►
From YouTube: 2022-10-03 meeting
Description
Open Telemetry Meeting 1's Personal Meeting Room
B
C
C
C
B
Yeah,
that
makes
sense,
but
we're
about
two
weeks
away
from
V1
release
so
time
to
kind
of
buckle
down
and
figure
out
the
remaining
items
so
for
the
agenda.
I
kind
of
wanted
to
focus
on
that
in
general
for
in
terms
of
completed
release,
items,
sijo,
implemented.net,
runtime
and
instrumentation
libraries
last
week,
so.net
is
now
fully
instrumented
and
Pierre
was
doing
a
great
job
on
the
services
documentation.
I
think
the
only
thing
we're
missing
right
now
is
the
rust
service.
B
Yeah,
so
we're
just
trying
to
point
it
to
you.
Real
quick,
I
think
it's
in
yeah
Docs,
so
we're
just
trying
to
do
kind
of
an
overview
by
service
and
language.
What
pres
like
what
you
know,
features
are
present
and
then
doing
a
little
bit
of
explanation.
So
you
know,
even
though
it's
a
demo,
we
need
a
point
users
where
we're
actually
doing
things
as
a
thought
and
so
Pierre
put
together
kind
of
as
a
template
that
has
been
used
across
the
services
too.
E
B
Well,
right
here,
yeah
just
send
me
a
ping.
If
you
have
some
issues
with
it
or
just
let
Juliana
know
if
you
prefer
him
to
take
care
of
it,
no
issues
there
sure
and
then
for
python,
metrics
I
think
Austin,
Adriana
I
obtained
me
and
said
she
was
going
to
be
looking
to
do
it.
But
we
do
have
you
know
about
a
week
or
so
left.
So
if
you
could
maybe
just
kind
of
keep
a
pulse
on
that.
B
Okay,
great,
we
also
have
a
go:
metrics,
PR,
open,
it's
been
open
for
about
a
month
or
so,
and
I
tried
to
kind
of
pivot
it
to
doing
our
hero
scenario
gauge
and
just
getting
basic
go
metrics
implementation,
but
I'm,
not
a
good
guy,
so
I
think
we'll
definitely
need
some
eyes
on
this.
Pr
specifically
just
add
basic
metric
support
and
ziki
is
in
China,
so
he's
not
able
to
join
any
of
our
Sig
calls.
B
C
A
B
Yeah
and
what
kind
of
Tramway
I
need
to
Pivot
a
bit
Helm
custom
configuration
I'm,
not
sure
if
Tyler
or
Pierre
I
think
Pierre
said
he
wasn't
gonna
be
able
to
make
the
call
so
I
think
they'll
still
be
looking
at
that.
But
we
love
to
get
that
kind
of
landed
in
V1.
It's
also
something
we'll
have
to
keep
a
pulse
on
and
then
in
terms
of
default.
Dashboarding
experience.
B
So
we've
had
the
wonderful
Angus
volunteer
to
help
us
out
a
little
bit
from
the
grafana
team,
but
we
still
need
to
do
a
fair
amount
of
work
on
this
ourselves.
So
I
think
we
need
to
do
a
little
bit
design
on
kind
of
like
the
hero
scenario,
user
flow
and
then
how
that
you
know
is
plugged
into
our
kind
of
default,
dashboarding
experience
and
what
that
looks
like.
So
we
need
a
little
bit
of
design
there
and
then.
B
A
Worked
out,
but
there
is
just
one
thing
that
I
mentioned
in
my
last
comment:
yeah
either
we
we
need
to
update
the
readme
or
just
start
a
new
container
whenever
we
are
testing
because
of
the
quote
service,
so
that
change
it
Mikko,
spotted
that
in
his
test,
so
yeah
I
can
I
can
I
can
have
this
done
today,
so
we
can
finally
merge
this
PR
awesome
great.
Now
we
are
close.
B
B
I'm
gonna
pick
him
on
that.
I
didn't
want
to
step
on
his
toes
while
he
was
making
some
like
finalization
updates,
but
yeah
I'm
planning
on
either
asking
him.
If
he
wants
to
do
that,
or
just
doing
it
myself
and
then
I
think
he
also
was
using
like
the
term
Auto
instrumentation
and
some
of
the
docs
which
I'm
hoping
to
not
use
and
mostly
Focus,
around
instrumentation
libraries
outside
of
java
and
maybe
like
one
or
two
other
languages
yeah.
