►
From YouTube: 2021-04-19 Delivery team weekly
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
B
C
I
think
it's
the
capture
card
kicking
in
because
it's
it's
plugged
in
by
default,
but
I
think
like
auto
suspense
similar
to
like
any
usb
device
or
maybe
just
take
a
second
or
two
for
the
gopro
input
to
be
sent.
Initially,
I'm
not
sure.
A
A
Did
you
used
to
have
a
thing
when
I
was
young?
We
didn't
have
enough
tv
programs,
it
wasn't
like
24,
7
tv
scheduling
and
you'd
have
like
a
holding
page
where
they
just
put
it
up
for
like
some
hours
in
the
day,
and
I
don't
want
to
sit
and
watch
it
like
convinced
that,
like
sooner
or
later,
like
an
amazing
program
was
going
to
happen
in
probably
like
five
hours
like
it's
any
minute.
It's
gonna
be
great.
B
Yeah
and
it's
very
similar
to
the
thing
that
we
see
when
you're
opens
his
camera.
A
B
At
least
here
in
italy
was
in
in
a
square,
so
you
have
a
square
with
the
thing
inside
and
then
everything
outside
was
black.
So
you
had
this
kind
of
signal.
A
In
the
day,
in
the
olden
days
when
we
were
young,
welcome
everybody,
so
robert
is
out
getting
his
vaccination,
so
hopefully
he'll
be
back
with
us
later
on
and
is
that
all
of
us
feel
like
yeah
we're
all
here?
Otherwise,
so
let's
kick
things
off
mttp.
I
have
shared
it
up
there.
So
it's
looking
super
healthy,
which
is
great.
So
what
on
yorick?
Why
don't
robert
I'm
going
to
leave
the
target
as
it
is
for
this
week?
A
We
said
originally
that
when
we
had
two
weeks
of
incident
free
time,
we'd
review
this,
so
I
think,
based
on
the
interruptions
we
saw
beginning
of
last
week
from
the
security
vulnerability,
we'll
leave
the
target,
as
is
through
this
week
and
review.
Hopefully
it'll
be
a
super
quiet
week
and
we
can
go
back
to
our
regular
12-hour
target
from
next
week.
A
So
on
the
announcements,
so
I
just
put
in
the
engineering
q2
okr
issues
super
helpful
that
they're
all
just
in
one
place.
That's
probably
the
first
time
I've
just
seen
every
department
listed
out
there.
A
They
are
all
working
progress
for
the
next
couple
of
weeks,
so
just
be
aware,
but
still
particularly
infrastructures
they're,
definitely
not
finalized.
There's
definitely
draft
but
useful
to
have
some
visibility.
A
Pg
upgrade
was
postponed
till
may
8th.
So
the
main
thing
I
think
just
to
watch
for
on
this
one
is
firstly
like
thank
you
scobeck
henry,
for
the
help
you
gave
last
week.
We
may
be
asked
to
be
involved
in
other
things,
leading
up
to
this.
If
it
looks
like
we
are
involved
more
than
just
a
tiny
bit
we'll
review
the
release
management
schedule,
as
we
have
both
of
you
too,
in
the
next
rotor.
So
just
as
a
heads
up,
everyone
like.
A
Slightly
related
to
that,
I
think
skybeck
taking
defensive
family
day
this
friday,
so
thank
you
for
robert
for
covering
and
robert
is
hopefully
not
too
taken
out,
but
potentially
taken
out
later
this
week
with
vaccine
cover
super
thanks,
garbek
for
covering
that
stuff.
A
So
the
discussion
point
I
had
for
today
was
I'm
thinking
around
oh
chaos.
So
thanks
for
the
comments
so
far
on
the
proposed
q2
okr
issue,
one
question
I
had
really
was:
how
useful
is
it
to
make
deployment
slo
as
an
okr?
So
we
don't
have
to
have
it
as
an
okr.
We
do
need
to
do
some
work
on
it,
but
it's
really
so
let
me
go
through
kind
of
the
pros
and
cons
of
something
being
an
okr
on
the
plus
side.
A
It's
easy
to
prioritize
because
we
do
okrs
and
then
we
do
other
things
so
really,
in
that
case,
it'll
be
very
straightforward
for
us
to
decide
like.
Should
we
be
doing
this
slo
work,
or
should
we
be
doing
something
on
rollbacks
or
like
coordinate
deployments
or
something
like
that
or
other
right?
A
The
downsides
deployment
slo
may
not
be
a
full
quarter
of
work,
which
kind
of
is
a
question
mark.
I
think
I
don't
want
to
expand
it
unnecessarily
to
fill
a
quarter
just
because
it
is
an
okr
like
if
we
can
achieve
something
good
in
three
or
four
weeks
perfect.
