►
From YouTube: 2022-02-16 GitLab.com k8s migration 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
Let's
give
oh
wait,
not
eight
for
that
time,
I'm
gonna
say
I'll,
give
a
few
more
minutes,
but,
like
I'm
super
early
today,
so
they
definitely
give
a
few
minutes
for
everyone
else
to
actually
arrive.
A
All
right,
hey
everyone,
so
welcome
16th
of
february,
and
this
is
our
kubernetes
demo.
So
is
there
anything
that
anyone
would
like
to
demo
today.
A
Nope,
okay,
so
rehab
you
have
the
discussion
item.
B
Hi
thanks
for
having
me
in
the
meeting
today
thanks
for
joining
us,
my
questions
are
all
about
the
preparation
for
the
1.22
upgrade.
B
B
So
if
you
look
at
epic
688,
which
I
think
in
the
agenda,
you
can
see
there's
three
child
issues.
I
think
that's
the
correct
epic,
if
I'm
not
mistaken,.
B
Yes,
so
there's
three
child
issues
to
that:
epic,
showing
three
repositories,
and
my
understanding
is
that
we're
gonna
be
looking
at
these
repositories,
mainly
for
this
epic,
and
if,
if
we
do
have
the
capacity
after
that,
we're
gonna
look
at
other
repositories.
B
Is
this
correct?
Is
this
wrong?
Are
there
any
other
inputs
that
we
will
need
to
look
at
as
part
of
this
change.
C
None
that
I
know
of,
I
think
the
one
thing
that
excuse
me,
the
item-
that's
going
to
make
this
a
little
bit
more
difficult,
is
that
we
do
import
some
helm
charts
from
elsewhere.
So
the
prometheus
operator
is
a
good
example
of
this,
and
there
may
be
deprecations
that
we
may
need
to
force
us
to
upgrade
that
particular
helm
chart.
C
So
looking
at
the
results
of
the
objects
that
we
deploy
and
where
those
are
coming
from,
it's
probably
going
to
take
the
most
time,
I
think
for
any
item
where
we
construct
the
chart
ourselves,
such
as
the
gitlab
home
chart.
Those
will
be
easy
to
look
at
and
then
for
some
of
these
items,
both
the
tonka
and
the
helm
files
repose.
We
do
have
like
a.
C
What
do
you
call
it
like?
A
bare
bones,
helm
chart
that
we
just
kind
of
fill
in
the
blanks
and
we
make
it
deploy
objects,
so
there
may
be
something
that
is
easier
to
do.
That
would
be
inside
of
that
own
repository
itself,
but
these
are
the
correct
three
repos
that
you
would
want
to
look
at.
There's
no
more
that
I
know
of.
B
Okay,
so
the
other
question
I
have
is
so,
if
you
mentioned
that
we're
importing
other
hunt
charts
from
outside
of
these
repos,
so
is
that
what
graham
referred
to
as
getting
the
manifests
for
these
wrappers?
C
Yeah,
like
a
helm,
template
if
you're
using
the
gitlab
home
files,
for
example,
will
provide
you
the
necessary
output
of
what
we
intend
to
deploy,
for
example,
so
they're
relatively
easy
to
look
at.
B
Okay,
thank
you
for
that
and
another
question
between
these
three
repos,
which
one
do
you
think
is
going
to
take
most
of
the
time.
If
that's
even
like
a
valid
question,
I'm
just
like
curious
to
sort
of
rank
them
the
most
difficult,
the
easiest
and
the
medium.
A
What's
the
kind
of
I
guess
like
what's
the
goal
of
knowing
that.
A
B
Just
thinking,
maybe
other
people
would
prefer
something
easier
depending
on
on
what
they
want
to
like
the
level
of
effort,
maybe
or
maybe
their
availability
like,
depending
on
everyone's
availability.
They
would
take
the
easier
or
the
hardest
one
so
that
that
was
the
main
goal
like
wanting
to
distribute
these
issues
between
team
members
and
sort
of
assigning
weights
to
them,
so
so
that
we
know
like,
what's
the
time
we're
going
to
spend
on
their
needs,.