It.
A
B
B
So
what
we're
making
some
small
updates
here
but
yeah?
Essentially
it's
like
checkout
service
right
now,
which
I
don't
think
is
very
user
friendly.
It's
ideally
be
like
go
check
out
service,
or
something
like
that.
So
just
if
a
user
is
interested
in
specific
language,
they
can
get
right
there
and
not
have
to
check
like
our
architecture
or
something
like
that.
B
A
C
A
B
B
C
Yeah,
ideally,
we
would
I
mean
some
of
this
should
be
done
for
us,
but,
like
maybe
have
I
think
the.
If
we
were
trying
to
be
like
best
practices
with
it,
then
you
would
kind
of
have
the
the
big
dashboard
and
you
know
that
was
sort
of
like
oriented
around
like
slos
different,
even
if
we're
not
building
actual
slos
but
just
sort
of
like
hey.
C
C
Out
the
service
name
and
then
flip
between,
like
here's,
the
memory
usage
for
blah
I,
don't
know
how
feasible
some
of
that
is,
but
because
I
don't
know,
I
mean
I
would
assume
that
like
host
metrics
are
normal
like
host,
metrics
should
be
normalized,
you
know
or
like
runtime.
Metrics
should
be
normalized,
but
there's
like
go
GC
and
Java
GC
or
different
things.
D
A
D
D
D
I
have
started
with
the
published
him
shorts,
but
then
I
I
slightly
changed,
because
when
I
deployed
the
front
end,
the
JavaScript
instrumentation
from
the
client
side
is
trying
to
reach
out
to
localhost,
which
makes
sense
so
I
try
to
change
it
to
have
another
endpoint
for
the
JavaScript
instrumentation.
On
the
client
side,
we
have
like
a
Ingress
that
opens
rule
that
opens
to
The
Collector
for.
B
D
A
C
B
D
A
A
D
D
A
C
Like
is
there
a
way
that
we
could
I'm
totally
cool
with
I'm
very
happy?
Actually
you
are.
You
are
using
the
demo
for
this
and
I
don't
terribly
mind
that
it's
in
a
different
thing,
because
it
actually
might
be
just
from
a
like
keeping
this
working
group
kind
of
focused
like
if
we
can
stay
focused
on
just
sort
of
like
application
and
observability
concerns
and
kind
of
not
have
to
end
up
taking
a
lot
of
kubernetes
stuff.
C
B
A
D
I
can
clearly
upgrade
the
environment
using
your
latest
version,
then
so
then
I
will
have
the
latest
metrics.
One
thing
that
I
for
for
your
for
kubcon
I,
don't
know
if
you
would
be
interested,
but
we
can
also
add
to
the
collector
the
kubernetes
receivers,
to
get
some
metrics
so
basically
saying
that
we
do.
We
don't
use
Prometheus
exporters,
the
cubesat
metrics
don't
explore.
We
use
100
of
what
has
been
built
by
the
open.
C
Well,
we've
I
mean
like
on
our
side
it
lights
up
like
we
found
in
terms
of
actually
trying
to
like
do
production,
kubernetes,
monitoring
with
Hotel
you're,
still
kind
of
limited.
You
still
have
to
basically
put
in
Cube
prom
stack
and
then
delete
kubernetes
and
then
tell
the
collector
to
scrape
just
because
it
doesn't
have.
You
know
eating
Cube
State
metrics
and
you
need
the
other
thing
that
I
can't
remember
the
name
of.
C
All
right,
I
think
in
terms
of
here's
my
thoughts,
let's
focus
from
this
group
about
getting
grafana
dashboards
for
the
things
that
there's
no
way
anyone
else
would
have
graph
on
a
dashboard
for
like
the
application,
metrics
and
the
span
metrics,
and
things
like
that
right,
there's
just
things
that
you
know
nothing.
You
know
those
nothing
else
is
going
to
produce
those
but
us.
C
So
let's
focus
on
that
from
our
end
and
once
we've
got
those
grafana
dashboards
they'd
be
very
easy
to
integrate
into
something
like
this
from
the
observability
tag
side,
let's
figure
out
what
needs
to
be
done
to
like
either
our
Helm
charts
or
some
something
about
this,
so
that
we
can
kind
of
keep
these
things
more
up
to
date
and
in
sync,
with
each
other
and
I.