That
would
be
the
best
outcome
and
then
probably
the
biggest
question
mark
around
something
being
an
okr
or
not.
Is
everybody
in
the
department
is
pretty
slammed
right
now
you
may
have
all
noticed
they're
doing
a
little
bit
too
much
so
going
into
this
quarter.
A
D
May
I
ask
for
the
specific
meaning
of
this
slo
because
I'm
always
getting
con
confused
by
okrs
with
objectives
and
key
results
right
so
with
the
objective
be
that
we
hit
a
certain
slo
all
the
time.
So
that
would
be
the
goal
and
the
key
result
would
be
to.
I
don't
know,
set
up
slis
to
be
able
to
measure
this,
or
is
it
meant
in
a
different
way.
A
So
the
very
first
part
would
be
to
get
slo
fully
defined
and
in
the
handbook,
with
a
target
that
we
can
be
tracking
against
in
order
to
have
a
target
we'll
need
to
be
able
to
measure
to
some
degree,
because
what
ideally
we'd
do
is
say
like
over
the
last
three
months.
It
has
been
this
number
and
we
can
set
a
sensible
target
based
on
that,
but
that
would
be.
The
very
first
part
would
just
be
the
definition
and
a
target.
A
Not
less
well
yeah,
maybe
a
way
to
record
them
yeah,
so
I
think
the
objective
would
be
to
have
a
deployment
slo.
The
key
results
would
be
there's
something
in
the
handbook.
There
is
a
target
that
we
can
work
towards
and
yeah
they
will
be
visible
somewhere.
So
I
think
the
step
after
defining
this
deployment
slo
will
be.
A
D
Yeah,
so
I
mean:
if
we
set
some
target,
then
we
would
be
tracking
this
over
the
full
quarter
for
sure
if
we
hit
that
right
or
if
he
getting
above
it,
so
it's
something
which
easily
spreads
over
the
full
quarter.
It
just
depends
on
how
good
we
are
hitting
this
target
and
how
much
work
we
need
to
spend
on
it
right
to
try
to
stay
within
this,
but
the
objective
would
fully
would
be
meant
for
the
full
quarter.
I
think
right
to
try
to
reach
this
objective
of
a
certain
seo
right.
A
Well,
I
mean
it
depends
on
how
we
how
we
want
to
set
the
target
up.
So
I
think
from
what
I've
seen
with
the
way
mttp
has
worked.
Is
we
sort
of
set
a
a
sort
of
a
slightly
stretched
target,
but
something
which
we
can
hit
and
then,
when
we
want
to
lower
the
target
it
kind
of
comes
with
a
a
big
change?
So,
for
example,
we
are
currently
targeting
12
hour
mttp
and
that's
very
much
linked
to
rollbacks.
A
We
won't
go
beyond
12
hours
until
we
also
have
rollbacks
and
then
we'll
use
deployment
slo
in
a
similar
way.
So
say
our
target
is
well.
I
don't
know
the
answer.
Roughness
would
be
like
seven
hours.
Maybe
we
would
sit
around
that,
but
then
at
some
point
maybe
we
say:
okay,
actually,
we
want
to
lower
this
target.
How
do
we
get
this
to
five
hours
and
we'd
have
to
look
through
and
see?
Okay,
we
could,
I
don't
know,
change
the
pipeline
or
do
something
else
to
actually
lower
this.
D
B
Yeah
regarding
metrics,
I
think
that
there's
this
andrews
proposal,
which
is
really
nice,
which
is
to
frame
our
as
low
as
the
update
sle,
so
it's
kind
of
comparing
let's
I
I'm
rephrasing
one
of
his
examples
because
he's
using
commits,
I
think
merged
request
is
more
suitable
in
our
case.
B
So,
for
instance,
something
like
a
number
of
master
merge
requests
deployed
to
production
within
one
day
compared
to
total
number
of
master,
to
merge,
request
managers
to
master
in
the
same
amount
of
time
and
so
and
then
he
was
proposing
to
because
this.
This
doesn't
give
us
insight
into
the
details
of
where
things
are
just
going
south.
B
So
having
this
as
an
external
bucket
as
external
boundaries
of
our
of
the
things
we
are
measuring,
then
we
can
do
things
like
a
number
of
merge
requests
that
we
tagged
within
x,
amount
of
time
compared
with
number
of
mergers
merged
and
then
things
that
completed
qa
within
this
amount
of
time
things
things
always
being
merge,
requests,
so
merge,
requests
that
completed
qa
within
four
hours
compared
to
merged
request,
merged
to
master
and
then
again,
merger
requests
deployed
to
staging
within
x
time
compared
again
always
same
thing,
merge
to
master
and
then
cannery.