A
It
could
be
worth
having
a
look
at
them
or
from
a
like,
which
ones
absolutely
have
to
have
happen.
Like
a
kind
of
criticality
like
I
would
probably
do
tanker
loss
just
in
case
we
get
stuck
on
the
other
things.
B
Yeah,
I
think
that
answers
my
question.
Is
there
any
advice
you
want
to
give
me
for
this
upgrade
initially
we
thought
that
we're
gonna
use
the
epic
for
the
quarters
work,
but
now,
after
like
seeing
the
amount
of
work
that
we'll
be
doing,
I
don't
think
it's
going
to
take
a
full
quarters
work
at
all.
B
D
D
I
I'm
not
really
an
expert
in
this,
but
I
saw
that
dream
a
few
months
ago,
use
some
tooling,
which
is
automatically
checking
for
duplicated
communities
versions,
not
sure
if
we
could
maybe
use
this
again
or
automate
somewhere
around
this,
because
this
will
come
up
again
right
every
time
we
need
to
upgrade
to
a
new
minor
version.
A
Well,
this
one's
interesting,
because
it's
disruptive
so
they're,
not
all
disruptive
but
yeah-
it
could
be
worth
reviewing.
Let's
take
that
as
a
separate
discussion,
because
I
think
at
that
point
we're
kind
of
it's
a
slightly
different
project
and
it
might
be
worth
us
reviewing
at
that
point
like
what
other
projects
are
we
comparing
it
against.
C
Google
has
also
been
very
good
at
sending
us
notifications
ahead
of
time,
saying:
hey
you're,
using
deprecated
apis;
here's
how
to
figure
that
out,
here's
the
projects
that
are
impacted,
etc.
So,
I'm
not
saying
we
rely
on
those
emails,
I'm
just
saying
that
we
at
least
do
get
a
heads
up
that
stuff
like
this
is
occurring.
D
A
C
And
I'm
pretty
sure
one
of
these
epics
that
rahab
linked
to
actually
contains
yeah.
D
B
Yeah,
so
this
this
was
an
issue
and
I
promoted
it
into
an
epic.
So
it
closed
the
issue
automatically
and
the
email
is
in
the
body
of
the
issues
with
the
original
email.
B
I
should
also
update
the
acceptance
criteria
because
initially
we
had
some
understanding
now
we
have
a
different
understanding,
so
that's
not
like
100
up
to
date.
Just
yet.
I
need
to
update
that.
A
A
Okay
and
feel
free
to
like,
if
you
want
to
do
async
as
well
like
feel
free
to
add
them
into
your
gender
and
graham,
usually
adds
comments.
So
he
may
come
back.
B
Cool
just
to
say
about
addressing
the
automation
I
I
actually
did
think
about
this
in
the
beginning,
but
I
I
I
think
it
will
be
a
separate
project.
I
mean
we're
still
going
to
discuss
with
with
the
managers
what
we're
going
to
do
after
closing
this
epic.
So
maybe
we
I
will
propose
that
we
we
attempt
to
automate
this
now
that
we
know
what's
the
effort,
but
like
no
promise,
we
still
don't
know
the
operatives.
C
B
Yeah
I
mean
in
a
different
world
so
like
we
could
create
some
sort
of
an
ai
that
would
go
through
the
change
login
like
can
be
deprecated
api's
and
then
determined
through
our
ripples
and
the
manifest
just
do
this
automatically.
I
actually
didn't
think
about
that.
I
thought
more
about
like
a
pipeline.
Maybe
that
would
attempt
to
list
the
resources
that
are
deprecated,
at
least
that
one
step,
and
then
we
can
do
the
rest
manually,
which
is
upgrading
how
about?
If
we
that's
just
one
random
question.
B
So
if
we,
if
there's
a
way
to
see
what
resources
are
being
changed
by
kubernetes
when
they're
actually
applied,
because
I
know
that
kubernetes
would
automatically
readjust
the
version
if
it's
deprecated
right.
D
C
And,
depending
on
how
flexible
a
helm
chart
is
what
they
will
attempt
to
do
is
modify
the
version
that
is
being
deployed
of
a
specific
object,
depending
on
the
version
of
kubernetes
that
you
are
running.
