►
From YouTube: Configure team meeting - 2020-12-07
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).
B
C
That's
that's
pretty
low
yeah,
that's
like
almost
17
degrees
celsius,
17
celsius,
yeah,
that's
pretty
low.
I
wear
like
my
sweatpants
and
my
hoodie.
These
are
pile
lined
actually
from
uniqlo
so
like
these
are
warm
like.
You
wouldn't
normally
wear
these
indoors.
A
C
It's
he's
just
trying
to
save
money
here.
Yeah
I
mean
it's,
not
our
house.
This
is
how
it's
been
done
so
far.
Yeah.
I
have
to
just
get
up
real
quick
to
let
the
cat
out.
D
D
D
This
I'm
showing
my
disheveled
house
how's
everyone
doing.
E
D
D
A
A
C
Well,
there's
a
website
which,
where
only
one
person
has
to
buy
the
game,
it's
called
boardgamearena.edu.
G
C
Would
check
it
out?
It's
it's
a
bit
like
I
mean
it's
a
website,
so
it's
very
accessible
works
on
all
computers.
You
need
a
big
screen
for
some
of
these
board
games
to
be
like
playable,
but
it's
I
used
it
in
the
past.
I
I
subscribed.
I
I
cancelled
because
I
didn't
use
it.
We
I
used
it
for
like
a
couple
of
months
with
friends,
it's
only
five
dollars
a
month
and
you
there's
no
binding
period,
so
you
can
just
subscribe
and
then
cancel
right
away.
E
A
What
what
do
we
expect
to
ship
in
1307?
I
took
the
planning
issue
and
I
know
a
few
items
that
we
do
expect
to
ship.
So
I
did
not
add
those
to
this
list.
A
I
just
added
the
ones
that
where
I
have
question
marks,
the
first
is
about
whether
cos
is
being
enabled
by
default
or
not
in
omnibus
and
in
hand
charge,
because
I've
seen
from
joao
that
he's
working
on
that.
But
I
don't
know
where
we
stand
with
it.
A
Just
okay,
let
me
last
draw
as
well,
probably
just
if
you
know
the
the
second
is
the
actions
in
the
terraform
state
listing
view.
B
D
Then
should
should
we
wait
on
that?
One
then
victor
until
we
have
all
all
three,
it
might
just
be
a
better
release
post
and
a
more
clear
yeah
announcement.
G
Yeah,
I
think
it
should
be.
I've
been
splitting
it
up
into
as
many
pieces
as
I
can
because
there's
just
heaps
of
lines
to
move,
but
I
don't
see
any
reason
why
not
cool.
A
Think
about
the
release
post
that
if
you
can
ship
something
amazing,
that's
what
people
are
most
likely
to
read
and,
for
example,
when
we
ship
guitar
from
features,
we
got
users
who
upgraded
the
gitlab
instance
because
they
saw
those
features
in
the
release.
D
All
right
so
looks
like
we
got
two
feature
issues
to
discuss:
we're
going
to
continue
to
discuss
the
reverse
tunnel.
D
To
begin
with,
I
think
we're
trying
to
drive
towards
a
a
definition
of
done
for
this.
I
haven't
gone
through
all
the
latest
kind
of
threads
here
can
can
anyone
sum
up
their
vision
for
what
their
like?
How
this
could
work.
G
F
F
And
that
means
there
may
be
multiple
agents
that
allow
to
access
them
from
a
ci
job
and
then
because
ci,
the
like
the
philosophy
behind
permissions,
the
permissions
model
is
the
ci
job
is
getting
the
permissions
of
a
user
who
kicked
it
off,
so
that
would
mean
the
agents
should
for
any
actions
it
performs,
should
impersonate
the
user.
Who
is
you
know
who
kicked
off
the
job
this
just?
F
F
If
I'm,
if
I'm
not
misunderstanding
something-
and
here
we
have
I-
I
guess
this
is
a
gener
general
case,
and
now
we
need
to
think
maybe
we
can
start
with
something
you
know
simpler
like
it's
a
ci
job
for
this,
for
the
configuration
project
itself,
and
then
you
don't
need
any
permissions
like
explicit
grant
of
permissions,
then
perhaps
this
whole
paper
pipeline
all
the
jobs
in
the
pipeline,
get
access
to
the
cluster
managed
by
the
agent
in
this
same
project,
and
that
was
a
suggestion
by
victor.
