►
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
So,
according
to
the
agenda,
where
I
just
modified
it
question
about
how
many
reviews
I
think
I
asked
a
couple
weeks
ago,
and
it
was
like
two,
but
all
the
docs
say
one.
So
it
does
seem
like
it's
one
review
unless
of
course
well
depending
upon.
If
you
look
at
the
newer
doc,
unless
of
course,
your
working
group
only
has
like
two
people
and
then
it's
zero
or
something
three
or
few
are
people.
A
B
C
Yeah
most
of
the
and
I
think
like
the
the
two
was
based
on,
like
the
so
I
believe
at
some
point
in
salsa.
They
were
saying
something
about
that
like
doing
having
two
two
approvals,
but
then
that
was
clarified.
I
think
we
we
had
a
whole
discussion
in
that
Branch
protection
thing
as
well.
The
other
RFC
and
I
think
it
was
not
really
practical
for
a
lot
of
working
groups.
A
Oh
I
haven't
told
you
that
b
stuff
doesn't
work.
No,
you
told
me
I
told
you
that
it
doesn't
work.
C
Yeah,
you
said
that
you
cannot
give
it
access
to
a
repo.
That's
not
you
don't
personally.
A
C
A
Doesn't
doesn't
work
at
Orcs
at
all,
yeah
all
right
get
to
PRS
and
we
will
see
some
things
that
mostly.
D
D
A
It
just
opens
the
new
window
anyway.
This
is
there's
like
six
of
these
PRS,
which
have
already
been
reviewed,
so
I'm,
probably
just
going
to
click
merge
unless
anybody
has
objections,
but
the
the
version
metadata
is
not
wrong
but
predate
so
a
lot
of
it
included.
A
A
lot
of
these
PRS
are
about
removing
this
stuff,
because
this
messes
up
scanners
like
cift.
Well,
that's
not
true
sift
find
well,
CFT
finds
a
version,
but
when
you
give
it
something
like
this,
it
declares
the
version
to
be.
A
0.0.0
basically
ignores
this,
and
it
uses
the
other
metadata
about
the
timestamp
of
the
commit
and
makes
a
a
version
of
000000
D
timestamp
do-
gsha
or
something,
whereas
if
you
have
main.
version
it,
it
very
clearly
knows
that
is
the
version
which
works
well.
If
you're
do
using
get
mod
because
then
you
can
go
get
that
thing
at
that
version
and
it
all
works,
whereas
the
version
that
sift
gets.
If
you
give
it
something
like
this
is
not
something
you
you
can
go
Mod
download.
A
So,
in
this
case,
we're
actually
duplicating
it
because
in
the
Bosch
CLI,
the
version
label
is
actually
used
by
the
command
the
version
command,
whereas
this
is
just
some
magical
place
that
doesn't
really
exist.
It
seems
like
there's,
no
there's
no
version
variable
declared
in
the
main
here,
but
this
is
the
way
that
go
metadata.
If
you
do
a
Go
version
DM
on
a
binary,
it
will
look
for
this
and
then
say
Oh.
This
is
the
version
of
that
thing.
A
So
I
think
the
the
bin
build
still
has
functionality
because
it's
like,
if
you
don't
specify
version
label,
it
uses
default
version
which
is
Dev
build.
So.
D
A
Still,
you
can
still
do
Dev
releases,
you
just
still
need
some
sort
of
version
thing,
and
all
of
this
data
is
actually
included
in
the
binary
anyway.
Go
now
when
you
build,
will
include
the
git
revision
and
the
timestamp
as
part
of
that
metadata.
Just
not
doesn't
call
it
the
version
field,
so
all
the
data
is
still
there.
It's
just
in
a
well
better,
structured
way.
A
This
all
of
this,
which
is
everywhere
in
our
code
bases
by
the
way,
seems
to
all
be
historical
stuff
before
go
mod
before
get
revisions
were
a
core
part
of
the
version
of
a
get
thing
or
a
go
thing
all
right.
