►
From YouTube: Charts Chat 20180306
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
B
A
C
You
so
yeah
be
quite
often
have
the
problems
that
it
can
give
you
yes
charts.
That
would
be
jobs
that
need
like
something
like
keys
stuff
like
that,
which
we
don't
really
wanna
commit
to
the
real
chart.
So
we
will
be
able
to
blue
like
this
and
extra
test
units
they're,
just
not
use
required
values,
because
you
just
can't
provide
them.
So
this
will
actually
choir
and
doesn't
134
tell
you
it's
just
not
possible,
so
we
need
to
have
some
way
to
specify
values
and
for
for
tests
in.
C
So
this
is
my
proposal
here,
but
antitrust
maybe
have
a
separate
folder
where
we
can
put
tests,
Yama's
files,
jungle
files
in
for
each
file,
so
this
would
even
allow
multiple
tests
with
different
configurations,
if
necessary,
for
certain
shots
just
on
the
test
with
any
younger
file
in
order-
and
this
is
just
a
little
little
best-
script-
I
put
together
just
to
illustrate
how
it
how
it
could
work.
So
this
is
not
really
I
didn't
look
at
the
actual
account,
so
this
would
have
to
be
integrated
somehow,
but
it
shouldn't
be
too
difficult.
D
D
C
D
As
a
side
note,
I
will
say
that
I
think
charts
should
really
stop
using
required
because
in
general
it's
a
bad
user
experience.
It's
almost
better
to
use
kind
of
what.
If
you
look
at
what
minecraft
does
or
some
of
the
other
charts
where
it
happily
accepts
value
to
a
wrong
but
and
goes
and
tells
you
how
to
how
to
fix
that.
D
I
mean
helm,
has
it
but
I,
don't
think
I
don't
know.
I
I
would
say
that,
as
in
kubernetes
charts,
we
probably
don't
want
to
use
it
it's
there,
but
you
know
I,
don't
think
it
really.
This
is
just
my
own
personal
opinion.
I
mean
we
can.
We
haven't
really
talked
about
it,
but
I
think
that
it
it's
probably
better.
D
If
we
do
do
it,
the
way
you
know
minecraft
does
it
with
the
EULA,
then
support
the
require
thing,
and
we
should,
if
we
city
charter
using
required,
we
should
either
change
them
to
use
the
other
approach.
Or
you
know
when
we
get
a
PR
with
someone
using
required,
we
can
ask
them
to
use
the
different
approach,
but
it
yeah
again.
That's
just
my
opinion.
I,
don't
know
what
other
people
think
about
it.
But
for
me
it's
like
a
not
a
great
user
experience.
D
The
way
minecraft
does
it
is,
if
you
say
you
can
do
helm,
install
stable
minecraft,
but
then
you
actually
don't
get
any
deployments
and
then
you
get
in
the
notes
or
txt
you
get
a
printout
saying.
You
know
error,
you
need
to
pass
in
Euler
accepted
equal,
true
or
whatever,
and
then
once
you
do
a
helm
upgrade
with
with
that
come
on
you
can,
then
it
goes
in
in
sort
of
appointment
and
then
you
get
pots.
C
Like
this
I
mean
some
charts,
you
probably
will
have
to
do
that
like
this,
but
you
need
some
kind
of
license.
You
can't
really
install
anything,
but
we
also
have
two
charts
that
just
and
there
was
a
TR
wherever
matt
is
also
on
each
of
that's.
It
suggested
to
remove
any
keys,
because
we
don't
really
want
to
have
keys
in
there,
which
is
not
good
and
have
keys
in
there.
So.
C
A
A
D
I
should
have
mentioned
that
that
was
the
the
user
experience.
Fitting
is
a
separate
issue
right.
We
still
need
two
values
file
for
CI,
oh
yeah
yeah
in
case
of
minecraft.
Yes,
exactly
we
need
to
set
that
unit
just
to
be
able
to
test
those
pods
actually
gets
tied
up
in
the
same
case,
it's
the
same
case
with
ghosts.
I
think
where
you
need
to
provide
like
a
ghost
host,
which
you
know.
Presumably
we
just
want
to
set
something,
like
example
con
just
so
that
we
get
a
deployment.