I
think
it's
an
excellent
suggestion.
F
And
then
in
the
comment
victories
asking:
how
do
we
ensure,
like
only
some
people,
are
authorized?
Some
users
are
authorized
to
pre
to
basically
make
use
of
this
connection
to
an
agent,
but
it
follows
from
how
it
works
if
we
impersonate
the
user.
That
means
any
any
mechanisms
that
are
in
place
already
to
restrict
what
the
user
can
do
should
be
just
it
should
just
work
naturally,
so
in
a
convenience
cluster
that
would
mean
airbag
or
any
other
authorization.
F
Things
in
place
like
admission,
workhooks
and
customer
base,
or
maybe
whatever
you
have
there,
that
decides
whether
something
is
allowed
or
not.
Those
things
can
just
do
their
job
properly.
So
it's
a
matter
of
for
the
cluster
administrator,
it's
a
matter
of
putting
the
proper
checks
in
place
and
we
don't
need
to
have
another
layer.
On
top
of
that,
I
think
that.
F
Adds
any
restrictions,
so
what
we
could
do,
what
we
could
do
is
we
could
to
the
identity
of
the
user.
We
could
add
extra
information
which
can
be
considered
by
the
authorization
mechanism.
On
the
kubernetes
side.
There
is
a
special
fee,
not
field,
but
the
group
of
fields
extras
and
then
we
can
also
not
maybe
can
we
we
really
should
do.
This
is
to
populate
users,
groups
with
gitlab
groups,
they're
part
of
so
that
again
authorization
mechanism
can
take
into
account
this
this
stuff
as
well.
F
F
I
think
there
is
also
a
project
called
oop
open
policy
agent.
I
think
it
allows
you
to
write
a
more
kind
of
advanced
logic
to
perform
authorization,
checks
and
user
like
cluster
admins
could
use
this
as
well.
So
there
is
so
much
so
many
things
that
allow
you
to
configure
permissions
in
a
very
fine-grained
way.
I
don't
think
we
need
to
build
yet
another
one.
I
think
our
job
is
to
just
propagate
the
identity
of
the
of
the
user,
and
I
hope
it
can
be
deduced
somehow
from
the
xi
ci
token.
E
E
Like
a
lot,
it's
a
common
feature
request,
so
I
don't
know
so
I
think
that
all
sounds
okay
and
then
I
think
we
also
have
mvc,
which
is
great,
and
if
we
want
to
kind
of
that,
that
implies
that
currently,
the
the
developer
role
in
gitlab
is
is
allowed
to
access
the
cluster,
and
if
you
want
to
restrict
it
further,
then
it
sounds
like
the
terraform
thing
where
we
can
have
protected
branches
as
well,
where
it
can
only
run
on
particular
branches
to
restrict
the
maintainer.
E
The
other
thing
that
I
haven't
seen
discussed
is:
do
we
want
to
allow
any
runner
to
be
able
to
access
your
cluster?
E
E
But
I
and
there
might
be
a
separate
follow-up.
I
don't
know
any
thoughts
on
that.
E
Yeah,
it's
default
yeah.
It's
default
currently
gitlab
defaults
to
allowing
share
runners.
There's
a
button
you
can
press
in
your
project.
Settings
to
disable
share
runs
so
yeah.
F
E
Oh
right,
yeah
there's
no
link
between
jobs
and
and
runners.
Unfortunately,
so
any
runner
can
pick
up
the
ci
job
yeah,
so
the
link
gets
made
when
the
runner
picks
it
up.
So
we
we
don't
know
and
the
only
yeah,
the
only
kind
of
place
where
we
can
put
a
stop
to
that
is
at
a
project
settings
level.
There's
no
other
place
well.
F
F
So
if
ci
drop
is
fine
to
be
run
on
on
a
particular
runner,
then
like
we
attached
to
the
ci
job
not
to
run
after
the
runner,
yeah
yeah
make
sense
like
it
would
be
a
duplication
of
functionality
so
to
some
extent,
if
we,
if
we
also
support
restricting
it
on
this
on
this.