That
was
a
lot
of
rambling
by
me.
Any
other
questions.
A
No
remote
asked
the
question:
oh
yeah,
all
right,
I
click.
A
A
Well,
it
declares
it
as
part
of
the
version
we
looked
at
the
history
of
this,
so
we're
building
runy.
Here
it's
built
from
this
exact
version
of
the
runy
tarball.
A
We
do
include
the
libc
comp,
so
it's
like
maybe
different
than
the
official
one
in
some
way,
but
we
also
yeah
I.
Don't
know
we
looked
at
the
history
of
the
version
and
at
one
point
they
moved
it
from
the
front
used
to
be
in
the
front
BPM
Dash,
and
that
turned
out
to
break
everything
apparently
and
the
version
before
that
it
was
just
called
BPM.
A
The
version
was
just
BPM,
it
was
the
BPM
version
of
runc,
so
this
will
once
again
allow
scanners
to
more
easily
know
what
version
of
runc
this
thing
was,
as
opposed
to
some
odd.
A
A
Oh
Le,
bran
I
missed
the
button
all
right,
I'll
clean
those
up
later
I
got
a
lot
of
stuff
on
my
machine
checked
out
to
delete.
This
is
all
the
The
Blob
store
CIS
and
they
have
the
exact
same
thing.
They
collected
up
this
data
and
then
combined
it
into
this
version
string
and
then
did
that
same
exact
problem.
A
B
A
Work
also,
so
all
of
the
commit
messages
are
a
lie:
I
mean
it
will
allow
that,
but
really
we
want
to
go
download
them
so
that
we
can
get
the.
A
Source
all
right
issues,
oh:
where
are
we
with
issues
so
last
meeting
was
21s.
B
A
A
Yeah,
but
he
was
only
doing
that
to
try
to
reproduce
something
but
yeah
whatever
he's
trying
to
reproduce
is
likely
the
vm's
still
there,
it's
still
heartbeating,
so
the
agent's
not
going
to
solve
your
problem.
This
is
sort
of
a
problem
that
it
would
be
nice
for
Bosch
to
try
to
solve
in
some
way
or
at
least
make
it
easier
for
people
to
solve.
I,
don't
know
what
people
do
like
really.
This
is
a
we've
had
this
request
from
like
Diego
a
while
back
or
for
Diego
to
like.
A
Oh,
there
was
a
problem
mounting
the
disc
or
the
disc.
Something
happened
to
the
disc
and
like
we
should
have
the
agent
check
the
disc
too
and
see
if
it's
healthy
and
it's
like
what
is
what
is
a
healthy
disc
like
you're,
going
to
end
up
recreating
VMS
that
aren't
actually
broken
more
times
than
you're
going
to
fix
things,
or
at
least
some
percentage
of
the
time,
which
is
probably
bad.
It's
really
hard
to
tell
the
health
of
the
system
from
inside
of
the
system.
A
C
That's
a
bit
of
a:
why
did
it
go
in
that
state
to
begin
with
and
will
recreating
it
actually
fix
it?
Yeah
feels
like
a
human
needs
to
be
involved.
A
C
Cases
yeah,
and
it
also
feels
like
it's
the
responsibility
of
the
the
process.
That's
spinning
up.
All
these,
like
so
Concourse
in
this
case,
should
do
better
resource
management
right
and
kill
processes
that
are
going
beyond
what
they
are
allocated.
A
A
B
It
now
it's
you,
you
also
get
these
issues
when
there
is
like
a
really
Noisy
Neighbor
in
one
of
the
is
on
their
cluster
or
their
wck
or
whatever.
D
A
They
haven't
responded
back.
They
were
there
three
days
ago,
so
I
a
little
more
time.
A
Yeah
is
this
back
in
my
ballpark
again
or
his
I
forget?
Oh
look!