A
D
I
mean
I
mean
this
talk
of
like
things
like
JSON
schema
and
you
know
over
other
ways
of
representing
values
so
that
you
can
hint
as
to
what
about
you.
Might
you
know
this
schema
for
value
should
be
or
something,
but
without
that
information
right
now
the
required
thing
is
kind
of
good,
because
helm
doesn't
actually
tell
you
doesn't
give
you
a
good
error
message
like
why
this
thing
is
needed,
right,
yeah
or
what
it
might
be.
The
nice
nice
thing
about,
do
it.
E
D
G
F
A
Imma
take
silence
means
we
were
here
generally
accepted,
so
the
next
steps,
I
think,
would
be
to
have
a
go
past
Vic,
because
he
knows
this,
this
part
of
the
CI
pretty
well
and
to
make
sure
that
everything
looks
good
there,
which
I
don't
see
an
issue
and
just
also
to
run
it
by
the
hell
maintainer,
because
just
to
see,
if
there's
any
feedback
or
anything
there
from
them,
I
don't
see
an
issue.
But
if
we
like
commandeer
directory
I
want
to
make
sure
that
in
the
helm,
three
thinking
we're
not
thinking.
A
You
know
that
the
directory
schemas
and
everything
line
up
for
this,
but
overall
I
like
where
this
is
going.
If
some
of
this
may
get
more
complicated,
when
we
deal
with
things
like
timeouts,
you
know
timeouts
and
assisting
because
we've
already
got
that.
So
we'll
have
to
be
careful
when
we
start
doing
more
than
one
config
file
and
how
this
impacts
timeouts.
On
some
of
these
things,
but
other
than
that,
it
looks
like
something
we
can
go
forth
with
I
like
it.
D
C
E
C
A
Can
and
and
so
a
couple
of
things
here-
yeah
four
different
ones.
We
can
do
different
things
also
if
we
can
probably
upgrade
to
2.8
dot,
one
which
is
out
now
and
in
two
one
weight
is
fixed.
So,
instead
of
constantly
checking
for
things,
we
can
probably
use
weight,
uninstall
and
then
the
stuff
should
come
up
and
that
should
help
with
some
of
the
issues
we've
had
and
our
CI
testing.
Oh
sweet.
B
Okay,
just
quickly,
while
we're
on
the
topic
of
CI,
is
there
an
established
way
way
or
place
to
put
extra
manifests
to
actually
say
run
a
job
to
actually
do
integration
testing
against
the
deployed,
for
example,
I've
been
working
on
for
the
Makoto
chart,
actually
running
some
stuff
to
insert
data
in
and
making
sure
that
data
is
inserted
into
the
database
a
very
established
way
or
place
to
put
that
sort
of
stuff.
Yeah.
A
D
A
No,
that's
all
right.
The
docs
are
not
always
the
easiest
to
find,
but
yeah
there's
home
testing
capabilities
built
in
and
our
CI
will
actually
run
when
we
do
so.
There's
the
circle
CI
stuff,
where
we
don't
run
home
tests,
but
when
we
actually
install
it
and
we
do
the
deeper
testing
the
end
end
testing,
we
will
install
it
then
we'll
run
home
tests
on
it,
which
will
run
whatever
tests
are
included
that
are
documented
here,
and
so
we
actually
test
this
upstream.
Every
time.
C
Yeah
really
I
came
across
the
the
TPL,
our
template
function,
which
is
I,
think
a
nice
addition.
They
was
playing
around
with
it
and
you
can
actually
to
some
degree
templating
in
values.
Younger
and
I,
was
thinking
if
we
should
kind
of
encourage
using
that,
or
especially
since,
more
and
more
charts
use
the
Tamil
function.
C
C
E
C
Yeah,
so
I
would
really
like
to
see
templating
in
values
about
that
would
yeah
like
a
first
election.
F
A
E
C
C
A
So
for
reference,
I've
added
this
to
the
discussion
for
the
how
meeting
later
this
week
in
two
days,
so
we
can
start
talking
about
it
there,
if
you're
not
able
to
make
it
I
can
try
to
represent
this
I
think
I've
got
it,
but
we
can.
We
can
have
that
discussion
there
as
well
to
see
what
the
path
would
be
to
getting
it
in
to
helm.
