►
From YouTube: 2021-08-04 meeting
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
Hi
richard,
like
your
new
haircut
new
style,
thank
you
was
this
part
of
the
the
off-site
you
guys
had.
B
No
gotham-
and
I
had
planned
to
do
this
for
like
two
or
two
and
a
half
years
now
and
he's
going
to
visit,
so
we
we
decided
to
go
all
in
wow.
That's
awesome.
A
Hi,
sorry,
I
bounced,
I
think,
my
connection
bounced,
but
but
richard
what
I
was
saying
is
richie
that
you
have
the
grafana
colors.
So
all
color
coordinated.
B
Depends
in
this
in
this
brightness,
for
depending
on
like
you
need
to
the
hair,
gets
quite
soft
after
you're
through
with
the
whole
process.
So
it's
not,
it
doesn't
get
fatty
very
quickly
and
ideally
you
don't
wash
it
when
you
shower
like
or
only
really
really
need
to,
and
if
you
do
this,
you
can
retain
that
amount
of
color
for
three
to
six
days.
Maybe,
and
then
it
starts
fading
slowly.
B
C
Very
cool:
is
he
around
yes
he's
visiting
from
berlin.
A
A
Let's
get
it
started
yeah.
I
think
we
have
about
eight
folks
now
and
let's
take
two
minutes
to
add
an
agenda.
A
I
think
so
I
think
there
are
a
lot
of
folks
out
today,
but
what
we
wanted
to
do
is
you
know,
at
least
for
the
couple
of
projects
we've
been
working
on.
Do
a
demo
of
you
know
our
progress
and
the
walkthrough
of
the
target,
allocator
updates
that
we
have
done.
A
D
A
A
D
D
So
we
have
actually
submitted
two
pull
requests
as
of
now,
and
I
think
he
has
started
reviewing
the
first
one
and
the
so
first
one.
Basically,
the
logic
remains
the
same.
The
only
additional
logic
is
that
we
had
added
a
roll
and
roll
binding
to
actually
help.
D
I
mean
because
it
uses
the
cluster
config
inside
the
image.
So
we
need
a
roll
and
roll
binding
to
actually
allow
it
to
watch
for
the
collector
parts.
So
that
was
added,
and
then
there
was
a
discussion
going
on
if
it's
actually
needed
or
not
so
that
so
then,
after
that,
I
think
jurassic
is
fine
with
having
it
in
the
code
or
it
just
the
cluster.
I
mean
if
it's
a
role
and
role
binding,
not
the
cluster
rule
and
trust
full
binding.
So
it
should
be
working
with
that.
D
Hopefully-
and
that
is
the
only
change
which
was
added
before
the
comments,
but
then
yeah,
the
discussion
jurassic
was
kind
of
not
convinced
initially
and
then,
after
that,
he
kind
of
mentioned
that
he
wanted
it
and
the
other
changes.
Actually,
the
major
change
is
that
the
code
restructuring,
so
basically
a
lot
of
code
is
being
reused
now
from
the
collector
components,
so
jurassic
kind
of
wants
it
in
one
folder
a
separate
I
can
just
mention
it
somewhere.
A
That's
just
that's
just
a
change
in
terms
of
where
the
you
know,
module
sits,
but
are.
D
Logic
would
not
change,
but
then
there
would
actually
be
a
kind
of
a
refactoring
going
on
because
a
lot
of
code
I
mean
the
collector
and
the
target
allocator
both
will
be
using
a
lot
of
the
same
functions
right
now.
So
because
of
that
some
a
few
things
will
change,
and
then
the
testing
also
will
change
a
little
but
logic
wise.
Nothing
is
changing
actually,
so
that
is
another
change
which
he
wanted.
Then,
apart
from
that,
it's
just
a
few
small
changes
here
and
there,
such
as,
instead
of
actually
checking
it.
D
So
what
we
were
doing
earlier
was
that
we
were
checking
for
if
it
was
a
stateful
set
and
if
the
mode
was
enabled
inside
the
reconcile
function.
But
what
do
I
see
wanted
was
that
we
actually
do
not
create
the
cr.
I
mean
accept
this
er
itself,
if
it,
if
it
is
not
valid
so
in
in
the
initial
run,
stated
stage
itself,
we
kind
of
give
out
an
error,
so
small
changes
like
that
and
a
few
other
things
suggesting
quotes
and
all
that.