Think
in
terms
of
you
know,
division
of
like
ownership
or
responsibility.
C
C
C
B
C
I
mean
we
should
probably
assume
I.
The
thing
is
I'm,
assuming
that
exemplar's
work
in
hotel,
and
so,
if
we're
creating
the
metric,
if
we
create
the
instrument
and
we
start
doing
measurements,
it
should
start
attaching
Exemplar
it
should.
It
should
fill
in
the
exemplars
in
the
data
structure
and
then
a
Exemplar
aware.
C
I
mean
I
think
we
figured
it
would
in
this
case,
because
the
cash
reset
the
it
would
only
grow
under
a
certain
like
either
region
or
customer
ID
or
some
something
would
where
in
aggregate
you
wouldn't
see
in
aggregate
all
you
would
see
is
line
go
up,
but
if
you
were
able
to
Pivot,
you
know
if
you're
able
to
like
say.
Oh,
let's
go
into
the
trace,
you
could
see
like
oh
well.
C
B
C
Dashboard
for
the
app
metrics
and
we
can
use
template
variables
to
like-
let
you
flip
between
like
because
because
we're
doing
spam
metrics
we're
doing
the
spam
message
processor
on
The
Collector.
So
we
should
have
radar
iteration
for
everything,
and
so
we
should
be
able
to
just
put
a
template
variable
on
there
for
service
name
and
then
be
able
to
go
right.
D
B
So
we
have
that
deep
for
default
experience.
We
also
have
the
tracing
components.
I
can
just
go
straight
from
that
once
an
Exemplar
is
ready
and
also
see
that
in
Jaeger
I
think
that
mostly
covers
the
hero
scenario.
Besides
doing
some
documentation
around
kind
of
like
the
guided
component
of
it
does
anyone
else
have
any
kind
of
comments.
Matt.
B
Oh
Tyler's,
here,
hello,
Tyler
we've
been
talking
about
the
home
charts
I
think
we
have
a
couple
items
I.
Think
one
thing
is
we're
looking
to
improve
the
like
custom
configuration
experience
so
say
like
you're,
a
vendor
like
honeycomb
or
dinotrace.
How
can
we
make
it
so
you
can
like
easily
customize
it
and
then
I
think
there's
also
some
conversations
around
this
observe
Kate's
demo.
If
you
haven't
seen
that
yet,
but
essentially
they've
are
kind
of
working
art
demo
and
doing
like
a
best
practices
observability
for
kubernetes.
Here.
E
I
came
to
give
an
update
on
that
first
part,
which
is
the
configuration
I
just
merged
into
something
to
The
Collector
chart
that
will
allow
us
to
do
templating
in
the
configuration
but
long
story
short.
We
should
have
the
collector
as
a
sub
chart
of
the
demo
chart
by
the
end
of
today.
Fingers
crossed,
in
which
case
you'll
get
to
have
you'll,
be
able
to
just
set
any
collector
configuration
you
want
and
it'll
take
it
like
you
can.
You
can
change
the
collector
configuration
using
this.
The
existing
rules
of
the
collector.
E
Is
go
ahead,
we'll
go
on!
Go
on
your
caveat
first,
my
caveat
is
that
The
Collector
has
a
default
configuration
that
includes
logs
and
metrics
and
with
the
collector
charts
current
implementation
in
the
way
that
Helm
uses
subcharts
and
nulls
and
values.yaml.
We
can't
turn
that
off.
So
like
the
when
someone
deploys
the
demo
Helm
chart,
The
Collector
that
gets
deployed
will
be
the
collector
charts
version
and
it'll
technically
have
an
otlp
receiver
in
metrics
and
logs
that
won't
be
doing
anything
it'll
just
be
there.
C
E
D
I
think
there's
even
an
issue
on
the
front
end,
because
I
changed
I
created
two
variables
for
The
Collector
at
the
backend
services
and
the
and
then
another
exporter
for
the
front-end
services.
And
even
if
you
do
that,
there
is
a
weird
Behavior.
So
I
think
there
is
something
we
need
to
resolve
on
the
front
end.
B
B
C
B
I
think
we're
on
track
for
it.
So
the
the
main
concerns
for
me
are
the
go
metrics
PR
I
mean
that's
really.
The
the
big
kind
of
outstanding
development
item
I
think
python.
Metrics
is
that
nice
to
have-
and
you
know
ideally,
we
do
have
it,
but
if
that
doesn't
land,
it's
not
the
end
of
the
world.
