►
From YouTube: Community Meeting, August 23, 2022
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
I
didn't
see
a
lot
of
topics
come
in,
so
if
you
have
anything
just
throw
it
in
the
chat,
I
wanted
to
quickly
start
off
with
a
comment
on
our
structured
logging
revamp.
So
as
of
the
last
rebase,
we
have
the
kubernetes
contextual,
like
structural
logging,
like
all
of
the
libraries
and
whatnot
are
in
there,
so
I've
been
slowly
changing
stuff
over
if
anyone's,
got
free
cycles
and
wants
to
spend
a
little
bit
of
time.
I
would
appreciate
some
help
on
this.
A
It's
very,
very
boring,
but
once
we're
done
kind
of
the
grand
vision
that
I
was
really
hoping
for
us
to
get
to
is
when
an
incident
test
fails
or
you
know
some
bug
comes
up,
we
should
be
able
to
go
through
the
logs
and
filter
down.
A
I'd
like
to
see
all
the
logs
for
a
particular
reconciler
for
a
particular
logical
cluster,
maybe
even
particular
keys
and
whatnot,
and
you
know
not
spend
as
much
time
sifting
through
stuff
so
getting
there
would
be
super
awesome.
We're
pretty
close.
I
think
some
of
the
virtual
workspace
directories
and
the
syncer
are
all
that's
left,
so
any
help
would
be
appreciated
before
we
jump
into.
B
A
Cool
paul
did
you
have
anything
you
want
to
talk
about.
C
With
the
questions,
one
quick,
general
announcement
that
I
put
on
the
issue,
the
second
september,
2nd,
is
closing
date
for
0.8,
so
that
means
we're
two
weeks
out
and
we
need
to
start
thinking
about
0.9
items.
I've
linked
in
our
work
packages
document
as
usual,
let's
put
names
to
topics
on
there,
I've
copied
over
the
items
that
will
still
be
in
development,
so
I
think
we
have
news
by
most
things.
C
A
It
looks
like
we
I'll
follow
up
here.
Something
internal
didn't
happen
correctly.
We
didn't
get
the
annotation
put
on
there
correctly
for
the
owner.
B
C
B
Just
try
to
read
through
the
description
of
that
stuff.
It
looks
cloth
related
assign
it
to
me.
Maybe
I
can
have
a
look
at
this
one
at
least
comment
and
see.
What's
going
on,
there.
A
A
A
A
Awesome
and
I
think
the
workaround,
because
this
they
hit
this
in
quota,
I
think
the
workaround
was
just
like
not
quota
the
virtual
resource,
but
yeah.
A
I
can
assign
myself
on
this
one.
This
was
something
that
I
hit
trying
to.
A
B
Unless
nobody
answers
yep
most
most
of
them
are
highly
privileged,
in
fact
yeah.
So,
in
fact,
I
don't
know
if
you
saw-
and
that's
also
why
I
pulled
you
in
steve
and
the
the
par
with
the
with
the
deep
tsar
in
fact,
already
starts
using
sort
of
like
completely
unprivileged
users
and
creating
roles
and
role,
bindings
and
stuff
like
this.
So
I
guess
this
is
a
gentle
start
towards
it.
B
I
totally
agree
we
should
you
leverage
more
service
accounts
and,
and
the
implied
permissions
feel
free
to
assign
this
one
to
me,
although
I
feel
this
is
like
a
very,
very,
very
broad
topic,
and
I
think
it
makes
sense
to
look
at
authorizes
and
like
we
have
a
test
e2e
package,
where
an
e2e
run
which
explicitly
tests
all
authorizers
and
does
pretty
much
unprivileged
calls
in
there
and
think
we
should
add
more
stuff
in
there,
but
we
should
also
think
about
concrete
e2e
test
scenarios.
B
What
we
want
to
test
because
there's
like,
I
guess,
like
a
combinatorial
explosion
of
things,
we
could
theoretically
test
with
with
constraints.
So
we
should.
We
should
judge
by
a
case-by-case
basis.
I
guess
that's
my
that's
my
call
and
like
the
test,
e2e
authorizes
packages,
the
canonical
source
of
things
where
we
want
to
hit
things,
I'm
privileged,
so
to
say,
but
again
like
the
deep
star.
