►
From YouTube: 2021-06-12 Delivery team weekly EMEA/AMER
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
Awesome
so
we're
out
time-
and
also
I
see
I
didn't
put
mtgp
in
the
agenda,
but
I
do
want
to
highlight
it
because
it's
looking
pretty.
A
A
So
mttp
at
the
moment
for
july
is
12.5,
so
huge,
huge
shift,
which
is
awesome
main
thing
here-
is
we've
tweaked
the
branch
times
a
little
bit
so
they're
fitting,
certainly
for
my
side,
they're
fitting
really
nicely
on
my
day,
but
also
graham,
has
started
the
playing
as
well,
which
is
making
a
significant
difference.
So
that's
pretty
awesome
to
see
so
what
I
think
we'll
do.
A
I
made
a
great
suggestion
last
week:
if
we
get
a
couple
of
weeks
around
this
target,
then
we'll
drop
the
target
back
to
12
hours
and
continue
on
from
there.
So
that
is
big
news.
One
additional
thing
I
learned
last
week
around
mttp
is
on
this
chart.
So
this
is
the
one
that
links
out
of
the
handbook.
A
We
actually
only
track
deployments
on
days
that
have
a
deployment,
so
it's
excluding
all
weekends
naturally,
which
I
think
is
a
really
good
thing
and
it
will
exclude
automatically
exclude
family
and
friends,
and
things
like
that.
It
does
mean
there's
a
side
side
effect.
If
we
have
a
really
big
incident
and
we
don't
deploy
anything,
it
will
also
exclude
that
day,
but
that's
a
reasonably
like
rare
event,
so
I
think
it's
kind
of
fine,
so
yeah.
This
is
tracking
mttp
at
hours
for
days
that
have
at
least
one
deployment.
A
Awesome
so
loads
of
announcements
I
just
was
through
shout
if
there's
anything
that
you
want
to
highlight
so
henry
is
out
now
until
the
26th
seo
about
to
go
out
as
well.
This
friday
is
family
and
friends
day.
So,
if
you
haven't
got
that
planned
off,
consider
taking
that
one
last
week
you
will
have
had
a
number
of
alerts.
I'm
sure
we've
got
we're
doing
an
updated
check-in
on
the
engagement
survey,
so
ruby
you
most
likely
didn't
get
these.
These
just
were
the
infra
department.
A
So
that's
fine
on
that
one,
but
for
everyone
else.
If
you
haven't
done
that
yet
please
do
and
most
excitingly
on
wednesday.
We
have
an
ama.
A
So
if
you
have
any
themes
you'd
like
to
go
through,
then
let
me
know
otherwise
I'll
just
announce
it
tomorrow
as
a
kind
of
a
regular
one,
and
we
can
just
open
it
up
see
if
it's
anything
we
want
to
seed
it
with
or
if
there's
any
questions
we
want
to
ask
each
other.
A
So
discussion
things
are
in
progress,
so
thank
you
for
the
comments
of
so
far
on
this
one.
A
If
there
is
anything
you
think
we
need
to
additionally
consider,
please
feel
free
to
drop
it
in,
but
one
thing
I
was
thinking
around,
so
it
looks
like
we're
sort
of
moving
towards
having
a
single
okr,
which
will
be
most
likely
something
around
hardening
the
kubernetes
cluster,
so
we're
quite
close
we're
in
the
progress
of
having
the
web
migrated,
and
that
will
mean
that
pretty
much
all
of
our
stateless
services
are
there
pages
is
the
only
other
one
that
that
will
come
soon,
but
one
thing
I
wanted
to
think
about
as
we
as
we
sort
of
move
into
this.
A
What
comes
next
is
what
what
do
we
need
to
change?
So
what
does?
Having
pretty
much
our
whole,
like
everything,
we're
deploying
so
we'll
be
doing
like
all
these
sort
of
jobs
that
we
have
at
the
moment
around
the
web
fleet
will
be
in
kubernetes.
A
B
B
And
this
has
been
a
long
conversation
around
this
with
distribution
as
well.
The
problem
here
is
charts,
so
charts
are
not
designed
for
this
type
of
things
and
when,
when
we
started
this
conversation
we
said
yeah.
If
you
are
actively
looking
into
an
operator,
we
we
would
need.
We
would
like
to
have
some
kind
of
hook
into
the
operator
to
where
we
can
instrument
things
like
that
right,
so
some
specific
checks
or
adjustability
to
how
to
stop
a
deployment
and
say
yeah.
A
Yeah
yeah
yeah,
the
operator
stuff
is,
is
progressing
probably
end
of
the
year,
so
yeah
we're
a
little
way
out
for
that
one:
okay,
yeah!