What
blockers
or
any
details
on
that
direction
from
the
other
helm
course,
especially
those
far
more
familiar
with
the
ins
and
outs
and
pitfalls
of
this
I.
H
Kind
of
seemed
seems
like
a
strange
anti
pattern
or
another
kind
of
layer
of
inception,
but
one
of
the
one
of
the
examples
here
I'll
just
paste
the
link
to
it
of
one
of
the
use
cases
just
for
for
when
that
longer
conversation
happens,
not
you
is
to
potentially
call
functions
on
a
bunch
of
files
on
a
range
of
files.
I
mean
it
seems,
like
one
year's
case
yeah.
H
C
H
I
definitely
come
from
its
School
of
Engineering,
where
you
want
to
or
when
your
code
has
any
indirection
in
it.
You
need
some
kind
of
comments
explaining
what
the
hell
is
going
on.
You
know
why
you,
why
you're
doing
this,
what
it
isn't
that
doing?
If
it's
not
if
it's
really
not
obvious,
you
know
mostly
why,
but
maybe
that's
something
that
we
could
recommend
as
a
possible
point
of
discussion
or
charts.
A
And
I
think
one
of
the
things
about
this
is
how
to
make
it
an
easy
user
experience
which
either
makes
it
means
it
needs
to
be
intuitive
or
explicitly
documented.
How
to
do
it,
and
so,
whatever
the
path
is,
it
should
either
be
something
that's
just
so
implicit
anybody
can
figure
it
out.
Minimal
documentation
of
pointers
or
an
example
with
documentation
should
be
very
clear,
so
we
have
a
best
practice
or
yeah.
H
A
But
then
we
got
best
practices,
then
we
can,
you
know
PR
review
against
those
all
this
fun
stuff
right,
so
that
would
be
what
I
would
be
going
for.
Is
we've
got
a
clear
direction
on
how
to
handle
this
coming
out
of
our
discussions,
and
then
it
is
either
PR
is
made
against
him
to
make
it
change
to
enable
some
of
this
or
a
PR
is
made
against
the
helm,
documentation
to
add
examples
on
how
we
do
this
in
details.
B
C
D
So
I
actually
didn't
know
that
this
existed.
My
take
is
that
this
is
gonna,
make
value
Falls
like
really
really
complicated,
specifically
for
a
use
case.
This
outline
here,
the
affinity
use
case
I
mean
at
the
same
point
where
we
have
to
draw
a
line
between.
You
know
every
single
thing
that
you
can
configure
in
in
any
kubernetes
spec.
That
is
what
we
make
available
in
the
values
at
Yammer,
and
this
is
potentially
crossing
that
line.
But
you
know
that
the
files
globbing
use
cases
is
an
interesting
one
as
well.
D
B
A
The
next
one
and
I
dropped
a
Lincoln
is
the
Jenkins
failing
CI
test
for
those
of
you
don't
know.
This
is
the
test
that
runs
every
hour
against
the
tip
of
kubernetes
master
and
the
Jenkins
test
is
failing.
It
has
been
for
some
time.
We've
done
things
like
extend
timeouts.
We
probably
need
to
get
in
and
add
some
more
troubleshooting
code.
A
Locky
was
the
one
who
was
last
working
on
this
last
week.
Actually
he
was
poking
at
it.
So
he's
do
we
have
anybody
who
we
I
mean?
We've
got
a
few
things
we
can
do
it
so
I
mean
we
can
pull
the
Jenkins
test
out
until
we
get
it
fixed,
although
it
would
be
nice
to
have
it
fixed
and
start
putting
some
other
tests
in.
A
H
A
If
you
go
to
any
of
these
test
runs
and
you
go
over
like
a
red
spot
on
one
of
the
blocks
that
failed
and
you
click
on
it.
It
will
actually
open
up
to
the
run
for
it
and
you
can
see
the
artifacts
bill
blog
and
all
of
it
and
so
I
dropped
in
to
chat
here.
It's
the
UI,
for
this
is
in
Coober
native,
and
you
can
actually
see
by
the
way
you'll
see
they've
exposed
any
time.
The
word
error
is
used.
It's
a
really
simple
function.