E2E
is
a
good
canonical
example,
where
we
already
start
introducing
these
subtleties
with
the
privileged
accounts.
A
Yeah-
and
I
think
like
outside,
of
the
authorizer
ees
we'd
be
less
interested
in
the
combinatorics
of
it,
and
I
think
it
just
gives
us
a
really
good
insight
into
like
user
experience
more
so
there.
B
Actually,
what
I
want
to
do
is
that's
going
to
be
one
issue
later
with
currently
the
the
our
back
messages
that
we
get
are
partially
very
confusing
right.
So
what
I
want
to
do
is
when
I
will
improve
the
messaging
actually
provide
an
e2e
test
which
will
assert
concretely
our
back
error
messages,
orbit
reasonings,
which
come
back
in
in
case
of
authorization
violations.
So
I
guess
that
could
be
a
nice
part
of
it.
A
A
B
Yeah
I
edited
myself
in
the
deep
star
stuff
so
again,
like
I
think
the
deep
star
pr
is
good
enough
now
and
I
will
jump
into
exactly
this
issue
to
improve
myself.
A
I
I
was,
I
think,
yeah
I
was
trying
to
bind
an
api
and
the
cluster
workspace
that
the
export
existed
in
didn't
exist
and
we
are,
there
might
be
more
that
we
can
write,
but
what
we,
what
we
have
in
the
error
messages,
is
also
conservative
not
to
give
the
user
too
much
information
right,
and
I
wonder
if
there's
do
we
have
a
good
enough
support
in
the
audit
traces
now,
so
that
when
an
admin
hits
this,
they
can.
B
Okay,
exact,
that's
the
other
issue,
I'm
working
on,
so
we
have
some
stuff
that
andy
merged.
I
believe
this
week
or
last
week,
but
I
want
to
go
through
the
remaining
authorizers.
I
I
completely
agree.
We
have
to
have
like
a
good
line
of
like
we
want
as
much
fine-grained
information
as
we
can
just
have
in
the
audit
logs.
B
B
So
I
guess
like
this
will
be
like
quite
like
a
little
bit
of
juggling
and
bike
sharing
on
the
prs
that
I'm
going
to
submit
but
yeah
it's
twofold,
like
the
I'm
working
on
the
audit
as
well
and
on
the
user
facing
messages-
and
we
just
have
to
you-
know,
look
at
the
pr
and
and
see
if
we
have
a
good
line
where
we
don't
need
too
much
stuff
on
the
regular
reasons
that
come
back
to
the
user.
A
Awesome-
and
I
think,
there's
I
imagine-
there's
probably
already
tooling
around
this
somewhere
but
like
a
very
quick
cli
to
query
the
audit
to
figure
out
to
go
from.
I
was
the
user
making
this
request
with
this
user
agent.
All
the
way
to
like
that's
my
audit,
you
had
here's
the
reasons
why
I
failed
automating
that,
in
some
sense,
might
be
also
good,
but
cool
awesome
see,
I
think,
on
on
those
two
fronts.
That
will
be
very
useful.
A
A
B
I
didn't
look
at
this
one
just
to
sign
this
one
to
me,
and
maybe
I
will
swing
back
with
okay
with
chris.
I
wanted
to
meet
him
anyways
and
then
I
can
put
in
more
details,
but
it
sounds
also
like
a
good
issue
to
me,
but
I
will
I
will
go
back
and
talk
with
chris,
because
I.
A
D
So
this
one
I'm
currently
answering
in
fact
I
didn't
push
the
comment
button
for
now,
but
in
fact
it
has
never
been
it
has.
I
mean
update
for
the
in
the
workspaces.
Virtual
workspace
has
never
been
implemented
from
the
beginning,
so
just
the
method
is
not
here
in
the
rest
implementation.
D
D
A
Awesome
david:
do
you
have
an
idea
of
milestone
for
this
one.
D
Well,
I
don't
know
how
critical
it
is,
in
fact
I
mean
as
soon
as
we
as
long
as
we
don't
have
precise
or
critical
use
cases
to
add
labels.