So
this
caused
a
blocker
for
auto
deploys
one
time,
because
the
nginx
ingress
controller
api
version
changed
between
clusters
and
in
between
auto
deploys.
C
That's
going
to
be
one
of
those
ones,
that's
a
little
bit
challenging,
but
it's
worth
having
the
ability
to
detect.
Excuse
me
the
ability
to
detect
that
ahead
of
time,
and
that's
also
going
to
require
a
little
bit
of
work
inside
of
autodeploy
to
help
try
to
prevent
something
like
that
from
occurring,
which
we've
got
an
open
issue
to
figure
out
how
to
address
that.
B
C
It's
a
little
sluggish
we're
at
a
step
now.
The
last
thing
I
know
that's
been
accomplished
is
igor
has
completed
the
6.2
upgrade
on
all
of
our
virtual
machine
infrastructure.
C
C
It's
recommended
by
redis
that
we
do
not
mix
configurations
the
help
chart
forces
you
to
use
host
names,
so
we're
going
to
try
to
switch
over
virtual
machines
and
do
the
same
I'm
running
into
a
problem
with
my
test
environment
where,
for
some
reason,
sentinel
loses
its
mind
and
I'm
trying
to
figure
out
why
and
when
that's
breaking.
So
I
don't
have
that
happen
in
production
so
that
part's
still
a
work
in
progress.
C
Igor,
has
started
working
on
the
monitoring
side
of
things.
I
don't
know
how
far
he's
made
it,
but
we
want
to
make
sure
that
we
are
able
to
monitor
both
our
virtual
machines
and
the
kubernetes
deployment
at
the
same
time,
and
as
we
do
that
transition
and
then
our
epic
still
has
a
whole
slew
of
other
issues
that
we
have
to
work
through.
C
D
C
D
Garbage
if
you
have
some
time
tomorrow,
maybe
we
can
soak
up
because
right
now,
I'm
still
reading
and
going
through
the
epics
and
work
that
we
did
so
far.
And
maybe
we
can
sync
up
tomorrow
and
you
can
get
me
up
to
speed
where
we
are
and
so
on.
Very.
A
And
if
you
do
like,
I
I
I
do
think
yeah,
my
I
think
matt
is
close,
if
not
sort
of
maybe
yeah,
I
think
close
to
being
able
to
join.
So
if
there's
something
on
the
bms
that
having
him
pair
up
with
you,
will
help
them
get
shout
and
see
if
he's
around
yeah.
A
Do
you
know
like
what
are
the
kind
of
steps,
the
broad
steps
we
have
before?
We
think
we
might
have
a
kind
of
a
testing
set
up
on
pre.
C
A
Okay,
would
it
be
helpful
to
have
someone
from
quality
help
us
with
wellness
once
it's
on
pre,
so
quality
volunteered
for,
like
john,
is
available
to
help?
If
we,
if
we
need
any
extra
kind
of
like
test
guidance
for
this.
A
C
Something
is
definitely
wrong,
so
I
think
it's
going
to
be
more
along
the
lines
of,
let's
make
sure
that
we
are
ready
to
go,
and
then
you
know
test
this
and
pre-prod
as
much
as
possible.
You
know
we're
rolling
forward
and
backward
and
making
sure
qa
is
happy
in
both
directions
and
also
just
making
sure
that
we
don't
miss
anything
and
cause
an
accidental
outage.
A
Okay,
so
tomorrow
we
have
the
apac
emea
demo
is
at
8,
00
am
utc.
A
One
thing
graham
mentioned
he's
going
to
demo
is
he's
been
playing
around
with
a
poc
for
maybe
how
cows,
as
a
first
service,
could
go
into
like
a
dedicated
node
pool,
or
you
know,
get
us
a
bit
closer
to
being
able
to
kind
of
separate
things
up
inside
the
clusters.
So
he's
going
to
try
and
demo
that
tomorrow-
and
we
may
be
talking
about
some
of
the
other
work-
he's
been
doing-
to
remove
kind
of
some
of
the
chef
dependencies
from
the
kate's
workloads
as
well.