A
A
Is
a
timeout
you'll
see
that
here,
they're
actually
using
we're
using
weight
and
it
times
out
you'll
actually
notice
that
helm
install
by
the
way
takes
a
very,
very
short
period
of
time
like
42
seconds
now
this
is
a
Java
app
and
when
we
run
it
in
our
poll
requests
it
takes
a
lot
longer
than
that
to
start
up,
but
in
this
cluster
it
starts
up
super
fast,
which
makes
me
wary
on
it.
Starting
up.
What
we
probably
need
to
do
is
update
this
to
add
more
both
debugging
when
something
fails.
A
G
C
It's
interesting
that
10-spot
is
closing
error.
We've
seen
this
on
our
cluster
a
lot
recently
in
connection
with
Sun
weight
so
actually
and
most
usually
the
the
helm
install
go
through,
but
right
in
the
middle,
the
transport
closes,
so
the
chart
installs
correctly.
The
releases
smart
is
pending
or
failed
and
the
transport
closes-
and
this
happened
since
we
switched
to
regional
clusters
of
Vietnam
which,
in
our
own
country,
ke
I'd,
never
seen
this
before,
but
just
for
quite
a
while,
since
the
early
regional
cluster,
this
happens
might
be
I.
Don't
know.
C
A
H
G
H
H
G
B
H
Reconnect
yeah.
A
We
probably
need
some
reconnect
logic
if
this
happens,
that
probably
wouldn't
be
a
bad
pull
request
against
he'll
try
to
reconnect
to
tiller
if
we
had
a
disconnect
of
some
kind.
Somebody
wants
biologic
now.
Is
this
G
RPC
connection
and
I'd
be
curious?
Why?
Because
G
our
PC?
Doesn't
it
have
the
reconnect
with
back
off
in
the
client.
B
A
A
E
A
E
A
A
There
was
a
pull
request
against
the
testing
to
add
some
better
output.
I'll
say
this
on
the
Jenkins
CI
test
failure,
but
it
actually
fails
under
a
whole
bunch
of
situations
it
passes
for
some
things
like
if
you're.
If
a
pod
only
has
one
container,
you
can
grab
pod
logs
and
it's
fine
because
it
sees
there's
only
one
container
and
it
does
it.
A
But
if
you've
got
two
or
more
containers,
it
throws
an
error
and
says
you
got
to
choose
the
container,
and
so
you
actually
have
to
look
at
each
of
the
containers
and
then
iterate
over
them
to
make
it
work
in
all
situations,
and
it
didn't
account
for
things
like
that.
So
we
haven't
gotten
feedback
on
the
pier,
but
there
is
one
to
get
some
better
test
information
and
we
probably
need
to
jump
on
that.
If
anybody
wants
to
get
in
the
loop
on
it,
it
would
be
appreciated.
D
D
D
G
D
B
B
A
H
D
May
also
want
to
align
the
scripts,
that's
doing
the
test
on
kubernetes
communities
with
the
same
one
that
we're
doing
NPR's
just
so
that
you
know,
if
we
add
debugging
capabilities
like
getting
the
logs
of
the
containers,
we
get
that
in
both
places.
So
we
may
want
to
talk
to
foxes
about
whether
that's
possible,
because,
right
now
it
does
it's.
It's
done
by
a
go
script,
which
is
in
the
repository.
A
A
C
A
Of
the
things
that
does
happen
with
the
current
script,
we'd
probably
need
to
update
it
for
here
and
just
make
it
similar
is
when
something
fails.
It
automatically
exits
all
future
testing
and
we
wouldn't
want
that.
The
current
tests
are
out
pretty
nice
because
the
way
that
they
do
it
the
way
they
do
it
tells
you
which
ones
pass
and
which
ones
fail
of
the
charts.
You're
actually
gonna,
see
a
is
it
the
go
file?
It
actually
creates
the
right
XML
files.
A
D
D
A
D
A
A
All
right,
so
we
do
have
two
more
things:
Oh
writing
it
in.
So,
if
we're
okay
with
this
one,
I'd
like
to
jump
to
the
next
one,
which
is
the
network
policy,
best
practices,
which
is
what
I
think
Joe
is
here
to
talk
about,
yep.
F
F
A
C
F
Don't
know
from
my
ops
perspective.
Putting
in
firewall
rules
for
an
application
has
always
been
part
of
the
application
installed,
so
I
mean
the
the
one
thing.
