►
From YouTube: Kubernetes Community Meeting 20190307
Description
We have PUBLIC and RECORDED weekly meeting every Thursday at 6pm UTC.
Check this out for more information: https://github.com/kubernetes/community/blob/master/events/community-meeting.md
A
B
D
C
Right
all
right,
great
and
so
hi,
my
name
is
Jesse
and
I.
Have
my
colleague
Alex
here
on
the
line
and
today
we'll
be
talking
about
Argos
CD,
so
Argos
CD
is
part
of
the
Argo
project,
which
is
a
collection
of
open
source
utilities,
controllers
and
DevOps
tools
for
kubernetes
we're
most
known
for
our
go
workflows,
which
is
a
batch
processing,
workflow
engine
analogous
to
kubernetes
jobs.
C
Basically,
our
goal,
CD
introduces
a
concept
of
application
and
I'll
show
you
the
spec
of
that
application,
and
an
application
describes
two
things,
one.
What
you
want
to
deploy,
which
is
a
reference
to
a
location
in
git
and
to
where
you
want
to
deploy
to,
and
this
is
a
kubernetes
server
in
the
namespace
and
inside
your
git
repo.
You
can
describe
your
manifest
in
any
format
you
want.
C
C
So
this
is
our
go
CD.
So
the
first
thing
you
might
know
this
is
that
we
support
single
sign-on.
The
single
sign-on
is
powered
by
Dex,
which
used
to
be
managed
by
core
OS,
and
this
allows
Argosy
to
integrate,
with
virtually
all
forms
of
SSO,
including
sam'l,
to
LDAP,
OAD,
cor
and
so
on.
My
instance
happens
to
be
configured
to
log
in
using
github
ooofff,
alright.
So
now
that
I'm
logged
in
you
can
see
these
are
all
the
apps
that
I
currently
have
deployed
using
our
NGO
CD
and
I'll,
be
creating
a
new
guest
book.
C
App
from
example.
Repository
that
we
that
we
have
this
is
just.
This
is
basically
a
git
repository
of
a
bunch
of
guestbook
apps
in
various
forms
like
a
case
on
it
version
a
helm
version
customized,
one
customized
to
and
I'll
be
just
doing
the
customized
guestbook,
so
I'll
go
ahead
and
create
a
new
application.
C
Yes
book,
I'm
referencing
that
git
repository
once
I
have
that
git
repository
it
scans,
the
repository
for
all
like
helm,
charts,
customized
apps,
and
it
presents
it
in
this
drop-down.
So,
let's
use
the
customize
I'm
gonna,
deploy
it
to
the
same
cluster
that
Argus
CD
is
currently
running
in
and
I
don't
have
I'm,
not
gonna
customize
any
other
parameters.
C
Alright.
So
now
that
I
created
the
app
you'll
see
that
it
starts
out
in
a
out
of
sync
condition,
and
that's
just
because
we
haven't
deployed
it
to
sync:
it
I
can
either
manually
sync
it
or
I
can
enable
automated
sync
and
automated
sync
is
essentially
whenever
Argosy
need
to
tex
a
difference
from
what's
and
get
to
what's
running
in
the
cluster
it'll
just
automatically
apply
those
changes,
so
I'll
do
that.
I'll
go
that
route,
I'll
just
enable
all
those
things,
and
so
as
soon
as
I
enabled
this
oops
me
reduce.
C
Let
me
get
rid
of
these
filters
yeah
as
soon
as
I,
enable
others
think
you
saw
that
it
applied
those
manifests,
and
this
is
essentially
the
application.
You
can
see
it's
comprised
of
a
service
in
a
deployment,
and
it
obviously
shows
those
relationships
between
the
other
objects
like
a
replica
set
in
the
iPod,
and
so
now,
once
this
is
running,
let's
actually
make
a
change
and
get
to
update.
Like
the
image
tag,
so
I'll
go
ahead
and
change
which
bump
up
replicas
to
five
I
also
want
to
maybe
introduce
grace
deadlines.
C
Okay,
so
I
committed
that
repoing
right
away,
that
you'll
see
that
there's
a
web
hook
configured
from
get
to
our
go
CD
and
as
soon
as
I
made
that
change
in
get
our
go.
City
was
notified
about
that
change
and
it
automatically
started
deploying
that
because
automated
sync,
but
something
went
wrong.
