►
From YouTube: Octant Community Meeting - May 26th, 2021
Description
Octant community meeting is held weekly. We discuss and talk about the current state and future of Octant, demo upcoming features and releases, and preview new ideas we are considering for Octant.
Meeting agenda: https://hackmd.io/CzaPxtmXT_SW8nEpdwvGzw?view
A
A
C
All
right,
it's
10
a.m,
so
we'll
get
started
so
hello.
Everyone
welcome
to
the
may.
26Th
I've
been
community
meeting
now.
This
is
time's
flying
by
so
pretty
much
end
of
may
now,
and
I
think
we
have
a
short
agenda
today.
So
I'm
just
gonna
go
and
start
sharing
my
screen
and
we'll
get
started.
C
So
what
are
we
looking
at?
The
first
item
is
an
item
that
I
put
on
the
list
on
the
agenda
and
for
anyone
who
doesn't
have
this
link.
This
is
the
hack
and
d-link.
It's
also
the
same
agenda
link
that
you
can
find
on
the
slack
channel,
and
this
is
open
to
editing
for
everybody.
So
if
there
are
any
items
to
be
added
for
this
agenda,
just
feel
free
to
throw
it
in
just
quick
note
again,
there
is
also
a
youtube
stream.
C
So
if
the
zoom
meeting
is
inconvenient
or
if
you
just
don't
want
to
use
zoom
for
any
reason,
you
can
just
use
the
stream
link
instead
cool.
So
we
have
this
first
item
here
is
the
re-prioritize
npm
error.
So
this
has
been
a
problem.
That's
been
bugging
us
for
a
really
long
time,
really
just
sir
okay
logs
no
longer
available,
but
essentially
we
have
this
npm
error
and
it's
not
very
clear
on
why
this
is
happening,
but
it's
responsible
for
a
lot
of
our
flakiness
and
testing,
especially
around
anytime.
C
We
do
npm
related
things.
I
think
we
see
it
more
on
mac
os,
but
it's
also
possible
that
it's
happening
on
windows
as
well,
and
it's
certainly
a
piece
of
the
tool
chain
that
we
want
to
go
away
because
it's
really
annoying.
It
doesn't
look
great
and
we
running
to
us
so
yeah.
So
this
is
an
example
of
what
it
looks
like
at
first.
We
thought
it
was
a
caching
thing,
so
we
actually
verify
the
cache
and
well
then
we
deleted
the
caching
and
see
our
environment.
C
So
now
there
is
no
caching
whatsoever
and
we
still
see
the
same
error
so
we've
rolled
out
caching,
so
the
next
approach
might
be
some
sort
of
npm
or
node
version
mismatch.
So
I'm
I'm
reading
about,
I
guess
enforcing
engines
in
the
package
in
the
package.json
file.
So
hopefully
that
resolves
it
so
we're
trying
this
out.
C
I
think,
as
we
discussed
before,
this
is
just
really
annoying
to
solve,
because
github
actions
don't
have
local
test
runners,
so
the
best
way
to
go
about
this
is
just
to
push
code
a
lot
to
iterate,
and
so
that's
just
an
unfortunate
part
of
the
situation
here.
But
I'm
going
to
reprioritize
this
I'm
going
to
it's
kind
of
been
around
for
like
the
last
three
four
months,
and
I
really
want
to
get
this
away
or,
if
not
just
modify
the
tool
chain
so
that
we
just
work
around
this
issue.
C
But
I
don't
like
this
is
getting
the
point
where
it's
not
acceptable.
So
I'm
gonna
move
this
into
the
spring
cool.
Any
questions
about
that.
C
Cool
no
questions
we
will
move
on
to
the
next
item.
Auction
are
back
in
informers,
so
this
one
is
another
older
issue
and
it's
really
the
thing
that
I
personally
have
been
working
on
the
majority
of
the
sprint.
C
But
the
idea
is
that
if
you
are
running
object
on
a
restricted
r
back
environment,
so
like
you're
using
a
service
account
that
can
only
view
certain
resources
instead
of
being
able
to
have
like
clustered
list
permissions
you'd
end
up
with
a
screen
that
looks
like
this,
where
you're
pretty
much
like
seeing
an
empty
octane
dashboard,
and
that's
just
not
a
great
experience,
and
this
has
like,
depending
on
which
version
of
object
you're
using
you,
would
be
able
to
see
certain
things.
C
But
I
think
we've
changed
the
way
that
the
how
we
basically
talked
to
the
kubernetes
cluster
using
informers
over
the
like.
I
think,
like
over
the
past,
like
several
months
so
and
I
think
like
if
you
look
at
the
history
of
the
dynamic
cache.go
file,
we've
rewritten
that,
like
maybe
three
or
four
times
so
this
is
going
to
be
something
that
we're
gonna
fix.