Is
you
don't
necessarily
have
to
have
network
policy
for
a
church?
If
it
doesn't
make
sense?
Obviously
you
wouldn't
be
required
to
just
have
an
empty
Network
policy
or
something.
The
other
thing
is
that
the
practice
that
I'm
advocating
has
just
like
we
do
with
our
back
there's
an
enabled
flag.
So
you
can
say
no
I
don't
want
the
charters
myself
yeah.
E
F
H
F
F
A
This
is
also
one
of
those
things
where
helm
is
a
package
manager
and
this
kind
of
walks
the
line
of
configuration
management,
which
is
a
tangental
issue,
and
so
we
should
be
careful.
You
know:
mixing
configuration
management
with
package
management.
I
just
put
that
out
there
to
consider
as
we're
doing
this
all
right,
thanks
for
getting
involved.
H
There
was
a
question
about
staple
sets
and
why
those
were
removed.
There
was
someone
answered
it,
but
it
wasn't
very
clear
and
I
think
the
question
was
yeah.
Why,
ultimately,
or
is
that
something
that
we
just
chalked
up
to
say?
Okay
sounds
good
enough,
or
does
anyone
really
understand
what
the
problem
is
there
I
don't
know
if
we
have
time
for
that
here.
A
H
E
A
G
A
A
H
A
A
A
G
I
can't
approve
my
own
thing,
which
I
think
is
probably
a
good
thing,
but
I
don't
want
to
have
to
bug
you
guys
so
I've
been
talking
to
people
about
getting
involved
and
they're
interested,
but
I
was
talking
to
Matt
actually
on
the
side
and
they
need
to
be
part
of
the
Kerber
Nettie's
organization.
Is
that
is
that
kind
of
just
a
paperwork
thing?
Or
do
we
really
want
people,
an
organization
that
are
strong
contributors?
G
A
Want
people
in
the
organization
who
are
trusted
so
so
the
big
reason
to
get
people
to
be
part
of
the
organization
for
these
purposes
is
because
we're
not
running
our
CI
system,
so
there's
charts
or
we
use
circle
CI
and
some
other
things
use
Travis
and
stuff
like
that.
But
kubernetes
operates
its
own
CI
system
and
that
CI
system
is
built
on
top
of
kubernetes
clusters
that
we
stand
up
and
manage
some
of
them.
We
stand
up
and
tear
down
many
times
per
day.
B
A
A
trust
issue
are
these
people,
contributors,
reviewers
people
who
are
engaged
in
part
of
the
project
and
do
we
trust
them?
And
if
that
is
the
case,
then
we're
okay,
giving
them
access
to
things
like
our
CI
system,
so
that
their
stuff
automatically
runs
without
needing
it
a
test,
and
they
can
be,
you
know,
being
owners,
files
and
merge
stuff
in
because
they've
proven
themselves
to
be
trustworthy
and
they're,
actually
contributing
and
involved.
So.
G
G
Don't
know
he's
he's
with
cloud
posse,
I
don't
know,
I
can
I
can
look
into
it,
but
I
just
I'm
wondering
what
the
level
of
technicality
is
for
a
membership
MultiJet.
A
G
A
Org
is
proper
owned
by
kubernetes
as
a
whole
at
the
moment,
so
it
is
just
kubernetes
as
an
organization
not
as
a
github
org
but
as
an
organization
owns
a
whole
bunch
of
github
orgs,
and
that
is
one
of
them.
And
so,
if
you
point
off
to
there
that
should
qualify.
And
if
anybody
complains
about
it-
and
let
me
know.
A
Going
once
going
twice
going
three
times,
thank
you.
Everyone
have
a
wonderful
week
and
remember
that
next
week,
in
the
same
timeslot,
we
have
our
office
hours.
So
if
you
just
have
general
questions
or
you
want
to
come
along
for
the
ride
and
help
us
review
charts
or
get
into
the
whole
reviewing
and
try
and
squash
down
the
PRQ,
please
same
time
next
week,
we'll
be
back
with
that
focus
it's
office
hours
and
for
anybody
to
come,
ask
questions
and
talk
to
us
or
to
help
us
review
stuff,
otherwise
have
a
wonderful
week.