So
we
have
this
company
so
right
now
our
go.
Cd
is
reporting
a
status
of
degraded
and
one
feature
of
our
go
CD
is
that
we
have
a
concept
of
Health
for
various
resource
kinds.
C
So
as
the
built-in
kinds,
new
kubernetes
kinds
like
deployments
stateful
says
demon,
said
services
ingress
jobs,
pods,
we
have
built-in
health
checks
for
those
and
and
just
recently
we
added
support
where
you
can
actually
define
your
own
health
assessment
logic
on
custom
resource
definitions
in
the
form
of
Lua
scripts.
So,
for
example,
just
recently
we
added
health
checks
for
a
certain
manager
certificate
in
issuers.
So
with
that
feature
Argosy
to
control
like
health
of
absurds,
so
to
debug
this
we
can
click
on
a
resource.
C
We
can
look
at
its
events
and
right
here,
I
can
see
that
it
says
image
pulled
back
off
and
I
know
I
happen
to
know
that
there
doesn't
exist
a
0.3
guest
book.
So
let's
fix
this
I'm
gonna
disable
auto
sync,
because
I
want
to
roll
back
and
so
from
the
now
from
the
history
of
the
application,
I
can
see.
I
can
see
what
was
deployed
and
rolled
back
to
the
previous
version.
C
Ok,
so
as
as
you
see,
things
are
happening,
it's
going
back
to
the
previous
version
and
we
are
now
back
in
a
healthy
state,
but
we
are
currently
out
of
sync-
and
this
is
expected
because
there
are
desired.
State
in
get
is
showing
those
changes
that
I
made,
where
I
bumped
the
replicas
to
5
and
change
the
image
tag
and-
and
you
can
see
the
diff
on
one
of
the
resources.
C
Ok,
so
that's
a
very
quick
overview
of
Argos
CD.
Some
of
the
other
features
that
I
wanted
to
highlight
where
Argo
city
supports
it,
concept
of
pre
and
post
sing
hooks.
These
are
similar
in
concept
to
helm,
lifecycle
hooks
and
it
gives
you
a
way
to
interject
custom
logic
in
the
form
of
like
jobs,
pods
workflows
before,
during
and
after
sync
and
internally,
people
use
this
to
do
things
like
database
migrations,
as
part
of
us
saying
also
into
it,
has
a
lot
of
compliance
and
regulatory
think
so.
C
B
E
E
There's
that
I
made
some
slides,
I
also
chopped
some
things
in
a
dock
in
the
meeting.
That's
so
you
can
follow
home
there.
If
you
would
like
me
to
the
thing
where
mustard
on
the
screen
for
a
second,
let's
see
here,
do
that
do
that?
Okay,
so
this
is
week,
nine
of
the
kubernetes
114
release
cycle,
we're
still
in
burn
down
burn
down
is
where
we
try
to
burn
everything
down.
E
So
if
I
click
this
link
here,
it
takes
me
to
the
docs
that
describes
what
burn
down
is,
but
as
a
reminder,
this
is
where
we
start
to
blow
that
up
for
a
second,
you
can
join
the
kubernetes
sake,
release
mailing
list
to
get
an
invite
to
burn
down
meetings
and
we're
going
to
start.
You
probably
can't
hear
me
very
well,
let's
see
if
that
helps
we're
going
to
start
to
ask
that
people
from
certain
SIG's
show
up,
if
necessary,
just
to
help
us
identify
what
issues
really
need
to
be
dealt
with
and
how.
E
So
there
are
some
of
you
out
there
who
are
probably
thinking
I,
should
write
some
docs
for
my
features.
Specifically,
there
are
probably
seven
of
you
out
there
who
have
Doc's
pr's
that
are
sitting
in
a
work-in-progress
State,
and
this
is
because
you
are
all
super
amazing
and
brought
out
your
Docs
in
time
for
March.
E
First,
the
next
deadline-
you
should
be
aware
of-
is
March
11th,
that's
Monday,
March
11th
monday
on
the
Pacific
time
schedule
we'd,
really
like
you
to
have
these
Doc's
ready
for
review,
so
the
docs
team
can
take
a
look
at
them.