D
Piece
of
chain
right,
my
my
my
intuition
is
saying
just
leveraging
existing
like
authorization
like
fundamentals.
It
sounds
like
we
have
fundamentals
around
who
can
access
what
job
and
what
users
can
access?
What
projects?
E
So
I
think
that
all
looks
good
is
for
mvc,
though
I
don't
know,
I
don't
really
understand
what
happens
if
we
have
multiple
projects.
Is
it
like
something
like?
We
will
have
multiple
contacts
and
then
the
user
has
to
choose
between
one
or
something
like
that
is.
F
That
right
maker,
so
that's
what
we
are
discussing
so
for
mvc,
we're
talking
about
granting
access
to
ci
of
the
configuration
project
and
nothing
else.
F
F
C
F
F
So
the
problem
with
defaulting
is-
and
I
described
it
in
the
issue-
is
if
you
have
multiple
agents
that
grant
access
and
you
had
just
one
before
which
was
the
default.
Now,
when
you
have
multiple
what
like,
which
one
will
be
the
default
one,
so
we
can
have
a
default.
That
means
your
previous
job
that
relied
on
the
default
now
suddenly
stopped
working.
C
We
can
make
the
default
always
the
oldest.
I
was
thinking
about
this
in
the
in
the
context
of
doing
the
same
for
for
the
old
cluster
integration,
where,
like
we
have
a
uniqueness
constraint
on
the
environment
scope,
but
so
only
one
cluster
can
have
a
given
scope.
I
think,
but
or
we
enforce
it,
so
you
can
only
ever
get
one
one
cluster,
but
we
could
in
theory,
send
you
all
clusters
to
the
matrix
scope,
but
you
have
to
somehow
do
it
in
a
backwards
compatible
way.
C
F
I
think
you
see
there
is
not
much
difference
between
one
and
all
like
you.
We
need
to
have
a
reverse
index
from
a
project
to
what
agents
allow
this
project
to
access
them
somewhere
in
the
database,
and
once
you
have
the
index
like
you,
just
look
it
up
construct
all
the
contacts.
F
D
F
I
think
there
should
be
an
environment
variable
that
cuba
ctl
respects,
which
tells
it
which
contacts
to
use
and
the
the
path
to
the
file
and
which
context
from
the
file.
So
I
hope
there
are
two
variables
I
I'm
pretty
certain.
There
is
a
variable
to
where
you
can
specify
the
path
and
we
can
set
that
automatically,
but
with
the
context.
E
So
so
we
meet
the
so
we
create
a
file
and
then
we
emit
like
a
variable,
with
a
path
to
the
file
which
is
kind
of
what
we
do
for
the
current
kubernetes
management
anyway,
and
then
we
don't
set
a
default
cute
contacts
and
then
we
have
some
sort
of
instructions
to
say:
here's,
here's,
the
format
of
the
context,
names
and
then
you
can
set
whatever
your
context
you
want
and
then
it
will
magically.
H
F
H
H
I
mean
we
have
kind
of
talked
about
shipping
k3s,
just
basically
change
omnibus
into
a
k3s
distribution
and
gitlab
ctl
becomes
rapper.
D
All
right
so
do
we
feel
so
I
guess
to
me,
like
I'm
feeling
clear
on
the.
D
F
F
D
H
F
H
D
Okay,
all
right!
Well,
I
will
I'll
take
all
that
and
craft
up
the
definition
of
don
for
this.
F
So
I
think
the
unknown
is
what
how
to
make
this
happen
in
in
the
runner,
because
it
should
be
the
thing.
That's
creating
the
configuration
file
right.
C
Or
might
be
well
this
one,
we
already
have
a
precedent
for
with
the
old
cluster
integration,
so
it
would
just
be
picking
together
the
agent
like
the
authorization
like
it's
all
on
the
gitlab
backend.
Basically,
so
the
way
it
works
is
that
you
can,
when
a
build
is
being
created,
you
can
like
choose
which
environment
variables
it
gets,
and
then
at
that
point
you
know
this
build,
has
agents
or,
like
this
project,
has
agency
check
for
that,
and
then
you
add
the
extra
variables.