That's
an
interesting!
That's
a
good
point.
A
A
Okay,
I
will
so
I
follow
up,
because
one
thing
I
want
to
make
sure
we
are
sort
of
thinking
about
going
forwards
is,
as
as
we
make
decisions
on
kind
of
the
kubernetes
cluster
set
up,
we're
tying
these
into
deployment
things
like,
I
think
in
the
past
it
was
slightly
less
critical
because
it
wasn't
they
weren't,
basically
totally
the
same,
but
as
we
go
forwards
now,
I
know
there's
conversations
around
how
we
apply
helm,
changes
and
helm
does
auto
rollbacks,
which
we
need
to
like
need
to
be
careful
of
like
that,
could
have
unexpected
things
around
how
we
do
deployments
so
I'll
open
up
a
there.
B
But
we
are
using
helm
with
weight
so
from
coordination
perspective
so
from
the
from
the
deployer's
side,
either
it
applied
or
it's
or
it's
reverted,
because
that's
the
point
right.
So
when
you
are
operating
by
default,
you
can
say
just
apply
this
and
if
you
don't
ask
him
to
wait,
you
would
just
say
yeah.
This
is
the
new
manifesto
I
published
it
and
then
eventually
the
cluster
will
come
to
the
new
state
or
roll
back.
But
we
are
not
using
this
feature.
B
We
wait,
which
means
that
either
it's
applied
or
it's
rolled
back,
but
we
know
this
because
we
are
waiting
for
the
pipe.
We
are
checking
it
for
the
pipeline
from
the
deployer.
B
Now
you
better,
I
was
asking
if
you
had
some
specific
use
case
in
might
so
in
that
sense
it
sounds
like
it's
better
than
than
virtual
machines,
because,
let's
say,
let's
think
about
some
incident,
that
we
had
during
last
release
management
ship.
So
we
had
machinery
were
not
able
to
start
not
able
to
install
right.
So
in
a
virtual
machine
world
we
end
up
with
a
failed
deployment
and
machine.
They
are
not
actually
running
because
they
are
outside
of
the
load
balancer
and
they
are
broken.
B
While
here,
I'm
not
entirely
sure
about
how
we
handle
the
load
balancer-
and
I
see
no
one
from
the
sre
team
of
our
team
here
to
just
help
me
on
this.
But
basically,
let's
assume
that
we
are
doing
the
kubernetes
magic
here.
So
we
do
health
check
on
pod
level
and
things
like
that.
So
in
that
sense
the
the
deployment
fail
because
we
are
not
able
to
upgrade,
but
the
cluster
rolls
back
to
the
previous
known
state.
So
we
are
not
losing
capacity
on
the
fleet
yeah
it
sounds.
B
A
Yeah,
it
should
always
be
in
one
of
two
states
right:
it's
not
just
out
of
rotation,
so
cool!
Okay,
that's
helpful!
Thank
you.
I
will
start
pulling
this
stuff
together
because
yeah
once
we,
the
kind
of
the
future
of
deployments,
is
we'll
be
able
to
use
more
kubernetes
stuff.
So
we
should
sort
of
like
start
thinking
about
what
what
do
we
want
that
to
look
like
like
how
do
we?
How
do
we
pull
that
stuff
together
awesome?
A
So
the
other
piece
I
had-
and
this
is
more
of
a
awareness
thing,
but
I
put
it
in
discussion
just
in
case
people
have
comments,
is
so
registry
is
getting
close
to
being
ready
for
deployment
and
then
the
migration's
starting.
A
So
I
shared
here
the
timeline
they're
going
through
so
henry
has
been
drying
on
this,
but
with
him
being
on
pto
and
then
release
managing
that
will
change
a
little
bit
so
just
as
awareness
it
will
most
likely
affect
graham
and
scarbeck,
but
just
as
awareness
like
some
things
may
re-prioritize
around
this,
so
we
can
support
package
through
to
through
their
migration.
B
Today
I
learned
that
we
can
do
gun,
chart
and
merge
request
descriptions
in
issues
and
epic,
never.
A
I'll
be
expecting
it
from
all
of
you
on
all
your
projects.
I
used
a
gantt
chart
once
on
one
of
my
projects
just
for
fun,
and
it
was
like
the
most
disastrous
project
I've
ever
been
involved
in.
We
basically
had
to
revert
the
entire
thing,
because
that's
it
I
was
like
I'm
never
going
to
use
one
again.
It
clearly
like
it
was
just
too
fixated
to
curse
the
whole
thing.