So
if
you
know
anybody
who's
working
on
docs
for
things
like
Kubb
ATM
upgrade
to
Bay
DMH,
a
documentation,
CSI
raw
block
to
beta
things
of
that
nature,
probably
make
sure
that
they're
aware
of
this.
If
members
of
the
release
team
haven't
already
reached
out
to
them,.
E
I
wonder
why
this
slide
didn't
load?
Oh
because
it's
huge,
so
code
freeze
is
coming
code.
Freeze
is
actually
here
today.
It's
end
of
day,
March
7th
I'm,
not
sure
how
many
of
you
are
actually
listening
to
what
I'm
saying
right
now,
because
of
the
amazing
lyrics
that
Jefferies
Jaffe
wrote
one
of
the
release
team,
Docs
people
and
Catharine
Barry.
Who
is
one
of
you-
need
slack
admins,
put
a
wonderful
cinema
meme,
but
this
is
pretty
amazing.
E
E
So
if
you
want
to
find
out
more
information
about
what
code
freezes,
it's
in
our
release,
Doc's
code
freeze,
all
enhancements
must
be
code
complete,
including
tests
by
Thursday
March
7th.
At
this
point
or
after
this
point,
only
release
blocking
issues
and
PRS
will
be
a
into
the
milestone
and
that
that
tweet,
if
you
want
to
see
it
his
linked
here,
you
can
get
to
the
slides
here.
You
can
get
a
look
for
Twitter
thread.
E
Big
round
of
applause
to
Katherine
and
that
Jeffrey
second,
for
that,
okay,
other
things
about
the
release
schedule,
their
release,
schedule
intensifies.
At
this
point
we
have
a
lot
more
meetings.
We
start
daily
burn
down
meetings
Monday
through
Friday
starting
next
week.
The
plan
is
to
have
the
Tuesday
and
Thursday
meetings,
be
at
a
European
friendly
time
of
8
a.m.
Pacific.
So,
as
I
said
earlier,
join
the
kubernetes
release
team
calendar.
If
you
want
invites
to
that.
E
The
reason
these
intensify
is
because
we're
going
to
start
pruning
a
lot
more
aggressively,
so
we're
going
to
be
pruning,
enhancements,
we're
going
to
be
pruning
issues
and
we're
going
to
be
in
pruning
PRS.
So,
let's
talk
about
enhancements
real
fast
of
the
31
enhancements
that
are
currently
slated
for
release.
16
of
them
are
listed
at
risk,
I
believe
the
majority
of
these
are
listed
at
risk
because
it's
unclear
whether
or
not
they
have
test
plans
or
it's
unclear
whether
or
not
the
docs
that
they
need
are
landing.
E
If
anybody
from
the
enhancements
team
has
reached
out
to
you,
please
be
sure
to
respond.
I
believe
that
Claire
and
lucky
and
create
and
I'm
forgetting
everybody
on
the
enhancements
team,
I've
started
reaching
out
to
people
directly
on
slack,
because
github
notifications
can
be
very
noisy,
but
please
do
be
responsive
not
being
responsive.
To
requests
on
these
things
is
what
causes
us
to
start
deciding.
Maybe
we
shouldn't
talk
about
your
enhancements.
E
E
I'm,
going
to
caution
that
maybe
some
of
these
maybe
aren't
super
ready
to
be
in
here.
Maybe
we'll
start
talking
about
taking
the
8
size,
XXL
pers
that
still
haven't
managed
to
land,
because
I
don't
know
they're
failing
their
tests
or
they
still
have
merge
conflicts
or
things
of
that
nature,
maybe
are
no
longer
really
considered
ready
for
this
release.
So
just
look
for
a
lot
more
attention.
E
Please
work
with
us
and
remember
that
the
goal
of
code
freeze
is
to
take
the
existing
set
of
code
that
we
have
and
make
it
more
stable
and
make
it
ready
for
release
on
a
related
note,
I,
don't
know
if
anybody,
if
any
of
you
have
looked
at
our
CI
signal
lately,
but
there
seem
to
be
a
number
of
flaking
and
failing
tests.
You
may
be
aware
of
this.
E
One
of
the
things
I'd
like
to
bring
to
our
attention
is
the
fact
that
the
one
of
them
like
more
bog
standard
jobs
is
flaking
continuously,
so
this
is
called
GCE
cos
default,
whatever
it
stands
up
a
kubernetes
cluster
and
it
runs
all
of
the
end-to-end
tests
that
are
not
disruptive.
