►
From YouTube: Octant Community Meeting - April 29, 2020
Description
April 29, 2020
Agenda
Editing PR #885
Terminal streaming update #887
More logging improvements #879
The duality of Octant
Electron experimentation continues, branch using electron-builder created.
Identifying work that progesses both the Electron application and in-cluster forward is continuing.
TENTATIVE Octant Hackathon
June 17th - 23rd
Open Q&A - add your questions here
How is YAML editing validation going to work?
Open API v2
Inline validation integration with Monaco
A
B
Yeah
sure
so
a
few
things
to
talk
about
today,
one
of
them
is
pretty
neat
Bryan,
put
up
a
PRF
to
allow
some
simple
editing,
so
the
a
while
back,
we
introduced
the
Monaco
yeah
Mille
component
as
a
read-only
component
for
the
yamo
view
and
Bryan
has
gone
through
and
and
done
some
of
the
changes
necessary
to
enable
editing
of
resources
using
that
yellow
component.
It's
it's
similar
to
the
first
pass
that
I
talked
about
at
the
last
agenda.
B
Where
it's
a
simple,
you
know
it
enables
the
editor
to
be,
and
it
only
you
or
edit
mode
and
then
there's
an
update
and
a
cancel
button
at
the
bottom.
That
allows
you
to
submit
your
changes
and
yeah.
It's
I'm.
I
was
actually
testing
it
out
this
morning
and
it
and
it
works,
works
pretty
well,
so
that's
that's
gonna
be
landing
in
in
master
very
soon.
It
currently
doesn't
support
custom
resources,
but
that's
something
we'll
be
looking
into.
B
It
also
doesn't
have
any
of
the
smart
editing
capabilities
that
would
that
we
want
to
have
long
term,
but
this
is
a
it's
a
step
in
the
right
direction
and
it
gets
us
moving
towards
that
goal.
So
this
is
a
great
great
PR
that
Bryan
put
up
and
yeah
I'm
really
excited
about
this
feature.
So
you
got
anything
to
add.
Brian
yeah.
C
So
it
does
support
CR
DS.
We
were
drawing
the
yamo
for
CR
D
separately,
but
Machane
should
I
put
in
last
week
where
I
consolidated,
how
we
draw
the
tabs
realizing
we're
only
using
one
yellow
thing.
So
it's
nice
work,
Oh
excellent
I,
didn't
realize
that
until
I
was
talking
to
Sam
yesterday,
oh
my
god,
yeah.
B
B
C
B
So
this
change
now
allows
it
to
function
in
an
in
cluster
environment.
The
terminals
themselves
operate
independently
when
you
open
new
browser
tabs,
there's
no
more
muxing
of
the
session,
and
we
also
changed
how
we
actually
it
stream
the
content
and
and
connect
and
disconnect
from
the
terminal
sessions.
So
it
now
follows
a
similar
pattern
that
we
were
using
with
with
the
the
new
logging
implementation.
So
this
is
a
great
change
that
we
wanted,
but
we
also
need
for
in
cluster
work,
so
that
is
we'll
be
landing
soon.
B
B
And
then
the
next
one
is
work
that
Mulan
has
done.
This
is
all
around
the
logging
improvements,
so
we
released
the
new
log
in
UI.
Folks
gave
us
a
bunch
of
great
feedback
and
Milan
jumped
on
all
of
that
and
and
implemented
I
think
pretty
much
all
of
the
the
requests,
and
so
now
it
it's
it's
pretty
great.
The
the
regex
expressions
they
support
the
the
filtering.
B
The
the
search
itself
is
now
by
default.
Case-Insensitive
timestamps
have
been
turned
off
by
default
since
most
things
in
the
the
most
things
that
are
logging
are
also
injecting
their
own
timestamps.