C
F
Okay,
so
when
it's
a
path
to
file,
you
can
also
provide
the
contents.
Okay,
okay,
then
I
guess
another
thing
that
comes
to
mind.
Is
this
reverse
index,
where
it's
the
agents
that
grant
access
to
a
project
or
multiple
projects,
but
then
the
ci
runs
for
a
project,
so
how
a
project
would
find
out
which
agents
it
is
allowed
to
access?
So
I'm
sort
of
yeah.
E
F
H
E
Right
so
yeah,
I
need
to
draw
this
on,
but
let
me
explain
it
visually
so
when,
when
we
create
a
ci
build,
we
emit
a
whole
bunch
of
variables
right
and
one
of
the
variables
is
acute
config
file
and
the
q
config
file
will
have
two
things.
One
is
the
url.
The
url
is
going
to
be
like
cast
something
agent
or
some
I
don't
know.
E
I
won't
forget
that
ddr
and
then
the
second
thing
is
some
sort
of
password
as
well,
which
is
going
to
be
the
ci
job
token,
which
identifies
the
job
that
it's
running
on.
It
also
identifies
the
project
by
the
way,
which
is
good
so
when
cass
receives
this
is
so
that's
going
to
post
the
cash
when
cass
receives
his
cash
is
going
to
take
the
ci
job
token
and
go
back
to
gitlab
and
ask
hey
is
is
what
what
is
this
for?
E
Is
this
ci
job
token
authorized,
and
so
gitlab
is
going
to
say
that
ci
job
token
belongs
to
this
agent,
and
this
is
where
the
agent
id
is
really
important
and
it
matches
the
it's
the
same
project
of
that
thing.
So
so
it
matches
that
so
it
sends
back
200
back
to
class
and
then
cass
will
connect
to
the
agent
that's
responsible.
E
This
is
where,
hopefully,
we
have
the
right
agent.
This
is
where
we
might
need
the
which
agent
is
connected
thing.
F
C
I
guess
so
in
in
in
this
case
you
will
have
one
like
the
the
project
that
configures
the
agent
is
also
or
agents
actually
is
going
to
be
the
same
project
as
the
one
that
deploys
the
manifests.
Yes
and
it
will
read,
it
has
to
read
the
agent
configuration
file
to
see,
and
it's
just
going
to
see
it.
It's
like
it
needs
to
look
at
the
entries
that
map
for
itself
so
that
project.
C
C
Yeah
we
need
to
look
at
the
configuration
file
from
the
target
project
and
probably
we
don't
want
to
be
reading
from
italy
every
time
we
want
to
make
such
a
query,
because,
like
that's
like
every
time,
we
start
a
ci
job
for
like
that's
using
this.
So
maybe
we
would
cache
it
in
the
database
like
on
commit
or
something.
H
Or
maybe,
if,
when
the
agent,
so
when
the
agent
syncs
the
mana
or
syncs
the
configuration,
maybe
it
can
tell
the
internal
api
what
projects
it
thinks
it
has
access
to.
F
E
H
H
D
Yeah,
I
think
so,
do
we
want
architectural
blueprints?
Would
that
be.
F
D
C
It's
actually
really
similar
in
that
sense
to
the
cluster
management
project,
where
we
needed
to
keep
track
of
the
installed
helm
releases
where
we
had
like
a
git
action
that
had
to
trigger
change
in
the
database.
We
we've
done
most
of
that.
H
D
A
blueprint
document
we
can
clarify
everything
there
so
so,
and
and
just
to
just
be
clear
on
this.
This
architectural
blueprint
is
for
the
is
this
also
for
the
single
project
agent,
or
is
this
for
the
multiple
like
can?
Can
we
build
something
now
the
simplest
possible
nbc
without
an
architectural
blueprint.
F
D
Okay,
so
we
can.
We
can
begin
work
on
the
reverse
tunnel
as
a
bigger
idea.
Right
like
like,
we
might
not
have
a
deliverable
right
away.
F
Which
are
available
by
by
themselves
which
can
use
the
work
that
needs
to
be
done
anyway
for
the
reverse
tunnel.
They
are
agent
information
page.