You
know
for
end
users,
not
sure
it's
it's
heavily
critical.
D
A
D
Am
I
right
that
that
it's
it's
there
is
no
limitation
that
you
can,
through
an
api
binding
just
by
the
cluster
workspace
and
directly
work
with
it.
If
you
have
sufficient
permissions
yeah
so
that
that
might
be
sufficient,
I
assume.
A
B
Just
one
node
david,
you
effectively
need
the
bind
verb
permission
against
the
api
export
in
the
original
workspace
and
using
the
deepstar.
You
won't
need
general
access
to
the
api
exporting,
so
you
will
just
need
this
word
permission.
D
A
Cool
this
was
a
bug
with
the
sinker
deleting
the
root
ca.
I
don't
know
if
we
ever.
C
I've
merged
well
a
fix
for
another
issue
that
was
kind
of
related
to
this,
because
the
singer
was
missing:
name,
translation.
So
can
you
assign
me
this
and
I
will
check
if
okay,
thank
you
awesome.
C
D
D
It
seems
it
to
me
it's
finished,
but
maybe
I
would
let
him
just
close
the
issue
or
you
know,
if
it's
possible
to
bring
him
maybe
tomorrow
for
someone
tomorrow
morning.
It
seems
to
me
that
the
the
two
parts
have
been
done.
All
right.
I
just
you
know.
D
Then
there
is
sorry
yeah,
yes,
and
this
is
this
one
is
a
follow-up
of
the
previous
ones.
He
is
currently
working
on
it.
D
D
C
D
Oh
sorry,
excuse
me
this
one
is
yeah
yeah.
This
one
is
the
you
know
last
follow
up
after
the
previous
one,
that's
mainly
because
when
you
point
to
external
external
api
exports
from
a
thing
target,
there
is
a
check
that
the
corresponding
thinkers
correctly
support
correctly
imported
the
shimaz
in
the
in
the
sync
target
workspace,
and
then
we
check
that
the
shimmers
are
compatible.
The
idea
is
that
if
you
have
a
location
workspace,
you
know
a
sync
target:
well,
a
location
workspace
that
has
bound
to
external
apis.
D
We
have
to
check
that
those
apis
we
want
to
use
are
also
present
in
the
physical
cluster
and
are
compatible,
and
we
would
only
schedule
a
given
workload
of
these
apis
only
for
sync
targets
where
this
api
is
compatible
inside
the
the
physical
cluster.
D
So
that's
quite
important
in
case
yeah,
you,
you
have
a
thing
target
that
joins.
That
has,
you
know
mainly
possibly
a
distinct
cube
version.
Then
the
t,
the
with
incompatible
shima
corresponding
to
the
you
know,
common
cube
that
that
api
exports
that
we
will
rely
on
in
the
future,
for
example,
but
that's
the.
A
D
A
A
And
I
think
yeah.
I
think
that
was
all
that
we
had
so.
A
These
are
the
epics
we've
got
running
through
here.
It
doesn't
look
like
we've
had
an
update
paul.
Do
you
know
if
we,
if
we
have
an
update
on
priority
fairness?
Oh
I
guess
it
got
moved
to
zero
nine.
C
We
need
to
see
if
we
can
talk
to
sean
about
progress
on
the
permission
claims
one
yeah,
I
think
the
other
thing
to
mention
here
is:
if
you
do
have
something
that
is
not
going
to
land
in
the
next
two
weeks
or
you're,
pretty
sure
it's
not
going
to
land,
we
need
to
try
to
start
scoping
those
out
within
the
epic
itself
or
moving
the
epic
to
the
next
fixed
version.
A
Yeah,
I
guess
none
of
these
seem
to
have
any
updates,
except
for
this
one
that
got
punted
off.
So
do
you
think
it's
worth
us
jumping
into
these,
or
should
we
just
do
it
in
the
issues
themselves?.
A
Okay,
it
doesn't
sound
like
we
have
anything.
We
need
to
talk
about
right
now.
Then,
let's
follow
up
on
these.
I
chatted
with
sean
yesterday
paul,
so
I
can.
I
can
follow
up
with
the
expert
stuff.
I
know
andy
took
over
some
of
that.