So
the
the
you're
still
seeing
those
timestamps,
but
we've
turned
off
the
timestamps
that
are
part
of
the
kubernetes
api
log
entry
for
the
that
that
pod,
so
those
are
they're
still
available,
but
the
others
now
off
by
default,
and
then
we
also
added
the
ability
to
well
mui.
B
He
also
added
the
ability
for
searching
and
filtering
on
the
container
names,
so
we
we
recently
added
a
feature
where
you
could
now
see
all
of
the
containers
logs
combined
and
in
the
viewer,
and
now
you
can
search
and
filter
on
that
when
you're.
Looking
at
the
all
view,
there's
still
a
drop
down
to
let
you
select
individual
containers,
so
this
is
more
applied
to
when
you're.
Looking
at
that
collection
of
all
of
the
container
logs
and
I,
don't
know
if
Milan
has
anything
to
add
to
that.
D
B
And
so
then
moving
so
these
three
things
are
being
reviewed
right
now
they
are
they're
all
they're,
all
looking
good
and
will
probably
be
part
of
the
the
nightly
that
that
gets
produced
tonight.
So,
if
you're
interested
in
checking
those
out
early,
they
will
they'll
be
there
any
questions
comments
about
any
of
those
items
before
I
move
on.
B
B
Certain
aspects
of
these
things
are
in
direct
competition
with
one
another.
They
they
don't
agree
on
on
certain
things
when,
when
you
do
them
so
we
are,
we
are
planning
and
focusing
on
the
items
that
overlap
both
of
these
areas
that
allow
us
to
enable
getting
both
of
these
features
completed
in
a
way
that
that
continues
to
progress
so
we're
working
on
these
issues
in
a
way
that
will
continue
to
progress
each
one
forward,
and-
and
most
of
that
is
just
around
things
like
allowing
octant
to
start
without
a
cute
config
present
right.
B
That
is
useful
for
the
application,
and
that
is
useful
for
in
cluster,
allowing
folks
to
then
provides
like
so
that's
the
startup
part
of
it
and
then
allowing
folks
to
then
provide
some
type
of
method
for
authenticating,
a
user
so
that
it
can
access
a
cluster
right.
That
is
that
that's
a
piece
that
will,
on
the
application
side,
we've
already
solved
right.
B
That's
that's
it's
a
later
date,
undetermined,
TBD
and
also
generally
like
what
the
recommendations
are
going
to
be
for
like
in
cluster
deployment
and
what
that
yeah,
mole
and
images
are
going
to
look
like
things
like
that.
That's
also
TBD.
This
is
more
about
the
architectural
and
engineering
work
that
needs
to
be
done
to
even
support
those
ideas
at
a
later
date.
B
B
You
know
just
like
just
because
we
got
electron
builder
working
and
spitting
and
exe
does
not
mean
it's
actually
ready
to
hand
off
to
people
to
release.
So
we
are
we're
in
constant
experimentation
with
that
and
and
progressing
towards
something
that
that
looks
like
we
want
and
same
with
the
in
cluster
work.
We
are
anytime,
we
find
an
area
that
needs
addressing.
We
address
it
with
respect
to
supporting
you,
know,
session,
based
work
and
removing
global
context,
and
things
like
that.
So.
A
B
B
C
Ryan
yeah
the
electron
app,
it's
mostly
me
being
selfish
I
want
it.
So
if
it
runs
an
electron
that
means
I
could
put
it
in
my
status
bar
on
my
Mac,
which
means
that
I
can
have
it
running
and
I
can
use
it
to
change.
Namespaces
and
I
can
put
other
interesting
things
like
summary
information
which
we're
keeping
put
it
right
up
there,
because
right
now,
I
have
it
on
my
my
left
screen,
taking
up
half
of
it
and
sometimes
I,
don't
want
to
see
that,
but
I
still
want
everything
that
oxide
provides.
B
And
some
of
the
cuts
so
like
one
of
the
things
is,
is
that
that
I
find
useful
about?
