►
From YouTube: Kubernetes SIG Scheduling Meetings 20171109
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).
B
C
C
Yeah
well
the
same
issue
that
existed
forever,
so
we
need
a
little
bit
more
active
reviewers
in
the
community
sig
scheduling
because
there
was
a
lot
of
key
ours,
including
my
own.
I,
am
reviewing
a
lot
of
them,
but
I
can
only
be
my
own
pr's,
so
I
hope
we
can
find
more
people
who
are
willing
to
you
peers,
yeah.
D
C
E
Go
in
batches,
it
fits
in
spurts.
So
if
there
are
things
that
need
a
review,
it's
probably
better
to
ping
me
because
right
now
my
backlog
is
pretty
enormous
and
I
go
through
and
I
prune
it
like
once
a
week
go
to
go
through
all
of
them
that
needs
to
be
taken
care
of,
but
growing
the
reviewer
pool,
I
think
should
probably
be
part
of
our
efforts
to
because
I
know
like
we
spent
a
lot
of
time
getting
close
up
to
speed
and
now
he's
up
to
speed.
E
C
Yeah
yeah
I've
been
trying
to
help
review
some
other
people
up
here.
Ours.
There
were
a
couple
other
guys
relatively
active
in
the
community,
who
are
plenty
reviewers,
but
the
oh
I
believe
you're,
like
more
junior
engineers,
maybe
you're
not
fully
familiar
with
all
did
you
know
little
details
in
the
scheduler,
so
they
mean
a
little
bit
more
time
until
they
get
up
to
speed
and
they
become
a
familiar
fully
familiar
with
everything.
E
A
D
Why
don't
you
go
ahead?
Yeah
hi,
so
I
wanted
to
talk
about
that
issue
regarding
scheduling
of
best-effort
parts
based
on
node
uses,
I
think
in
the
past
the
reason
we
have
not
done
that
and
I
think
there
was
a
comment
from
Dame
previously
that
it's
blocked
on
the
matrix
API,
but
it
seems
now
matrix
if
API
is
had
been
part
of
cube
since
1.8.
C
Yeah,
so
you
think
it's
it's,
it's
actually
something
super
used
boy
I
I
briefly
saw
that
that
pure
that
you
sent
out
and
I
I
know
that
there
are
some
questions.
I
guess,
Muraki
asked
some
question.
Had
some
suggestions
on
@pr
but
mark
has
changed
roles.
I,
don't
expect
him
to
be
responsive
really
on
those
PRS
anymore.
So
maybe
you
can
go
ahead
and
and
implement
things
that
you
have
in
mind.
I
know
that
he
had
its.
The
suggestion
was
that
to
break
down,
did
the
stats
on
various
predicates
I
believe
right
or
priority?
C
B
C
D
A
C
D
D
I
think
I
would
say
that
that's
a
starting
point
because
anyway,
like
for
other
for
tech,
bustable
ports
of
guarantee
for
guaranteed
ports,
they
specify
resource
requirements
and,
in
fact,
when
we
calculate
load
uses,
we
also
compute
resource
requirements
of
those
spots.
So
I
think
it
seems
it's
right
now.
It
seems
it's
more
important
for
the
best
efforts,
because
we
don't
know
anything
about
their
uses
or
any
more
uses.
Yeah.
C
D
A
A
Probably
it's
a
good
time
between
that
and
1.9
for
folks
to
review
it
and
discuss
it.
Yeah,
but
I
agree
that
it
would
be
useful
and
I
haven't
been
following
the
metrics
API,
but
if
it
now
gives
the
kind
of
usage
information
you
would
need,
then
seems
like
that
would
be
it's
on.
It's
unblocked,
yeah,
Wow
I,
don't
know
how
much
history
like
does
it?
Does
it?
Does
it
capture
some
kind
of
average
over
some
time
period?
I
can't
remember
yeah.
D
A
C
A
Useful
so
I
don't
know
where
you
might
want
to
ask
the
folks
who
worked
on
that,
like
the
monitoring,
whatever
the
sake
that
you
say
whatever
it's
called
metric
like
if
they
have
a
plan
to
do
like
you
know,
averages
or
whatever
over
time
windows
and
expose
that,
because,
if
not,
you
might
have
to
do
that
yourself.
I'm,
not
sure
yeah.
D
Yeah,
so
yeah
they're.
Definitely
so
as
a
cop,
you
want
to
look,
do
all
those
things
and
they
definitely
we
want
to
talk
to
those
guys
also
yeah.
A
D
E
C
Freeze
I
will
also
try
to
address
the
pod
disruption
budgets
or
at
the
support
for
it
and
during
preemption
I'm,
not
so
hopeful
that
I
will
get
to
it
before
the
code
freeze
and
one
other
well.
There
has
been
a
couple
of
smaller
fixes
here
and
there,
which
are
not
very
notable,
that's
pretty
much
it.
That
I
can
remember.