►
From YouTube: Recap self-managed customer infrastructure challenges
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
Oh
hi,
everyone.
This
is
Mac
VP
of
informal
quality
and
then
Lyle.
Who
are
you
again?
Can
you
please
remind
me
who.
A
Great
so
this
came
up
our
one-on-one
discussion
and
also
to
follow
up
on
the
past
feedback
from
customer
success
on
the
challenges
of
self-managed
reference.
Architectures
Arts
help
manage
infrastructure.
We
were
reviewing
the
current
issues
that
were
raised
by
by
the
customers
and
how
we
can
dissect
that
into
channels
or
sections
and
then
and
Tackle
this
problem
more
iteratively
and
then
more
more
focused.
But
you
want
to
take
take
over
the
overview
that
you
you
mentioned
earlier.
Please
yeah.
B
B
You
know
a
choice
like
justice,
pulling
the
helm,
install
straight
and
not
reading
any
of
the
documentation
or
warnings
about
stateful
components
in
gitlab
and
then
their
one
of
their
one
of
the
very
important
pods
their
database
pod
got
killed
and
is
not
corrupted
so
that
expertise
piece
feeds
into
the
The
Rake
based
backup,
where
right
now,
gitlab
is
not
quite
opinionated
enough
about
what
to
do
at
that
level,
and
these.
B
These
customers
that
are
not
on
the
Savvy
scale
are
forced
to
make
some
guesses,
but
the
other
reasons
for
customers
who
aren't
able
to
run
GitHub
historical
choices.
That
happened
before
reference
architecture,
where
again,
customers
were
making
their
best
guess.
Some
were
operating
under
Professional,
Service
guidance
but
prior
to
reference
architecture,
and
you
know
in
the
in
the
in
the
wild
days
four
years
ago,
like
there
was
all
kinds
of
requirements
and
PS
did
their
best
job
and
so
I.
We
have
these
kind
of
older
gitlab
installations
and
then
I
could
say.
B
The
third
category
is
like
we
have
a
workload
that
isn't
really
accounted
for
in
reference
architecture,
you're,
seeing
this
in
places
like
rivian
or
Zoom,
where
the
reference
architecture
that
they
have
seems
like
it
should
should
keep
up,
but
they're
doing
something
special
or
weird,
with
a
mono,
repo
or
tons
of
Ci
or
something
else
so
I
think.
The
tasks
that
we
have
right
now
is
shrinking
the
pool
of
customers
that
are
in
this
state
who
aren't
able
to
manage
their
own
architecture
while
making
sure
that
customers
don't
fall
into
that
pool
and
I.
B
B
Like
that
who
blessing
story,
distribution
team,
who
makes
a
fully
featured
Helm
install
that
you
can
deploy
and
log
into
and
use
and
God
bless
our
customers
that
then
put
that
into
production,
because
you
know
it
works,
but
that
kind
of
that
would
be
a
reason.
B
Another
reason
would
be,
they
have
some
sort
of
specific
requirement
within
their
company
or
they
need
to
make
some
sort
of
trade-off
like
I,
definitely
see
this
in
really
large
customers,
where
you
have
a
networking
team,
a
storage
team,
a
VM
team
and
getting
a
VM
provisioned
takes
three
months
doing
it.
This
other
way
takes
another
few
months.
B
B
Variations
yeah,
so
it
I
would
say
that's
kind
of
that's
kind
of
the
way.
I
see
the
problem
did
I
miss
anything
Mac.
That
was
like.
A
A
little
bit
I
think
that
was
right.
They
only
mentioned
three
things
right
and
the
three
is
like
one
is:
customer
service
doesn't
have
enough.
Kubernetes
expertise,
I
think
that's
the
long
game
that
we
can't
solve
immediately.
A
This
is
how
you
actually
do
backup
and
be
more
opinionated
about
it.
So
I
I
think
we
would
want
to
expand
on
two
and
three
is
to
have
a
process
to
migrate.
This
historical
setup,
faster
with
validation
and
then
third,
is
the
third
one
is
within
that
list
of
unsupported
workload
or
workload.
That's
not
accounted
for.
Let's
just
come
up
with
a
list
and
let's
solve
the
the
easiest
clear
one.
First
and
I
acknowledge
that
in
three
there
there
will
be
unknown
unknown
unknowns
that
we
we
never
thought
about.
B
Yeah
so
I
mean
overall,
like
with
those
goals
in
mind.
I
I
just
want
to
right
size,
the
level
of
urgency
and
importance
here
we
do
have
a
large
set
of
customers
who
are
having
trouble
with
gitlab
because
of
these
sort
of
historical
reasons
or
expertise
reasons,
but
by
and
large
customers
today
are
choosing
reference,
architectures
and
I
think
they're
doing
well,
except
for
in
the
case
when
they
choose
poorly
or
don't
understand
their
own
workload.
Enough.
A
Okay,
yeah
thanks
thanks
for
that,
I
think
the
team
will
really
appreciate
that
yeah
I
think
the
the
quality
team
has
a
a
planning,
epic,
we'll
send
this
recording
to
the
team
to
provide
additional
Clarity
and
and
guidance
on
what
they
can
be
looking
forward
to
next.
Thank
you.
Thank
you.