D
But
then,
yes,
those
are
the
two
major
changes
here.
What
do
you
mention?
Apart
from
that,
there's
also
so
the
second
pr
where
anthony
had
left
a
few
comments,
which
I
think
everything
is
addressed,
except
one
of
them,
where
it's
just
again
the
where
it
should
be
placed,
but
then
it
can
be
done
in
a
separate
year.
That's
what
he
mentioned,
but
yes,
once
joyce
also
reviews
it.
D
Maybe
we
can
decide
how
to
proceed
with
this
and
there
would
again
be
a
third
tier
as
well,
which
is
a
very
small
change,
but
we
have
not
submitted
it
yet
because
the
because
of
the
changes
in
the
first
year
it
would
actually
change
the
structure
of
the
third
pr
as
well.
So
we
are
waiting
for
the
changes
of
the
first
pr
to
be
completely
finalized
before
we
actually
go
forward
and
summon
the
third
because
of
the
changes.
D
So
yes,
that's
the
updates
regarding
the
target
allocator
from
myself.
D
Yes,
we
are
actually
so
there
are
a
few
comments
left
today
based
so
that
is
still
being
worked
on
and
mostly
it
should
be
done
by
today
or
tomorrow
morning-
and
apart
from
I
mean,
but
yes
other
I
mean
if
david
or
anybody
else
wants
to
review.
Obviously
that
would
be
great
because
more
the
reviews,
it's
better
yep.
A
A
A
You
know
algorithm
for
allocation
as
we
look
at
the
general
collector
service
discovery
mechanisms
and
you
know
align
it
with
what
what
we
are,
what
you
know,
other
algorithms
might
be
used
there.
That
is,
I
think
that
this
implementation
will
get
affected
by
that,
but
it
could
stay
the
same
too.
A
I
mean
it
works
for
the
basic
cases,
but
doesn't
necessarily
you
know
it
could
be
more
sophisticated,
at
least
in
you
know,
in
in
our
conversations
and
discussions
that
we've
had
and
and
vishwa
it
would
be
nice
if
you
guys
could
also
take
a
look.
I
mean
again
it's
it's.
It
works
in
this
implementation
and
and
for
all
the
80
use
cases
that
we
have.
A
But
again
you
know
it
would
be
nice
to
actually
get
some
other
suggestions
in
terms
of
what
are
some
of
the
other
use
cases
that
may
not
work
totally
right.
I
mean
we're
making
some
basic
assumptions
on
how
labeling
is
done.
How
allocation
you
know
has
done
queuing
is
done
and-
and
I
think
that
david-
this
is
not
it's
not
exactly.
You
know
what
you
had
in
mind.
Initially,
when
you
know
you
had
discussed
the
allocator
as
a
separate
component,
but
again
just
it
works
so
for
now.
F
D
So
I
mean
presently,
the
number
of
sta
stateful
sets
is
just
obtained
from
the
custom
resource,
what
is
defined
by
the
user
so
based
on
that
number
it
could
be
distributed.
D
A
F
Yeah,
yes,
that's
what
I
thought
you
know
to
start
with:
it's
gonna
be
manual
user,
specified
number
of
shots,
and
then
hopefully
you
know
we
can
do
some
work
to
auto
scale.
It.
A
F
D
Actually,
it's
in
both
of
these
are
in
separate
branches,
but
then
yes,
there's
no
interdependency
among
both
of
them.
I
think
yes,
both.
D
Yes,
so
it's
these
two
actually
so
presently
this
again
a
third
year
so
presently,
the
only
thing
which
would
which
can
be
viewed
is
how
the
targets
are
being
allocated
among
them,
but
the
collector
still
does
not
have
the
capability
to
pull
it
because
that's
the
third
year.
A
Okay,
yeah
yeah.
Are
you
almost
ready
rahul
with
the
third
pr.
D
Yes,
actually
it's
done
it's
just
that
because
of
the
first
part,
changes
yeah.
C
D
Structure
of
the
third
year
would
also
change,
not
the
logic,
but
then
I'm
just
waiting
for
the
first
pr
to
be
completely
approved
before
because
and
finalize
the
changes
again
because
things
such
as
function
names
or
some
things,
putting
small
things
like
that.
So
I
don't
want
to
push
the
pr
before.