So
we
can
display
information
about
an
agent,
the
agents
that
are
connected
actually
to
some
cars,
because
we
will
need
to
track
them
for
the
reverse
tunnel
to
know
where
to
send
traffic
to,
and
if
we
track
them,
we
can
show
them
to
the
user,
and
it's
just
an
api
for
forecast
to
expose
and
for
gitlab
to
call-
and
this
is
necessary
for
the
reverse
tunnel
as
well.
D
Okay,
well
that
segues
really
well
into
the
next
issue.
We
need
to
discuss
so
this
next
one
is
management
interfaces
for
the
agent.
D
Maybe
I'm
totally
wrong.
This
is
more
just
read-only
information.
I
guess.
Does
anyone
have
any
other.
D
Based
on
this
issue,
does
anyone
disagree
with
that?
This
seems
like
it's
mostly
like
health
checks,
and
you
know
metrics
and
number
of
sinks,
number
of
connections.
H
E
It's
it's
actually
more
than
management
is
actually
information,
so
this
is
actually
the
actual
feature
that
people
most
people
want
actually
yeah
in.
That
is
what
here's,
what
the
agent
is
doing
for
you,
like
really
crucial
information.
I
would
encourage
us
to
think
of
it
as
more
than
just
one
page
like
it's.
E
It
might
be
a
series
of
pages
as
well
sort
of
like
how
ci
has
like
jobs,
a
page
for
jobs,
a
page
for
pipelines,
a
page
for
job
traces,
that
kind
of
stuff,
like
the
whole
series
of
things,
to
display
information
about
your
your
work
that
you're
doing
on
about
the
work
that
the
agent
is
doing
for
you.
H
Could
potentially
all
wrote
like
the
all
each
individual
thing
could
also
be
rolled
up
into
like
a
health
like
a
health
dashboard,
like
you
know
how
many
agents
are
are
running.
You
know,
unhealthy
would
be
like
not
connected
in
some
period
of
time.
You
could
have
you
know
healthy
clusters
and
unhealthy
clusters.
Clusters
that
are
out
of
resources
like
this
is
kind
of.
This
could
be
a
huge
yeah,
like
tong,
said
kind
of
a
lot
more
than
just
a
page,
yeah.
E
So
then,
those
those
health
checks
are
useful
for
a
certain
persona
and
then
and
then
there's
there's
more
different
personas
at
different
pages.
D
D
H
E
Yeah,
I
think
I
like
us
to
focus
more
if
this
is
just
me
personally
to
focus
more
a
lot
on
the
detail:
information
rather
than
roll
out
information
first
but
yeah.
This
is
my
preference.
H
F
Somebody
already
building
no,
not
yet
but
yeah.
It's
just
an
issue
about
the
idea
to
have
a
an
activity
stream
like
what
you
have
in
your
profile,
something
like
that
where
we
show
all
the
important
events
about
the
agent
now
and
that
might
include
all
those
things
like
deployment
status,
not
status.
But
what
was
happening
at
what
time
and
in
which
well
for
this
agent.
So
it's
not
a
page
or
tab
where
you
go
and
you
see
the
dashboard,
it's
a
stream
of
things,
and
one
doesn't
mean
we
can't
have
the
other.
F
I
think
both
are
useful.
This
activity
stream
is
useful
for
debugging
and
also-
and
this
is
where
it
it's
really
where
we
can.
F
To
our
benefit,
compared
to
other
solutions
where
we
have
sources
of
information
like
different
sources
of
information,
we
can
overlay
this
information
on
the
time
scale
and
show
what
was
happening
and
we
can
show
deploys
commits
like
agent
digitopsyncs,
various
monitoring
events
alerts
stuff
coming
from
celium
like
everything
in
a
single
timeline.
F
I
think
this
is
the
cool
cool
kind
of
outcome,
of
being
a
single
application
for
everything,
and
this
is
what
our
users
hopefully
want,
because
this
is
really
cool.
I
think,
and
can
be
really
useful
to
troubleshoot.
F
I
I
have
used
it
quite
a
bit
yes
and
yeah,
but
we
can
do
even
better
than
that.
I'm.
D
F
F
If
you
are
familiar
with
what
this
is,
there's
a
stream
of
events-
and
you
can
build
views
from
that
stream
and
to
materialize
views-
that's
how
they
are
called.