They
aren't
slow,
they
aren't
flaky
in
theory,
and
they
don't
require
any
extra
special
features
and
if
I
take
my
handy
dandy,
test
grade
options
here,
her
saree
size
and
turn
it
into
super
compact
just
to
try
and
drive
the
point
home
here.
E
The
top
line
here
is
all
red.
It's
almost
all
family
and
all
of
these
little
red
things
here
are
individual
test
cases
that
are
failing
at
one
time
or
another.
The
point
that
I'm
trying
to
make
here
is
that
I
can't
point
to
any
one
clear
and
obvious
test
as
the
cause
of
us
being
unable
to
successfully
see
this
job
go
red.
E
This
is
really
critically
important
for
us
because,
as
we
look
to
qualify
some
code
ready
to
go
out
the
door,
we
need
to
make
sure
that
that
code
can
actually
reliably
pass
tests
and
the
proxy
we
use
for
that
is,
can
it
test?
Can
the
tests
pass
three
times
in
a
row
for
the
same
commit
of
code?
And
the
answer
right
now,
for
this
is
no
another
way
that
this
is
especially
important
here
on
the
sig.
E
Sorry
I'm,
a
114
blocking
dashboard,
because
this
box
up
here
that's
changing
as
I'm
flipping
around
is
the
commit
on
the
release
114
bridge
it
changes
as
our
release
branch
manager
fast
forwards
daily.
So
we're
getting
good
data
on
like
how
flaky
this
particular
commit
is
I'm
just
pointing
out
this
one
test.
There
are
actually
many
other
tests
that
are
failing
and
I
encourage.
Those
of
you
who
are
reached
out
to
by
our
CI
signal
team
to
please
pay
attention.
That
is
all
that
I
have
any
questions.
E
B
B
D
A
Something
Guinevere
Sandra's
been
working
on
she's,
not
here
so
I
am
just
presenting
this
for
her.
This
is
her
and
Josh
and
Tim,
and
a
bunch
of
people
who
work
around.
Then
you
contribute
a
workshop.
So
this
is
kubernetes
thanks,
contributor
playground.
It
is
a
place
for
new
contributors
to
basically
practice
doing,
pull
requests
and
a
bot
will
talk
to
them
and
they
can
practice
all
the
kind
of
stuff
that
they
need
to
learn
before
doing
a
real
pull
request
on
a
kubernetes
repository.
A
So
what
we're
doing
is
socializing
that
people
can
practice
here
in
a
safe
place,
get
help
and
not
worry
about.
You
know,
clogging
up
the
official,
repos
and
stuff
with
spelling
fixes
and
whatnot.
So
mostly,
this
is
just
this,
isn't
something
that
most
of
the
people
on
this
call
probably
wouldn't
use.
But
if
you
could
help
us
socialize
that
this
exists
as
we
integrate
this
into
all
of
our
instructions
and
your
contributor
documentation
and
all
of
that
stuff,
it
would
really
help
us
out
and
that's
all
I
have
for
the
tip
of
the
week.
A
B
B
F
See
the
screen-
yes,
yes,
okay,
perfect,
all
right,
so
he
crease
hello.
Everyone
today,
myself,
Saleh
and
Richard
will
be
pouring
the
updates
for
say,
IBM
cloud.
So
what
we
did
in
last
cycle,
we
created
our
cig
charter
and
it
is.
It
is
marginals
available
out
there
thanks
to
steering
committee,
especially
Aaron,
for
reviewing
how
thing
does
make
it
good.
We
still
need
to
work
on
that
going
forward,
as
we,
you
know,
do
some
of
the
things
we
are
working
on.
F
The
cube
corn.
This
is
our
first
updates
after
cube,
corns,
so
I'm
just
mentioning
it
here,
you
can
see
Otto,
we
did
the
parties,
we
had
a
good
participation
with
the
sega
tivities
I.
Actually,
Richard
and
I
have
created
a
blog
post,
which
is
over
there.
The
link
is
in
the
presentation
and
then
presentation
link
is
in
the
community
meeting
agenda.
F
We
also
the
cloud
provider
specific
activities.
We
had
a
signal
spinner
there,
so
we
have
video
on
that
in
the
blog
post
and
also
the
sig
meet
and
greet
session
was
a
was
a
great
event
for
us
and
I
think
for
everybody.