A
Yeah
but
make
sure
you
I
mean
again,
given
that
you're,
you
know
you're
working
through
this.
Let's
make
sure
we
get
reviews
so
that
this
you
can
file
the
third
beer.
D
A
Okay,
so
sheba
do
you
want
to
go
next?
We
have
been,
as
many
of
you
may
already
know,
we
didn't
demo
in
the
collector's
sig
also,
but
wanted
to
kind
of
walk
through
we've
been
building
an
helm
chart,
we
added
a
helm
chart.
You
know
to
the
for
the
collector
and
the
operator
made
some
enhancements
in
the
operator.
A
As
you
know,
adding
stateful
set
support,
so
we
wanted
to
kind
of
run
through
and
full
pipeline
and
walk
you
through
what
we
have
done
with
the
hum
chart,
the
operator
and,
obviously
you
know,
deploying
multiple
instances
of
the
collector
on
eks.
So
shiba
did
you
want
to
share
your
screen.
E
Yeah
sure
so
so
first,
let
me
send
the
hem
chart
link
in
the
chat.
So
basically
this
is
the
hand
chart
we
have
built
for
the
operator
and
let
me
share
the
screen
yeah.
E
Okay,
so
let
me
work
through
so
do
you
want
me
to
walk
through
the
hamstring
first
or
do
you
want
me
to
yeah,
okay,
so
yeah?
So
so
this
ham
chart
is
for
the
gem
tree
operator,
as
I
have
proposed,
a
design
in
the
ultra
promiscuous
work
group
before
and
basically
the
users
could
first
install
this
hem
chart
in
their
eks
or
kubernetes
cluster
and
then
configure
their
own
characters
with
four
modes
to
instruments,
their
applications
or
their
eks
or
grenades
infrastructure.
E
Chart,
and-
and
basically,
as
you
can
see
here,
is
the
structure
of
this
hem
chart
like
all
the
configurations
yemos
and
we
have
oops,
and
we
have
tests
for
this
hamstring
to
test
the
service,
as
well
as
the
search
manager,
service
connections,
while
speaking
of
the
surrender
well,
just
as
open
to
entry
operator
requires,
we
use
the
server
manager
to
provide
a
self-signed
certificate
by
default
to
to
make
sure
that
to
make
sure
the
operator
has,
the
trusted
have
a
trusted
connection
chain,
one
one
connection
between
the
webhook
services
and
the
grenades
api
server
and,
as
you
can
see
in
this
well,
this
readme
also
contains
how
to
install
this
hem
chart
and
how
to
upgrade
it,
how
to
configure
how
to
configure
it
and
the,
and
the
important
part
is
how
to
configure
the
character
right.
E
So,
basically,
just
as
the
the
operator
does,
we
have
four
modes
to
deploy
the
character,
deployment
statement
set,
staple
set
and
cell
card
okay.
So
for
each
mode
I
have
fields.
I
have
written
some
configuration
examples
for
users
you
know
to
to
use
if
they
want
it's
just
very
simple
ones.
E
So,
okay,
so
basically
that's
it
for
the
ham
chart
structure
and
I
can
do
a
quick
demo
to
show
you.
The
entrance
user
experience
using
this
hand
chart
and
then
configure
the
connector
to
scrape
the
kubernetes
infrastructure,
and
I
will
also-
and
I
will
also
use
some
exporters-
to
export
the
metrics
data
to
amp
or
to
cloudwatch,
which
are
also
the
aws
services.
E
Okay,
okay,
I
will
do
a
quick
demo.
Okay,
that's
yeah!
So
currently
I'm
just
using
eks
cluster.
Well,
we
can
also
use
the
kubernetes
cluster
that
they
are
basically
the
same,
and
this
is
mycast
cluster
and
there
are-
and
I
have
also-
and
I
have
already
configured
the
star
manager
and,
let's
first
add
this
hand-
wrapper
from
the
open
to
empty
hand
repository,
and
we
want
to
update
it.
E
E
Kubernetes
cluster
well
well,
when
design
of
this
ham
chart
is
we,
we
will
install
the
operator
in
a
specific
namespace,
open
temperature
operator
system
and
we
have
to
code
it
to
to
keep
consistent
with
the
crd,
which
is
open.
Temperature
collector
yeah.