It
is
like
I
to
kind
of
go
along
with
what
Brian's
saying
a
friend
of
mine
he's
using
octant
that
they
work.
They
work
at
Red
Hat
and
they
used
oxen
to
look
at
kind
of
their
their
openshift
stuff,
and
the
thing
they
said
was
like
when
they
first
started
using
octant.
B
They
really
wanted
to
just
be
in
cluster,
but
now
that
they've
been
using
it
as
a
desktop
application,
they
could
they
like
they
definitely
don't
want
to
in
cluster
they're,
like
so
happy
that
it's
a
desktop
application
and
I
think
I
think
that's
kind
of
like
the
I
think
most
people
just
want
something.
That's
a
web
app
because
it's
easy
and
it's
in
their
brain
that
it
should
just
be
in
cluster.
B
B
So
that's
why
we're
still
doing
both
is
because
there's
there's
a
there's,
a
whole
class
of
people
that
want
it
to
be
a
desktop
application
and
find
it
very
useful
in
that
way,
and
then
there's
another
class
of
people
who
work
in
super
large
enterprises
that
have
don't
want
to
distribute
a
binary
to
all
their
employee
laptops,
but
want
everyone
to
use
octant
right
who
just
want
it
to
be
running
as
a
web
service
somewhere.
So
like
there's
the
need
for
both
and
we're
like
we're.
Gonna
address
both
of
them.
So
that's
what
that's?
B
E
A
very
romantic
use
case
that
makes
sense
with
browsers
where
you
want
two
different
views
open
at
the
same
time
and
you
may
have
to
re-implement
that
and
native,
but
just
have
two
tabs
or
two
windows
open
or
whatever
in
the
browser.
So.
C
Yeah
Tim
I
find
that
interesting
people
say
that
I'm
actually
looking
for
those
use
cases,
because
if
that's
a
use
case
where
someone
wants
multiple
things
open
at
once,
maybe
we
should
tackle
those
things
rather
than
saying
go
between
two
web
pages,
because
if
you
think
about
other
so
AWS
people
don't
generally
use
two
web
pages
or
Google
Cloud
console
people
are
not
using
two
webpages
there.
So
maybe
that's
a
that's
a
UX
thing
that
we
should
tackle
and
and
try
to
approach
it.
That
way.
That's.
E
Fair,
maybe
on
create
you
know.
Maybe
I
have
a
name
space
over
here
that
got
built
a
long
time
ago.
That's
got
a
lot
of
custom
stuff
and
I'm
trying
to
bring
it
up
to
snuff,
with
them
with
a
more
standard.
So
I
want
two
different
namespaces
in
two
different
windows
or
two
different
tabs
and
I
really
just
want
to
visually.
Compare
them.
That's
a
use
case.
I
just
came
up
with
just
now,
I'm,
not
saying
I.
E
D
D
C
So
this
is
this
is
the
trick.
It
was
a
reason
why
I
pitched
for
Monica,
not
because
it
was
a
good
editor,
because
I
knew
what
Monaco
could
do.
One
of
the
things
that
makes
a
vias
code
so
powerful
is
its
ability
to
to
be
an
editor
for
anything
and
the
basis
for
that
is
actually
Monaco,
and
so
what
we
can
do
is
we
can
actually
generate
what
a
so
if
we
know
what
an
object
is,
I
can
actually
and
I
know
if
I
know
what
an
object
is.
C
I
know
what
the
properties
using
opening
API
v2,
what
the
object
is
I
can
actually
build
a
tool
that
or
I
can
build
something
that
says
hey.
This
is
the
valid
I
can
say.
These
are
the
valid
values
and
I
can
actually
show
that
it's
invalid
right
in
line,
and
so
that's
that's
the
goal.
I
haven't
done
any
of
that
work
yet,
but
I've
done
it
all
and
I've
done
it
all
separately
in
other
projects.
So
I
know
that
it
should
work.