C
C
If
you
can
check
for
permission
to
have
that
resource
which
ends
up
slowing
you
down
so
sort
of
like
this,
like
speed
versus
accuracy,
trade-off,
I'd,
say
if
that's
accurate
and
I
think
we
we
have
a
pretty
decent
solution
like
we
figure
out
like
one
like
what
are
the
bottlenecks
within
the
client
go
informa
code,
but
also
what
we
need
to
do
to
avoid
hitting
these
are
back
issues.
C
C
I
know
those
a
pretty
generalized
way
of
talking
about
it,
but
if
anyone's
interested
in
like
the
nitty-gritty,
we
can
also
talk
about
that
either
offline
or
slack
channel
or
it's
some
other
point.
C
D
I'll
just
add
a
little
bit
more
in
saying
that
each
iteration
that
we've
done
on
the
the
dynamic
cache
has
surfaced
the
errors
in
kind
of
a
different
way
and
let
us
get
a
little
bit
more
introspection
into
what
is
actually
causing
the
issue.
So
now
we
have
this
error
message
coming
out.
That's
like
very
clear
of
like
this
service
account
is
not
allowed
to
start
a
watcher
for
this
resource,
instead
of
just
kind
of
like
a
random
failure
here
or
a
random
failure
there
for
a
pvc
or
or
some
other
resource.
D
So
that's
why
sam
was
able
to
actually
get
in
there
and
determine
a
a
good
solution
to
to
finally
resolve
this
error.
That's
been
kind
of
going
through
different,
like
the
error,
has
always
been
there,
but
the
messages
that
that
were
getting
spit
out
by
octan
kept
changing
as
we
changed
or
as
we
like
tried
to
patch
over
it
with
a
little
thing
here
and
there.
So
it's
nice
to
finally
get
good
clear
error
messaging
around.
Why
I
can't
do
things.
C
Yeah
totally,
I
definitely
agree
about
that
right.
We
can
even
see
like
this
original
pr
message.
Object
kind
is
missing.
An
unstructured
object
has
no
kind
like
that's,
not
really
an
actionable
thing,
so
certainly
a
lot
getting
better
and
I'm
we're
pretty
close
so
where
I
think
we're
gonna
yeah
we're
we're
getting
pretty
close.
I
I'm
pretty
sure
we're
gonna
get
it
right
this
time
around.
C
Cool
all
right:
well,
we
will
jump
on
or
any
other
questions
comments
concerns
about
this
one.
Oh
okay.
Well,
it
is
a
complex
and
ongoing
thing
and
like
we're
also
waiting
for
some
upstream
work
as
well,
so
we'll
make
a
pass
here
and
hopefully
in
the
future,
if
this
will
just
get
easier
to
do
and
if
not
there's
also
some
talk
about
us
potentially
building
our
own
watchers,
but
we'll
leave
that
for
another
day,
okay,
open
q,
a
let's
see:
why
is
resource
viewer
graph,
not
a
standalone
component?
B
Yeah
I
I
put
that
in
I
was
wondering
you
know
why.
While
I
was
working
on
the
on
the
some
of
the
lost
prs
and
on
the
resource
viewer,
especially
on
the
things
that
are
kind
of
plugging
others
facing
it
seems
that
resource
geograph
would
be
a
good
candidate
to
be
standalone
component.
I
know
there
are
some
plugins
out
there
and
k
native
might
be
the
only
one
currently
using
the
the
the
graph
and
it's
it's
a
little
awkward,
and
I
think
the
only
reason
they
can
do.
B
B
That's
a
question
for
sam
and
wayne
that
had
more
inside
than
that
and.
D
Mainly
that
there
hasn't
been
a
huge
demand
for
it
to
be
a
standalone
component,
and
since
it
like,
why
make
a
public
api
that
we
then
have
to
start
maintaining
and
conversioning
and
deprecating
when
there
is
no
incentive
to
do
so
is
mainly
is
mainly
it
there
was.
There
was
a
request
a
while
back
from
like
the
the
tecton
plug-in
they
were
wanting
to
display
some
some
graphs
or
some
some
flow
charts.
D
I
should
say-
and
ultimately
they
ended
up
using
the
the
graphics
plug-in
to
do
that
or
the
graphics
component
to
do
that
right,
so
they
just
kind
of
and-
and
so
I
think,
I
think,
plus
we've
added
some
other
new
components
that
address
some
of
the
specific
use
cases
of
why
someone
might
want
to
use
the
resource
viewer
component
like
timeline.
D
Is
there
now
and
some
other
things
that,
like
weren't
there
before,
that
someone
might
have
wanted
to
like
co-opt
and
use
the
the
resource
viewer
for
just
because
it
drew
lines
between
boxes?
And
so
I
think,
there's
no
reason.
It's
not.
But
the
main
reason
it
isn't
is
because
until
we
were
getting
like
formal
requests
to
make
it
a
component,
we're
just
gonna
keep
developing
it
and
being
able
to
modify
it
and
change
it.