So,
as
you
can
see,
here
is
my
operator,
and
if
we
want,
we
could
see
the
the
logs
of
this.
E
We
will
just
choose
the
manager
container
and
well
at
this
point,
the
hem
chart
installs
the
the
operator
version,
which
is
29
and
the
latest
one
is
31.
Well,
we
will
update
it
and,
as
you
can
see,
some
message
is
saying:
it's
registering
webhook
and
it's
starting
webhook
server,
something
like
that
and
so
yeah.
So
because
the
operator
currently
is
waiting
for
us
to
configure
the
character,
and
once
we
have
configured
one
the
quantities,
api
server
will
just
notify
it
yeah.
E
Well
let
let's
do
one
to
configure
this
character,
and
I
have
one
example
you
well.
You
know
one
very
simple
example
here
yeah,
so
I
have
one
very
simple
example
to
configure
the
character
as
the
deployment
and
scripts,
the
metrics
data
from
my
kubernetes
infrastructure,
using
the
permissions
receiver
and
and
exports
the
metrics
data
to
amp
and
to
cloudwatch
using
the
aws
permissions,
remote
right
exporter
and
devs
emf
exporter
respectively.
E
Yeah,
so
so
here's
the
the
logs
of
this
character
and
everything
is
working
right
now.
I
guess,
and
exporter
started,
provide
a
receiver
started
pipeline
started
yeah,
so
it
should
be
scraping
the
matrix
yeah.
So
so
it's
working
right
now
and
we
can
we
can
validate
if
the
metric
theta
has
made
this
way
to
amp
autocal
watch
with
separate
assets,
so
yeah.
So
how
to
so?
E
I
use
database
curve,
which
is
a
kerala
tools
to
to
send
queries
to
my
amp
workspace
you
using
the
amp
api
and
accuracy
if
this
matrix
script,
simple
script
has
made
its
way
to
my
amp
and
she,
of
course,
use
jq
to
pretty
print.
The
results.
E
E
My
amp
workspace
and
also
I
can
verify
if
if
they
arrived,
my
cloud
watch
law
groups
yeah.
So
here
is
the
law
group
name
and
the
log
stream.
Let's
find
the
latest
one.
What
time
is
it?
It's
8,
27,
okay,
yeah.
So
so
here
is
the
matrix
in
my
device
called
watch.
E
A
Yeah,
yep,
and
and
and
this
is
very
cool
because
you
know
it
uses
the
standard
components-
the
prw,
the
prometheus
receiver.
You
know
with
the
enhancements
that
we've
been
adding,
as
well
as
the
premier,
the
prw
exporter,
so
you
know,
can
write
to
any
service
backend
which
is
remote
right
enabled,
and
it
really
is
a
general
pipeline.
So
you
should
be
able
to.
You
know,
simply
be
able
to
enhance
this
by
adding
configurations
for
any
kind
of
back
back-end
framework
that
you're
talking
to
yes,
any
questions.
A
A
A
Yeah
and
and
vishwa,
maybe
you
could
also
add
you
know
some
of
the
other
configurations,
because
I
think
it
becomes
very
useful
for
an
end
user
to
be
just
taking
this
and
setting
up
the
pipeline
and
being
able
to
run
with
it.
G
A
G
The
most
prominent,
the
most
prominent
is
configuration
reload
by
sending
a
sig
hub,
and
there
is
an
open
issue
that
jurassia
had
opened
a
while
ago
like
two
years
ago,
except
yes,
yeah.
Let
me
I'll
paste
it
I'll
paste
it
here
in
the
chat,
except
you
know
not
much
got
done
to
it,
but
it's
useful
if
we're
gonna
have
compatibility
with
prometheus.
I
have.
G
Issue
for
compatibility
with
prometheus
yeah,
that's
good!
That's
so
far
what
I've
been
looking
at
early
this
morning
and
then
yeah,
I'm
basically
going
down
that
entire
list
of
configurations
and
doing
parity
comparisons
with
prometheus.
A
Again,
this
is
the
larger
you
know.
One
of
the
goals
that
we
have
with
achieving
metric
stability
on
the
collector
is
to
make
sure
that
the
prometheus
receiver,
you
know,
is,
is
hardened
and
can
be
used
for
the
current
service
discovery
and
scraping
that
it's
doing,
but
also
you
know,
any
any
open
issues
that
are
still
pending
are
addressed.