So
you
can.
G
F
G
F
Yeah
you're
just
viewing
the
details
of
a
particular
agent
okay
for
for
a
particular
project,
which
you
are
looking
at.
Basically,.
H
I
guess
yeah,
I
guess
later
once
we
have
the
piece
where
the
agent
like,
if
we
know
what
agent
is
allowed
to
a
project
like
that's,
not
the
configuration
project,
we
could
also
have
a
corresponding
page
in
the
the
non-configuration
project.
That
shows
probably
a
different
set
of
of
events
from
the
stream,
so
probably
more
focused
on
like
deployments
of
that
of
that
project,
but
they
they
could
end
up
being
very
similar
in
the
ui
like
in
terms
of
what
they
look
like
and
where
this.
H
F
So
actually
this
made
me
think
of
the
the
stuff
that
the
product
or
container
security
team,
what's
the
right
name,
what
they
are
building
with
celium.
This
is
also
an
event
stream
right.
What
what
they
are
doing
so
one
day
when
we
have
this
event
stream,
we
could
just
merge
two
things
together
and
there
event
stream
for
them
becomes
a
filter
on
top
of
our
event
stream.
D
Yeah,
I
think
that
makes
a
lot
of
sense
in
order
to
keep
keep
moving
here.
I
think
we
can,
if
you
have
any
other
thoughts,
please
add
them
to
the
issue
I'll
try
to
I'll
try
to
paste
a
lot
of
this.
These
comments
in
I've
got
the
next
few
items.
D
We
had
a
security
issue
or
a
some
people
that
kind
of
ran
into
our
thing
that
we
had
found
kind
of
known
about
with
terraform
plan
json
like
the
fact
that
we
put
into
an
artifact,
and
so
I've
been
communicating
with
them
a
bit
about
the
work
or
the
the
remediation
and
documentation
that
we
added
a
while
ago.
We're
already
matt
is
already
working
hard
on
fixing
the
actual
root
problem.
D
Here
I
believe
we
might
be
adding
a
little
bit
more
documentation
on
to
make
it
really
clear
that
you
should
not
put
your
aws
credentials
into
your.
You
know
you
should
encrypt
things
basically
and
that
we're
not
responsible
for
that.
D
So
just
fyi
on
that
also,
the
next
release
we're
gonna
be
starting
the
release,
preparation
on
the
15th,
so
oh,
okay,
tong
did
you
wanna.
D
Yeah
yeah,
we
already
have
it
noted,
but
once
we
split
the
docs
into
two
different
pages,
then
the
warning
got
removed
from
the
first
page.
So
then
they
were
like.
Why
isn't
it
on
the
front
page-
and
I
said
well
used
to
be
there,
so
I
added
it
back
to
the
so
that's
duplicated
we
can
put
it
in.
You
know,
add
a
blink
tag
or
whatever
we
want
to
it,
make
it
clear
so.
D
D
So
let's
just
stay
in
communication.
If
you
think
that
your
thing's
not
going
to
get
merged,
then
we
can
halt
on
those
release
posts.
It's
not
a
big
deal,
just
just
communicate
about
where,
where
your
thing
is
at
also
do
the
engagement
survey,
if
you
haven't
already-
or
I
should
say,
feel-
and
you
know
feel
free
to
do
it-
it's
not
it's
optional.
I
think
I
made
it
sound
like
it
was
mandatory.
It's
not
mandatory
but
feel
free.
D
All
right
so
we'll
jump
into
the
agent
breakout,
and
I
think
some
of
this
has
already
been
talked
about
I'll,
let
people
so
the
first
item
an
agent
breakout
is,
is
kaz
enabled
by
default
on
both
helm
and
omnibus.
It
sounds
like
matt
you're
saying
that
it's
it's
underway
or
it's
already.
It's
either
already
emerged
on
both
or
it's
going
to
be
merged
on
both
soon.
D
And
then
the
agent
production
rollout
I
wanted
to
talk
about
that.
I
met
with
anthony
sandoval,
who
is
the
em
for
sre
observability
today,
and
we
had
a
good
chat
about
the
agent
and
so
his
team
would
be
responsible
if
you're.