But
we
made
you
know
new
contributors,
especially
you
know,
folks,
like
Marquis
Jackson
from
sis
dig.
Is
we
found
it
that
the
Sigma
done
great
and
he
is
now
a
regular.
F
F
We
don't
until
a
bit
wrong
with
the
youtube
uploads,
but
thanks
to
George
again
a
big
thanks
to
him
and
and
jiffy
I
asked
Jaffe
about
his.
You
know
he
goes
by
Jeffrey
or
he's
like
now
go
with
jiffy.
So
that's
what
a
dimensional
yep
thanks
to
jiffy,
so
YouTube
video
link
is
out
there
as
well
plans
for
upcoming
cycles.
So
this
is
the
slides
which
I
will
be
talking
to
you.
Richard.
C
Those
changes
in
our
fighters
before
this
next
cycle
will
be
writing
the
cap
for
the
IBM
cloud
provider
and
then
we
continue
to
support
the
effort
in
the
cloud
providers
saying
to
consolidate
the
cloud
provider
sticks
into
under
the
state,
one
cloud
provider
sig
umbrella
with
a
proposal
having
user
groups
or
whatever
the
end
result
will
be
there,
but
that
is
put
together
as
we
speak.
So
those
are
our
areas
of
focus
for
the
upcoming
cycle
and
ODS.
All
I
have
for
now
sign
up
back
to
you.
Yep.
F
I
think
thanks
Richard
and
that's
actually
pretty
much
it.
We
have
the
Crete.
That
was
a
quick
update,
so
you
can
find
us
on
the
slack
College
Richard
and
myself.
We
had
three
college
right
now.
There
are
a
bunch
of
links
given
here
for
the
for
the
home
page,
and
you
know
selection
and
mailing
lists.
So
please
reach
out
to
us
if
you're
any
questions
on
to
have
participation
in
the
sea
and
all
the
great
stuff,
that's
pretty
much
it
Chris.
F
B
D
Okay,
everybody
so
I
work
on
six
service
catalog,
which
is
a
kundunese
incubator
project
that
extends
kubernetes
to
implement
the
open
service
broker
API.
So
what
do
we
do?
Last
cycle,
so
we've
been
coming
out
and
expanding
our
repertoire
of
custom
resources
and
our
extension
project
movements
and
extending
those
to
namespace
versions
of
those
resources.
That's
been
sort
of
the
big
feature
work
that
we've
been
working
on
since
the
last
meeting.
D
Still
in
flight,
though
we
also
presented
it
ensuring
deep
dive
sessions
at
coop,
consi
at'll.
I've
got
a
lot
of
participation
in
those
actually
to
the
point
that
we
actually
had.
Some
new
contributors
join
up
from
s
AP
from
the
interaction
we
had
at
Seattle.
So
that's
nice
and
they've
started
work
on
transitioning
to
see
IDs,
so
we
currently
have
our
own
API
server
and
oyster
net
city.
Just
because
of
how
old
the
project
is
and
we're
interested
in
transitioning
to
co-teach.
But
now
we've
actually
started
working
on
it.
D
What
are
we
looking
forward
to
do
in
the
near
future?
So
we're
hoping
to
finish
that
namespace
resources,
that's
gonna,
be
a
zero
to
zero,
which
is
probably
going
to
be
our
last
major
release
before
going
GA
after
that,
we're
looking
forward
to
1.0
whether
or
not
Co
DS
are
going
to
be
happened
before
one
dodo,
there's
still
a
content
of
discussion,
so
it's
kind
of
still
up
in
the
air
and
outside
of
the
CRD
question.
It's
mostly
just
minor
fixes
documentation
that
sort
of
stuff
most
of
the
major
feature
works
actually
already
done.
D
D
How
these
plans
affect
you,
Service
Catalog,
fairly,
self-contained
I,
didn't
really
know
what
to
put
on
here.
So
I,
don't
know,
probably
not
gonna
affect
the
other
SIG's
very
much,
except
for
maybe
teams
that
manage
do
these
stuff,
which
is
actually
something
I
wanted
to
talk
to.
If
anybody
here
knows
anybody,
who's
worked
on
Sienese
like
actually
using
them,
particularly
if
it
was
a
migration
of
a
non-si
OD
project
to
using
CDs.