C
C
So
but
it's
a
good
question
because
it's
like
you
know
and
that's
something
that
that
I
was
I.
I
was
talking
with
Wayne
about
earlier
this
week,
where
I
said
the
reason
I
put
that
editor
in
previews,
because
just
creating
and
editing
objects
is
is
a
thing,
but
when
we
get
in
and
there's
a
reason
why
I
did
not
create
I
created
the
editor
before
I
could
have
to
create,
because
there's
more
work
than
you
be
done
for
the
create
we
should
not
just
say.
C
Oh
here's,
an
empty
editor
here,
go
take
your
ya
mole
and
that
won't
be
great.
So
we
need
to.
We
need
to
approach
that
with
being
able
to
upload
a
piece
of
ya
mole
from
your
local
computer
or
if
we
know
that
you're
creating
a
deployment
cube
control
has
templates
in
there.
That
gives
you
the
basics
for
creating
a
deployment
and
thirteen
other
items.
I
think
I
counted
just
like
13
other
items.
C
So
we
can
get
those
for
free
and
the
ones
that
we
don't
have
like
for
staples
Sarah
stable,
set
replica
set
all
the
other
workloads
that
we
support
and
everything
else
is
on
the
sidebar.
Well,
it'll
just
create
defaults
and
we'll
do
it
the
same
way.
So
for
every
single
thing
that
we
can
besides
CR
DS,
we
can
just
say:
hey
I,
want
to
create
this.
C
We
fill
in
most
of
it
for
you
and
if
it's
something
it
hasn't
work
load
and
then
you
just
fill
in
the
maybe
the
most
you
to
change
is
a
container
image
and
that's
where
I
would
like
people
to
be,
and
then
you
can
upload
the
rest.
So
I
get
you
most
of
the
way
there
and
then
we
can
see
how
people
are
using
it
or
how
and
how
we're
using
it
and
see
if
that
actually
satisfies.
C
So
that's
that's
the
theory
there
and
then
other
were
something
else.
I
wanted
to
bring
that
so
I
was
talking
to
Sam
yesterday
and
he
brought
up
that
the
team
is
thinking
about
the
resource
viewer
and
this
is
another
piece
where
we'll
have
to
tread
lightly,
because
one
of
the
one
of
the
most
interesting
parts
of
octant
to
me
and
I
actually
gave
Sam
a
demo
of
this
yesterday
about
how
I
use
resource
your
and
why
all
the
things
are
in
there,
because
it's
hard
to
understand.
C
If
a
user
wants
to
see
everything
or
they
want
to
see
like
a
trim
down
view,
we
really
can't
make
it
so
wait
it's
easier
to
show
everything
and
then
be
able
to
pare
that
back
and
to
start
with
nothing
and
then
scale
it
forward,
because
what
are
you
yelling
it
to
so
this
will
be
an
interesting
I.
Think
one
of
the
biggest
pieces
for
oxen
going
forward
will
be
more
about
user
experience
and
actually
figuring
out.
C
How
do
we
engineer
a
better
experience
on
kubernetes,
not
just
like
features
and
what
not
just
like
but
I'm,
sitting
down
I'm
using
this
thing,
and
how
does
it
make
me
better
at
whatever
else
I
have
to
be
doing
so
that's
a
something
we
need
to
think
about,
and
it's
funny
I'm
wearing
my
oxide
shirt
today.
I
think
the
plan
will
be
on
this
call,
but
I'm
here
and
I'm
wearing
an
octant
shirt
so
serendipitous
or
once
it
ends
I,
don't
know.
A
F
C
Yes,
yes,
it
has
been
investigated,
it
has
not
been
solved
yet,
but
we
know
what
the
issue
is
and
just
for
people
in
this
call
I'll
explain
to
you
a
little
bit
about
how
octant
does
what
it
does.