B
So
we
have
the
big
thing,
which
is
we
reached
out
to
production,
and
then
we
all
have
different
sle
for
everything
that
happens
before.
So
we
can
see
what
is
trending,
which
one
of
these
is
trending
and
it's
running
outside
of
basically
the
boundaries
of
the
control,
and
we
can
see
where
we
need
help.
D
But
right
now
we
are
talking
about
deployments
right,
not
of
bringing
a
certain
mr
into
production
right
so
yeah,
but
it's
about
the
type
of
deployment.
I
thought
what
we
are
tracking
here.
B
D
It's
based
on
the
merge
request,
yeah,
so
the
small
mttp,
what
he
is
mentioning
there
right
in
time
to
production
from
from
a
merge
request,
but
this
slo
seems
to
be
about
yeah
you're
right
length
of
our
deployment
pipeline
right.
So
it's
a
different
thing.
I
think,
and
what
I
proposed
in
a
comment
before,
that
is
at
least
an
idea,
because
we
have
this
in
prometheus.
D
We
have
this
thing,
which
is
called
job
matrix,
which
is
a
fairly
easy
way
to
send
signals
metrics
to
prometheus,
with
certain
tagging,
to
signal
the
start
and
the
end
of
a
certain
job,
and
also
if
it
was
successful
or
was
a
failure.
We
used
us
for
database
backups,
for
instance,
and
so
we
can
track
via
this
kind
of
metric
when
something
started
when
it
stopped.
D
If
it
failed
and
using
these
metrics,
we
could
measure,
for
instance,
how
long
does
it
take
on
average,
maybe
just
an
idea
of
how
we
could
track
this
by
just
having
a
very
simple
call
to
prometheus
at
the
start
on
at
the
end
of
a
pipeline,
for
instance,
that
would
be
an
idea
how
we
could
maybe
come
up
with
metrics
for
this
one.
If
we
are
talking
about
the
length
of
a
delivery
pipeline
right,
I
need
to
go
to
the
door
for
a
few
seconds.
One
moment.
A
Yeah,
it's
so
robert
had
an
interesting
graph
from
one
of
his
mrs
last
week,
which
he
just
compiled
locally,
which
I
find
a
sec.
But
what
was
interesting
about
it
was
that
it
wasn't
just
the
the
current
number,
which
was
interesting,
but
the
trend
line
was
fascinating.
I
definitely
want
to
make
sure
as
we
go
through
either
I
don't
mind
if
it's
the
kind
of
the
start
or
after
but
somehow
related
to
deployment
slo,
we
need
to
be
able
to
say
like
what's
the
state.
A
Is
this
just
like
a
one-off
spike
that
we
dig
into
or
is
there
a
trend?
That's
going
up
that
you
know
we'll
want
to
dig
into.
B
Yeah,
maybe
building
on
top
of
what
henry
pointed
out
instead
of
counting,
because
I
I
was
so
taken
into
the
sle
idea
that
I
didn't
check
exactly
what
we
were
accounting.
So
this
was
exactly
a
good
point,
but
maybe
the
events
could
be
a
number
of
deployments,
so
it
could
be
if
we
say
that
we
would
like
to
have
deployment
to
production
that
takes
two
hours.
Let's
say
two
hours
is
our
goal.
Then
we
can
do
sle,
comparing
deployment
that
took
less
than
two
hours
compared
to
total
deployment.
A
Yeah,
I
think
I
think
it's
definitely
useful.
I
think
probably
it
comes
back
to
what
we
maybe
touched
on
a
few
weeks
ago,
which
is
the
what
questions
do
you
want
to
be
able
to
answer
with
this,
but
I'll
share.
A
In
so
robert
had
generated
this
locally,
and
it's
being
able
to
see
have
this
visibility
for
each
like
stage
of
the
pipeline
would
be
fascinating.
A
But
yeah
the
goal,
I
think
the
longer
term
goals.
So
this
is
the
piece
that's
in
our
control,
which
is
the
pipeline,
but
then
following
this
once
we've
got
an
approach
and
we
know
how
to
sort
of
validate
this
stuff,
we'll
add
on
the
building
and
also
the
qa.
So
this
bit
right
now
shouldn't
really
take
into
account
the
qa
stuff,
or
certainly
it's
not
a
key
piece.
A
That's
within
our
control,
so
we
should
make
sure
and
measure
kind
of
we
can
separate
that
because
the
goal
later
on
is
we'll
set
target
different
targets
for
those
three,
and
that
gives
us
some
sort
of
sla
that
we
can
work
with
quality
and
say:
like
hey,
you
know
tests
are
taking
longer
than
they
were
two
months
ago.