D
D
I
think
we're
approaching
the
point
at
which
the
resource
viewer
api
itself
could
be
well
exposed
like
how
to
add
an
edge,
remove
an
edge
things
like
that,
and
so
as
long
as
we
get
the
abstraction
right
and
and
we're
not
exposing
any
of
the
implementation
detail
that
we
are
likely
to
change
underneath,
then
I'm.
B
Yeah,
I'm
for
it
cool
that
makes
sense.
I
was
just
wondering
yeah
eventually,
I
mean
obviously,
if
there's
a
more
interest
in
community
we
can.
We
can
definitely
do
that.
I
I
feel
for
everybody
as
a
plugin
developer,
who
wants
to
use
that
would
definitely
be
huge
help.
So.
D
D
The
word
customize
is
option,
so
yeah
being
able
to
create
that
would,
I
think,
would
be
really
useful
and,
like
you
said,
the
canadian
plugin
is
able
to
do
it
now,
because
they're
javascript,
plugin
and
so
they're,
just
inserting
the
json
blob
manually
into
their
content
response.
D
But
it's
yeah,
I'm
not
yeah,
I
think
creating
an
issue
if
we
don't
already
have
one
to
make
it
a
standalone
component
and
getting
it
into
reference.dev
is
something
we
can
try
to
prioritize
in,
probably
not
the
probably
like
in
a
release
or
two
like
just
as
something
we
can
put
in
there
to
get
to
because
yeah
do
you
think
it'd
be
useful
awesome.
Thank
you,
yeah
I'll
open
the
issue.
C
Awesome:
well,
that's
any
other
follow-ups
for
questions
about
creasersphere
and
just
being
a
consumable
component.
If
yeah,
I
don't
really
have
too
strong
of
an
opinion
on
this
one.
I
think
it's
a
great
thing
to
have
that.
The
probably
the
interesting
piece
here
would
be
if
someone
were
to
use
the
resource
vr
as
just
an
arbitrary
graph.
What
would
they
be
graphing
if
not
a
kubernetes
object,
and
that
could
lead
to
some
very
interesting
spaces
right
like,
for
example,
cicd
workflows,
but
then
that
is
met
by
graphis,
so
we
do
have
some.
C
We
do
have
some
support
for
that
type
of
line,
but
maybe
there's
something
that
a
resource
viewer
could
do.
Potentially
that
isn't
covered
yet
so
I'm
all
for
it.
D
Yeah
and
I
think,
as
we
get
more
people
using
it
like
that,
it'll
probably
expose
more
interesting
feature
requests.
I
know,
like
the
direction
of
things
has
been
something
that
has
been
talked
about,
which
is
in
a
generic
like
doing
that.
Generically
is
very
difficult,
but
if
someone
was
going
to
create
like
a
like
a
bpf
plug-in
and
show
like
network
and
routing
stuff,
if
the
resource
viewer
was
available
to
then
like
have
show
direction
right
and
flow,
they
would
be
able
to
model
that
so
so
yeah.
C
Cool
cool.
Thank
you.
That's
that's
an
awesome
exploration
of
this
so
yeah.
Please
do
open
an
issue
for
this
and
we
definitely
have
to
capture
this
and
get
this
prioritized
all
right.
So
does
anyone
else
have
any
other
additional
q
a?
I
have
one
more
thing
that
came
to
mind.
This
is
just
more
of
a
housekeeping
thing,
but
I
think
we
bumped.
If
we
go
to
our
projects
page,
we
bumped
out
our
release
to
june
1st.
C
So
a
quick
heads
up
from
I
guess
it
just
gets
pumped
out
of
except
I'll.
Add
it
here.
D
C
D
Ignore
that
never
mind
nothing
to
see
here
also,
do
we
want
to
introduce
liam.
C
Oh
yeah
sure
yeah.
We
we
have
an.
We
have
an
intern
who's
going
to
be
working
on
the
octane
project
for
next
12
weeks.
Would
you
like
to
introduce
yourself.
E
Yeah
sure
so,
hi
everyone,
I'm
liam.
I
am
working
out
of
a
dc
suburb,
specifically
in
northern
virginia,
I'm
a
current
junior
at
wpi,
and
I
have
a
background
in
mostly
like
you
know,
a
front-end
web
development
with
node,
but
I've
recently
been
taking
some
classes
about
distributed
systems
and
go
so.
This
is
really
cool
chance
for
me
to
learn
a
lot
more
about
something
I'm
not
too
familiar
with.
E
Specifically,
I
I
am
interested
in
you
know
like
optimizing
code,
making
things
run
fast
on
hardware,
those
types
of
things,
and
I
think
I'm
going
to
be
learning
a
lot
about.
E
You
know
how
you
can
use
kubernetes
or
kubernetes
as
supposed
to
be
pronounced
to
do
so
so
yeah,
looking
forward
to
working
with
you
all
and
meeting
everyone
a
bit
more.
C
Twice:
okay!
Well,
if
no
one
has
any
other
questions,
I
think
we
can
call
that
today
and
we'll
see
y'all
next
week
next
time,
all
right
later.
Thank.