If
you
look
on
that
kubernetes
agent
to
production
environment,
rollout
issue,
he
said
that
his
team
will
take
responsibility
for
the
top
four
items
there.
The
tracing,
sentry,
login
and
metrics
they'll,
be
you
know,
they're
they're,
happy
to
to
kind
of
configure
all
that
for
us,
the
other
items,
the
rate
limiting.
D
He
said
that
it
would
probably
make
sense
for
whoever
was
going
to
deploy
it
to
do
that.
He
said
that
his
team
would
likely
not
be
the
team
to
deploy
it
because
they
don't
have
the
experience
with
kubernetes.
Actually,
so
he
said
that
graham
is
actually
the
most
experienced
person
either
that
or
the
delivery
team.
D
E
I
think,
graham
graham,
was
just
showing
us
some
pointers
to
the
helm
files,
so
we
can
always
make
a
hammer
to
to
do
that,
but
we
can't
see
the
results.
So
at
least
we
can't
debug
anything.
I
think
we
can
still
see
the
pipeline,
so
I
think
we
will
probably
we
can
make
an
mr
if
you
want
to
speed
things
up
and
then
get
delivery.
One
of
the
delivery
teams
just
kind
of
side
check
us.
D
Yeah,
that's
that's
what
I'm!
If,
if
we're
able
to
do
that
that
that's
probably
what
I
would
propose
with
delivery
is
that
we
can
go
ahead
and
create
the
draft.
Mrs
those,
mrs,
would
be
blocked
in
my
opinion,
by
the
observability
issues.
Here
you
know
we
want
to
make
sure
that
we
have
all
the
observability
stuff
in
place
before
we
merge
anything
so,
but
we
could
go
ahead
and
create
those
draft.
Mrs
now,
from
my
perspective.
F
Do
I
understand
correctly
that
the
observability
stuff
will
be
done
in
staging
first,
because
I
mean
yeah
in
production.
We
don't
have
it
so.
B
D
Yeah
yeah,
that's
what
they'll
definitely
do
anthony
did
say
that
the
staging
environment
is
not
really
on
par
with
production,
there's
not
equal
parity.
So
what
that
means
at
a
technical
level,
I'm
not
totally
sure,
but
he
kind
of
made
it
seem
like
we
can
test
what
we
can
test
in
staging,
but
he
was
also
curious
if
we
could
deploy
cass
in
production
but
not
actually
hook
it
up
like
deploy
it
but
make
it
inaccessible
so
that
they
can
test
it,
but
no
users
could
actually
hit
it.
F
G
D
Okay,
well,
I
can
follow
up
with
I
kind
of
feel
like
we're,
going
to
be
getting
more
involved
with
different
sre
teams
and
and
I'm
going
to
be
kind
of
vetting
it
out,
because
they
sound
really
busy
the
observability
team
sounds
really
busy.
I
don't
know
about
delivery,
they're,
probably
really
busy.
Also,
so
you
know,
I
will
influence
what
I
can,
but
we
have
to
be.
D
F
E
Yeah
I
was
wondering
if
we
should
open
issues
and
that's
probably
the
best
way
to
start
for
deliberating-
maybe
maybe
we're
missing
things
just
in
case.
D
Yeah
we
yeah
it's
possible,
we're
missing
things
that
we
need
so
the
for
the
top
four
items
anthony
said
that
he
would
open
a
infrastructure
issue
for
those,
but
for
the
the
deploy
we
could
just
open
an
infrastructure
issue,
which
is
what
I'm
I'm
guessing
that
deliver.
The
delivery
team
uses
the
infrastructure
project
or
do
they
use
gitlab.
E
E
No,
but
I've
got
a
fyi
about
the
release
tools,
so
I'm
working
on
getting
cast
and
omnibus
automatically
getting
the
correct
version,
so
it's
mostly
done,
but
it's
not
activated
yet
I'm
just
waiting
on
mikael's
update
to
get
us
past.
This
braking
change.
E
F
H
D
Great,
does
anyone
have
anything.
D
D
All
right:
well,
thanks:
everyone
appreciate
you
taking
the
time
great
conversation
and
have
a
good
rest
of
your
day.