Based
on
based
on
how
I
saw
like
the
python
metrics
kind
of
maturity
that
wasn't,
you
know
completely
sold
on
it.
I
guess
I.
A
E
D
A
C
It's
just
like
hey,
go
look
at
these
docs
in
GitHub
and
so
I
think
the
focus
should
be
on
documenting
the
hero
scenarios
in
GitHub
mm-hmm
and
that's
going
to
be
the
most
resilient
sort
of
way
to
actually
do
that.
That's
going
to
be
the
the
least
fragile
way
of
doing
this.
If
we
can
have
someone
write
up,
basically
like
a
walk
through
of
you
run
it
here's
the
pr,
you
know,
here's
the
problem
that
you
can.
You
know
here's
screenshots
of
grafana
showing
the
problem,
and
it's
like
okay.
How
do
we
solve
you
know?
C
How
do
we
investigate
this
or
how
do
we,
you
know,
and
how
to
and
understanding
how
the
Telemetry,
how
the
instrumentation
connects
to
what's
going
on,
what
we're,
seeing
in
terms
of
telemetry
like
what's
the
code
that
we're
doing
like?
How
does
this
Con
like
link
to
that?
And
you
know
how
does
this
all
fit
together
like
that's
what
we
need
in
terms
of
hero
scenario,
documentation,
I,
think
and
then
for
the
website?
It
can
just
be
a
for
now.
B
Yeah
I
think
the
ball
I
could
just
be.
We
already
had
the
announcement
blogs
that
it
was
to
be
like
a
V1
release
blog.
What
to
talk
about
some
examples.
We
have
and
then
yeah
just
point
them
there
yeah
so
I
should
I
should
have
no
problem.
Writing
that
Gary.
If
you
have
more
of
an
interest
on
like
the
actual,
like
technical
hero
scenario,
side
I'm
happy
to
switch
on
that
too,
but
we
can
just
talk
on
on
slack
as
well
sure.
Okay,
that
makes
sense.
That's.
C
A
C
The
scenarios
and
stuff
are
like
written
well
enough,
and
it's
evident
enough
that,
like
what
is
what
is
happening
and
what
we're
trying
to
get
across
that
when
people
that
maybe
like,
because
the
goal
is
for
people
who
maybe
have
heard
of
open
Telemetry
but
never
really
like
used
open
Telemetry
that
this
should
be
their
entry
point
into
the
project.
If
they're
not,
you
know,
if
they're
kind
of
a
more
of
a
lay
person
or
they're
not
like
an
observability
nerd
or
whatever,
so
we
just
want
to
be
very
thorough
and
kind
of
like
hey.
E
B
C
A
C
B
B
A
E
A
E
Take
care
of
that
yeah
we
have
there's
someone
who
submitted
the
helm
chart
initially
from
the
demo
like
Sig,
that
is
an
owner
for
it,
but
I
haven't
seen
them
around
in
a
while
wph.
Maybe
he's
an
interesting
I
forget
if
he,
the
expectation
was
that
that
owner
would
like,
when
the
demo
changes
come,
make
those
changes
to
the
demo
chart.
If
he's
not
able
to
do
that,
I
would
love
for
someone
else
in
the
demo.
Sig
who's
like
actively
aware
of
like
the
different
Services,
would
come
and
contribute
that
stuff.
B
E
B
B
A
C
A
C
A
B
C
B
A
A
It's
saying
is
here
so
I'm
on
this
conversation
here,
but
if
I
can
spin
up
the
application,
I
haven't
taken.
Look
in
the
refill
yet
I'm
guessing
there's
a
link
similar
to
the
application.
If
I
can
spin
that
up
and
take
a
look
at
what
metrics
it's
emitting,
maybe
I
can
make
some
suggestions
on
a
dashboard
as
well,
but
I'm
just
I
just
joined
the
conversation
on
Friday
evening
or
of
the
weekend
so
bear
with
me
like,
while
I
get
up
to
speed.
B
A
B
The
most
recent
release
wouldn't
have
some
of
our
metrics,
so
you're
gonna
have
to
build
just
from
like
the
source,
repo
itself,
so
there's
an
FYI
there
and
yeah.
You
should
be
able
to
see
the
default
experience
at
the
top
or
out
of
the
box,
and
just
let
us
know
kind
of
your
feedback
from
that
curious
to
learn
more.