We've
got
this
agreement
and
same
with.
A
Distribution
so
clearly
there's
a
lot
of
I
like.
I
agree.
I,
like
the
the
sort
of
updex
approach.
We
probably
have
some
questions
around
like
what
do
we
measure?
How
do
we
measure
those.
A
How
would
you
feel
about
deployment
slo
not
being
an
okr,
and
then
this
would
mean
that
on
the
board,
we
will
quite
likely
have
say
like
ideal
case,
we'll
have
like
one
or
two
tasks
that
are
deployment.
Slo
we'll
have
one
or
two
tasks
that
are
coordinate
deployments.
We
maybe
have
one
or
two
roll
backs
one
or
two
change
logs
I'll,
ignore
kubernetes
for
now,
because
that
definitely
is
going
to
be
an
okr,
but
all
of
these
others
would
sit
at
p3.
A
Does
that
feel
like
it
would
be
useful,
like
you'll,
come
to
the
board
and
pick
up
a
task
like?
Is
that
going
to
be
too
overwhelming?
Do
we
want
some
additional
priorities
in
there.
B
Works,
I
am
so
what
I'm
a
bit
concerned
about
the
department
strategies,
I'm
not
department
with
the
yeah,
whatever
so
the
strategies
that
we
have,
because
they
state
that
we
are
going
to
define
those
slo
for
by
the
end
of
q2.
B
B
A
Have
we
won't
have
it
by
the
end
of
the
quarter?
If
we
don't
make
it
an
okr.
B
What
I'm
thinking
is
that
everything
else
will
be
more
important.
We
have
a
list
of
things
and
priorities,
and
I
mean
ocr
should
be
what
defines
our
goal
and
our
targets,
and
I'm
just
thinking
about
what
happened
this
quarter
right.
Then,
registry
registry
was
not
the
register.
Migration
was
not
expected
and
it
came
out
and
even
if
we
had
okrs,
that
thing
was
more
important.
That
was
a
business
case
was
made,
and
so
and
so-
and
so
I'm
fine
with
this.
B
But
what
I'm
saying
here
is
that,
if
deployment
slo
is
not
even
an
okr
having
this
conversation,
next
quarter
will
be
even
harder.
B
Yeah,
maybe
even
if
we
are,
we
agree
that
it
will
not
take
a
full
quarter,
because
maybe
we
can
do
it
and
then
we
just
say
adjusting
numbers
or
just
checking
our
ability
to
actually
process
and
understand
those
numbers.
It
still
gives
this
more
importance
because
yeah
this
is
nokia
for
the
team.
Something
important
is
it
aligns
with
our
strategy
and
yeah.
So
this
is
my
idea.
D
D
D
A
Yeah,
absolutely
it's
definitely
difficult
when
we
have
other
projects
coming
in
often
we
have
some
room
to
prioritize
them,
but
but
certainly
not
always,
so
that
is
the
challenge.
Okay.
So
if
we
do
make
this
an
okr,
I
think
it's
fine
for
it
not
to
be
a
stretch
goal
like
they.
They
usually
are,
but
I
think
it's
fine
for
it
not
to
be,
but
I
would
very
much
still
like
it
to
be
as
small
as
we
can
make
it.
A
So
it's
because
we
do
have
lots
of
other
things
going
on
so
for
this
one,
we'll
just
state
that,
like
we'll,
keep
it
small
and
then
that
will
place
a
higher
priority
than
backs
coordinate
deployments
and
chain
shock.
A
Yep
everyone
happy
with
that
super
okay,
fantastic!
Thank
you
for
that.
I
will
go
ahead
and
get
the
okr
issues
open.
The
kubernetes
one
quite
straightforward,
we're
getting
very
close
to
the
end
of
the
stateless
migration,
but
I
won't
try
to
discover
it
last
week.
We
won't
use
the
words
like
finish
it,
but
we
will
have
like
continue
and
have
all
the
work
listed
in
there.
A
It
is
going
to
be
more
than
a
quarter,
but
it's
probably
not
like
infinitely
more
than
a
quarter,
so
that's
totally
fine
and
our
second
okr
will
be
a
deployment.
Slo
goal
objective
will
be
to
have
a
deployment
slo
in
place
which
meets
our
strategy
view,
so
we
can
track
against
this,
but
not
necessarily.
A
We
won't
necessarily
change
it.
We'll
just
set
a
target
and
start
measuring
against
that
fantastic.
Thank
you
for
that.
Is
there
anything
anyone
else
wants
to
discuss.