We
could
probably
use
some
help,
so
I'd
love
to
chat
with
you,
zoom
really
like
that
and
then
yeah.
D
These
are
just
the
notes
for
how
you
can
contribute.
The
easiest
way
to
get
into
things
is
probably
show
up
to
our
weekly
meeting
on
Mondays
I've
liked
a
couple
good
first
issues
here,
but
again,
I,
don't
know
it's
an
extension
project,
so
it's
kind
of
weird
and
different
from
the
rest
of
kubernetes.
So
you
probably
wanna
come
have
a
chat
for
you.
D
B
You
thank
you
Jonathan.
Would
you
mind
dropping
a
link
to
your
slide,
janky
notes,
yeah.
Thank
you
all
right
and
now
announcements
is
everyone
super
excited
for
announcements
because
I
know
I
am
alright.
There
is
a
LTS
working
group.
They
have
a
survey.
There
is
a
link
to
the
survey
and
the
community
meeting
notes.
Please
take
it
sure
it's
make
sure
that
it
has
wide
dissemination
and
that
they
get
all
the
possible
data
they
can
and
the
next
announcement
I
will
have
Jordan
Lake
himself
deliver.
Please.
G
Hey
this
is
a
multiple
release
ahead
of
time:
fair
warning
that
the
two
and
a
half
year
old
issue
to
stop
serving
extensions,
v1
beta
1,
is
making
progress
and
the
deprecated
API
is
in
that
API
group.
We
plan
to
stop
serving
by
default
in
116,
so
this
is
three
releases
out.
We
will
say
this
again:
we
will
say
it
loudly.
We
will
say
it
on
the
mailing
list.
We
will
say
it
in
the
release,
notes.
We
will
say
it
over
and
over,
but
it's
never
too
soon
to
start
migrating.
A
Yeah,
just
real
quick
we're
about
one
or
two
moderators
short
in
Europe
and
Asia
Pacific
still,
so
the
idea
is
as
soon
as
slack
fixes
their
exploit
wheels
should
be
able
to
turn
on
you
slack
signup,
so
we're
still
looking
for
Europe
and
a
pack
I
just
want
to
mention
that
the
job
isn't
just
sitting
there
signing
up
to
a
bunch
of
channels,
it's
more
of
a
reactor
rerolls
if
you're
sitting
at
work
anyway
and
you're
on
slack,
it's
very,
very
low
maintenance.
So
please
consider
signing
up.
Thank
you
also,.
A
B
Much
appreciated
all
right
shoutouts.
Let's
begin
Erica
Berger
shouts
out
to
amici
for
his
continued
review
of
great
amenities
tests
and
NPR's
he's
helped
land
over
30
be
ours.
This
release
cycle,
including
some
conformance
tests,
thank
very
much
a
lot
of
incentive,
shouts
out
to
bend
the
Elder
and
monitors,
don't
know,
I
said
otherwise,
and
the
kind
team
for
all
the
world
rate
work
that's
gone
into
the
kind
tools
and
docks.
It's
radically
changed.
B
My
interlude
for
testing
code
changes
in
kubernetes
thanks,
like
I,
also
shouts
out
to
Florence
for
being
an
awesome
enhancements,
lead
on
the
114
release
team
she's,
taken
the
time
to
coach
all
the
shadows
and
has
diligently
grooming.
The
33
features
hopper
for
114
cheers
from
your
fellow
and
handsome
shadows.
Paris
would
like
to
very
similarly
like
to
shout
out
the
upstream
mentoring
panelists
on
our
meter
contributors
meeting
yesterday,
I
explained
Carolyn
dims
DirectX
Man
12.
In
a
second
sorry,
I,
don't
know
your
real
names.
B
Dims
would
also
like
to
thank
Paris
and
George
the
mouse
for
the
meet
our
contributors.
Meeting.
I
would
like
to
personally
thank
Bob
Dylan,
first
stepping
in
to
take
notes.
The
past
two
weeks
and
the
carbonated
community
meetings
I
really
appreciate
it.
Three
con
parrots
to
you,
sir
and
Nikita,
shouts
out
to
Gigi
and
Catherine
for
the
wonderful
video
and
tweets
from
Erin.
Thank
you
very
much
for
that.
Are
there
any
other
announcements
or
shoutouts.