G
A
G
Yeah,
it's
it's
the
core
ones,
but
you
know
I
was
just
tabling
that
because
it
doesn't
seem
to
be
known.
A
I
know
I
know
you
haven't
had
much
bandwidth
perhaps,
but
you
know
it
would
be
great
to
actually
to
get
some
more
feedback
on
some
of
the
use
cases
that
you're
using
prometheus
receiver.
For
you
know
we're
looking
starting
to
look
a
bit
more
deeply
on
what
else
can
be
optimized
on
the
receiver
so
again
again
welcome
feedback
from
anybody
else,
who's
dug
into
the
code
and
just
thought
about
the
design
in
general.
A
Okay,
if
there
are
any
other
pr's
that
are
blocked
for
review,
I
think
the
wall
updates
got
merged
on
the
collector,
so
the
prw
changes
exporter
changes
got
merged,
so
that's
available
on
the
latest
release
for
the
collector.
But
again
grace
did
your
pr
get
merged,
or
is
it
still
in.
A
A
And
just
wanted
to
check
with
richard
richie.
Is
there
anything
else
that
you
wanted
to
bring
up
on
your
end,
any
any
details
on
on
on
areas
that
we
could
be
helping
or
working
on
in
the
prometheus
on
the
prometal
side,.
B
This
is
the
same
song
as
always,
ideally
with
the
testing,
but
there
are
no
tests
on
our
site
or
anything
new,
so
yeah.
A
Yeah,
we
we
have,
you
know,
started
running
in
regular
daily
run
of
the
prw
tests.
You
know
on
our
on
our
end,
but
we'd
like
to
add.
I
think
you
know
it'd
be
nice
to
kind
of
enable
that
on
the
collector
pipeline
itself
on
the
project
also,
but
we
do
run
it
and
we
are,
you
know,
making
sure
that
the
tests
are
compliant
all
running.
A
So
so
I
saw
an
thread
on
the
developer
list,
about
the
open,
metrics
tests
and
and
some
of
the
compatibility
tests
being
added.
Is
there
are
there
requirements
that
you
have
defined
further
in
terms
of
some
of
these
tests
that
could
be
built
out?
I
mean
do
those
does
that
full
suite
exist
at
this
point,
or
is
it
just
in
or
is
that
something
we
could
help
out
with.
B
B
B
A
Yeah,
well
I
I.
I
think
this
is
a
good
good
link,
because
I
don't
think
I'm
on
this
list.
I
usually
track
the
prometheus
developers
list.
A
All
right,
any
other
updates
richard
did
you
want
to
give
any
updates
on
prom
con
or
anything
else,
that's
planned
in
terms
of
some
of
the
testing
and
the
compliance
requirements.
Is
there
a
session
on
that.
B
Yes,
there
will
be
last
thursday
the
cncf
governance
board
agreed
on
the
on
the
conformance
program
as
a
as
an
abstract
thing
cncf
and
asked
me
to
to
finalize
the
logo
guidelines
and
all
that
stuff.
Basically,
what
I
said
last
week,
I
didn't
yet
hear
back
from
them
for
prom
con
itself.
The
plan
is
to
to
give
the
current
state
and
also
basically
announce
the
program
in
full,
ideally
with
with
the
starting
members
who
are
already
compliant
with
it,
and
before
that
we
need
to.
A
G
Hey
richard,
the
confirm
this
is
for
conformance
tests.
I
Not
really,
but
not
really
so
the
confirmations
we
are
talking
here
are
specific
to
prometheus,
or
are
they
specific
to
open
telemetry?
I
just
these.
A
Are
these
are
prw
tests
that
this
is
the
prometheus,
remote
right
specification
and
the
related
tests
to
them,
which
we
run
on
the
open,
telemetry
collector?
You
know
in
order
to
make
sure
that
the
prw
exporter
that
we
have
is
is
passing
all
those
tests
right.
Okay,
so.
A
Of
being
able
to
use
it
for
any
other
prometheus
remote
right,
you
know
compatible
interface.
A
Okay,
everyone
thanks
so
much.
I
think
we've
run
out
of
our
topics
of
us
I'll,
give
you
back
20
minutes
and
appreciate
everybody
joining
in.
I
know
we
are
hitting
vacation
time.
So
thanks
have
a
nice.