not
entirely
sure
how
much
we
can
share,
but
obviously
Rio
dynatories
have
data
which
cncf
products
are
used
more
or
less
together
to
build
platforms
and
and
how
frequently
they
use
the
media
based
on
our
Enterprise
customer
base,
and
we
discussed
it
already
internally.
We
just
have
to
make
sure.
F
E
F
E
F
Also,
the
container
registry,
a
link
for
container
Registries,
which
we
have
the
same
for
Harbor,
for
example,
Argo
folks
I-
think,
there's
lots
of
opportunities
of
people
who
to
talk
to
it's.
It's
still
a
lot
of
work.
I
mean
don't
don't
get
me
wrong,
but
they
might
have
already
good
contact
with
some
of
those
people.
F
G
Oh
no
worries
no
worries.
I
literally
just
wanted
to.
Let
folks
know
about
updates
in
the
adopt's
working
group,
but
not
it's
not
a
big
presentation.
It's
really
just
to
say
that
I'm
super
excited
that
flux
graduated
within
cncf.
It
was
like
a
really
a
really
long
and
important
process.
G
I
have
more
details
of
that
for
anyone
that
needs
to
that
wants
to
know
and
I
should
also
say
that
that
Argo
had
just
announced
that
they
graduated
as
well.
So
that
is
a
good
thing
too,
and
and
Netflix
we
on
the
flux
team.
We
waited
to.
We
waited
until
the
cncf
announced
this
and
coordinated
it
with
cncf
I.
Don't
believe
that
was
the
case
with
Argus
I
don't
have
as
many
things
to
point
to,
but
I
do
I.
G
Think
more
information
will
be
coming
out
soon
about
that,
but
yeah.
It
is
an
amazing
weeks
for
githubs
and
there's
a
a
certification
coming
up
for
get
Ops
through
the
Linux
Foundation.
That
is
a
collaboration
between
the
CD
foundation
and
cncf.
G
So
if
anyone
wants
to
know
more
about
that,
please
reach
out
to
me.
They
can
also
reach
out
to
Brad
McCoy.
There
are
other
people
as
needed,
but
that's
just
kind
of
what's
happening
in
the
landscape.
That's
big!
Of
course.
There
are
other
things,
but
we
can
save
the
other
updates
for
another
time.
F
G
Yeah
and
and
I
think
in
terms
of
industry-wide.
That's
definitely
the
case.
I
mean
all.
There
are
yeah,
so
they're
they're
many
different
Cloud
providers
and
major
organizations
that
have
already
folded
this
into
their
offerings
and
others
you
know
are
are
are
working
on
it.
It's
it's
yeah,
it's
pretty.
It's
pretty
exciting.
G
G
G
That's
just
about
to
start
and-
and
it
was
really
great
seeing
you
know
in
the
within
the
scope
of
of
of
get
UPS
working
group,
I,
I
suppose
also
it's
important
to
to,
like
you
know
like
say,
like
Thomas,
Tomas
and
Florian.
That
was
a
great
presentation.
G
I
know
that
Captain
it
expands
Beyond
githubs,
but
it
also
it
also
connects
there.
So
that's
really!
That's
really
cool.
C
F
We
will
work
on
making
this
available
publicly
again
again.
There
is
an
open
ticket
with
service
desk
that
we
need
to
follow
up
and
to
be
be
better
on
this
charge.
I
also
owe
you
the
details
on
the
mailing
list
with
trouble
share
with
you,
because
oh.
F
We
also
had
in
the
past,
but
that
kind
of,
like
slagged
as
well
like
combining
key
updates
around
app
delivery.
Related
topics-
I-
wouldn't
call
it
in
newsletter
because
that
would
kind
of
like
overdoing
it,
but
we
might
want
to
re-initiate.
We
need
to
instantiate
this
as
well
going
forward
to
obviously
drive
more
interaction,
which
I
think
these
are
mostly
happening
asynchronously
these
days,
because
we
see
people
joining
the
meetings,
obviously
when
they
have
something
to
present
or
when
they
have
time
but
more
obviously,
reading
offline.
F
Given
that
the
all
on
tight
schedules
these
days,
all
right
thanks,
everyone
yeah
most
likely,
not
the
reading
meeting
in
two
weeks-
very
unlikely
in
the
first
week
of
January,
but
in
two
weeks,
maybe
it's
just
a
smaller
group,
is
getting
together
to
discuss
some
of
the
organizational
topics
and
and
how
to
improve
organizational
efficiency
there.
So
I'll
put
this
in
the
agenda
that
this
will
be
the
main
focus
there
and
with
this
have
a
nice
rest
of
your
day.
Bye.
Everyone.