So
we
first
started
octant
almost
two
years
ago,
I
had
this
idea
of
using
Informer,
so
we
could
be
so
we
could
catch
this
information
on
your
local
machine
and
we
could
have
this
data
available
to
you
almost
instantly,
because
it's
in
memory,
then
you
know
we
move
forward
and
on
a
development
cluster.
C
This
is
perfect,
but
then
you
move
into
people
who,
when
you
start
using
this
on
production
clusters
and
and
we
don't-
you
can't
actually
say
what
the
state
of
permissions
are
going
to
be.
So
we
can't
assume
that
you
have
permissions
to
see
things
in
the
cluster,
so
we
have
to
do
off
checks
and
what
happened
is
the
amount
of
off
checks
that
we
were
doing
is
actually
I,
guess
on
a
larger
cluster
or
on
a
cluster,
with
a
lot
of
objects
that
it
can
actually
strain
that
controller
and
that's
what
happened
now?
C
The
fix
for
it
is
no
fix
right
now,
but
neither
a
reason
about
why
that's
happening
and
then
can
we
do
it
less.
One
thing
that
kubernetes
does
not
provide
us
is
a
way
the
changes
tell
us
when
off
changes.
So
we,
oh,
we
don't
know
you
have
to
pretty
much
check
all
the
time.
Every
time
you
view
an
object,
you
have
a
check,
you
can't
just
say
even
if
I
wish,
I
could
even
say
hey
for
the
next
five
minutes.
C
Give
me
this
token
and
I
will
use
this
token
and
it
will
expire
and
I'll
and
I'll
reset
it.
So
you
can't
do
that,
but
I
think
that
as
part
of
our
solution
that
actually
you
mentally
invent
that
ourselves,
where
we
just
go
and
we
register
short-lived
tokens
per
clients,
so
we
can
say
that
a
client
I'm,
a
client
will
live
for
a
minute,
and
during
that
minute
we
won't
deal
any
off
and
when
that
token
expires,
we'll
get
a
new
client
and
we'll
do
off
again.
C
So
we
will
spread
out
the
amount
of
off
calls
over
time,
rather
than
doing
them
every
time.
So,
for
example,
if
you
look
at
the
home
page
for
like
octant
and
you
have
like
I'm
just
looking
at
one
right
now-
I
have
two
deployments:
there's
seven
pods
two
replicas
sets
and
a
job.
We
do
a
lot
of
Hoth
calls
so
for
every
single
item
on
there.
We're
gonna
do
at
least
one
article
and
I
think
we
can
do
better
than
that.
F
C
Actually,
that's
a
if
we
when
we
run
in
cluster,
we
won't
have
this
problem
at
all
because
we'll
be
using
individual
client
and
whenever
the
client
gives
you
a
not
found
it's
easier
for
us
since
we're
going
directly
to
a
client
not
using
the
informer,
we
won't
actually
have
this
issue.
This
is
only
because
of
the
informers
got
it
yeah.
F
C
C
C
F
Tiddly
team
did
mr.
previous
meeting.
Has
anybody
talked
about
octa
son
or
the
octan
hackathon,
and
some
of
the
tentative
dates
that
we
had
on
that
yeah?
No
okay.
So
we
have
some
tentative
dates
who
have
on
that?
Maybe
if
I
know
you
want
to
stop,
maybe
put
some
blocks
in
your
calendar
is
June
and
again
this
is
sensitive
is
not
fully
committed,
but
this
June
17
and
of
day
June
23rd
and
we
can
go
add
a
plus
or
a
day
so
can
go
18
to
24,
for
example.
That's
an
option
as
well.
F
F
There
is
no
we're
having
it
spanning
week
is
to
give
folks
a
couple
of
business
days
and
a
couple
of
weekend
days
for
the
folks
that
can
take
time
off
work
to
contribute.
Then
they
can
do
that
and
the
folks
that
really
can't
do
that
didn't
work
day.
They
can
find
that
we
can
to
basically
play
around
a
little
bit
with
it.