It
was
me
again:
okay,
good
finally,
IM
meeting
where
I
didn't
come
to
saying.
I
should
get
back
to
this
okay,
so
he
hasn't
responded
back.
Assumably
he'll
get
back
at
some
point
soon,
but
it's
been
just
about
a
week.
B
B
C
A
C
A
Saying
Penny
was
in
Europe
for
a
week
and
he
just
got
back
Tuesday,
so
I'll
I'll
alert
him
that
wait.
Why
is
this
September
22nd?
The
last
response
was
in
July
or.
D
D
A
Right,
I'm
gonna,
move
on
from
that
this
one's
interesting,
so
Bean
you
I,
couldn't
actually
find
where
any
of
these
metrics
came
from
I
just
found
in
the
engine
X
how
the
traffic
was
routed
to
different
different
things
based
upon
the
different
endpoints
but
you're
saying
that
the
data
actually
comes
out
of
the.
E
Datase
now
I
mean
this
collector,
which
so
there
are
the
generic
metrics
which
are
provided
by
and
they
are
for
the
end
points
and
I
was
thinking
that
our
workers,
they
don't,
they
are
not
accessible
via
Endo
Bush
direct
API,
end
points
correct,
so
we
have
metrics
for
the
director
end
points
and
there
are
some
metrics
about
tasks.
So
the
metrics
here
is
this:
Resurrection
enable
total
POS
tasks
in
total
and
so
on,
which
are
actually
gathered
from
the
database
in
the.
A
Oh
yes,
yeah
yeah,
those
ones
that
makes
sense.
There
are
the
API
metrics,
the
API.
A
Metrics
are
those
are
just
like
there's
a
bunch
of
metrics,
so
I
didn't
I
didn't
try
to
hit
this
with
actually
a
a
Prometheus
client
but
I
just
curled
it
and
it
spit
out
a
bunch
of
like
each
endpoint.
Every
API
endpoint
was
listed
in
the
API
metrics
and
like
it
had
some
sort
of
like
data
about
number
of
requests
or
or
something
in
there
and
times,
and
those
are
probably
wrong
because
those
aren't
stored
in
the
database
and
we're
using
Puma.
A
A
Which
makes
the
whole
situation
terrible,
but
yeah
the
T
like
these
things,
these
things
come
out
of
the
database,
though
so
it's
going
to
be
fine,
these
these
are
all
yeah.
These
are
all
things
that
we
just
feed
with
the
database
INF
every
X
number
of.
A
A
Magic
it's
up
to
dat.
Well,
we
we're
removing
support
at
least
plug
in.
C
D
A
A
I
didn't
see
anything
they
were
doing
here.
I
didn't
see
any
code
that
seemed
to
solve
the
problem.
It
was.
A
A
It's
still
got
its
own
Global
interpreter,
lock
though,
but
I
don't
know
somebody
if
somebody
should
do
some
sort
of
benchmarking
test
with
and
without
this,
but
yeah
I
mean
this
is
basically
you
have
this,
where
you
need,
basically,
a
load
balancer
in
front
to
gather
the
metrics
or
you
need
to
store
the
metrics
in
a
the
metrics
in
some
sort
of
data
store
could
be
dis
could
be
something
more
convoluted.
A
Ever
we'll
see
if
they
want
to
come
back
on
that
with
any
more
info,
though
any
other
tabs
here,
I'm
supposed
to
look.
E
At
there
was
the
RFC
there's
a
draft
implementation
for
the
RFC
regarding
P
integration,
I,
don't
know
what.
C
E
C
A
Which
one
are
we
looking
at
here?
We
kep.
E
A
D
A
A
Well,
all
of
our
stres
are
undefined,
which
is
probably
a.
E
Problem
if
you
fall
DFC
mentioning
I
think
they
mentioned
that
this
is
ready
for
there
above
the
paying.
D
D
E
A
C
C
E
A
Filters,
oh,
does
anybody
have
anything.