►
From YouTube: 2022-08-29 meeting
Description
Open Telemetry Meeting 1's Personal Meeting Room
B
A
Did
you
do
the
dreaded
doctor
system
pruning.
C
Yeah,
I
actually
it's
still
building
on
my
droplets,
so
I
haven't
like
tried
it
again
locally,
but
it's
my.
B
D
E
E
C
E
Cool
well,
it
seems
like
we
have
some
good
leadership
representation
right
now,
so
we
officially
have
around
a
month
before
our
upcoming
release.
This
is
a
good
time
for
us
to
kind
of
discuss
what
we're
actually
looking
to
accomplish
and
when
do
we
want
to
kind
of
institute
a
future
freeze
and
work
on
just
improvements
as
well.
So
are
there
any
other
kind
of
items
of
interest,
or
you
know
you
guys
think
relevant
topics
over
the
next
month.
C
E
E
A
lot
of
open
pr,
so
maybe
maybe
we
should
take
a
look
at
that
real,
quick
too.
I'm
curious
what's
outstanding,
and
maybe
some
of
these
we
can
agree
as
a
group
to
just
say
we
don't
want
these
feature
editions.
C
A
Is
I
think
I
I
think
you
hit
it
carter
because
it
happens
on
mine
as
well,
I'm
using
honeycomb,
I
owe's
repo.
So
if
you
use
a
corporate
repo
to
to
create
your
fork
from
that
corporate
repo
is
locked
down
that
you
have
to
be
a
member
of
that
corporation
in
order
to
create
branches
on
it.
So.
C
E
C
C
E
E
A
Any
aprs,
I
think,
we're
austin's,
going
though
it
probably
makes
sense
to
understand
which
github
branch,
repo
protections
are
the
ones
of
the
culprit
here
yes,
and
set
those
as
guidelines
like
hey,
look,
we
we
love
your
pr's,
but
please
make
sure
you
don't
have
this
turned
on
your
repo
or
something
along
those
lines.
So
I
think
it
probably
makes
sense
for
us
to
investigate
which,
because
there's
several
different
options
on
those
github
protection
features
but
which
one
is
is
causing
that
being
hold
up
here.
A
F
I'm
looking
at
the
repo
in
question
right
now,
at
least
for
the
pr
that
you
all
that
flagged
for
me
here.
I
have
the
branch
protection
rules
open.
If
you
want
to
investigate
it,
it's
basically
just
require
approvals
and
require
voter
requests
before
you
merge
into
maine.
That's
the
only
brand
protection
I
have.
E
Okay,
well,
michael,
you
may
have
some
of
your
mprs
to
update
too.
I
see
the
health
check
and
the
yeah
two
health
checks
for
me
when
you
get
cleared
out
as
well.
I'm
not
sure
if
you'll
have
time
today.
A
A
One
yeah
and
against
older,
like
it's
got
to
be
redone,
so
I
would
be
you
know
we
could
just
close
stale
and-
and
somebody
should
pick
up
adding
metrics,
but
we
should
do
it
using
the
sdk
in
place
today,
not
on.
E
D
E
Happy
with
that
awesome,
okay,
so
we
have
some
good
good
coverage,
at
least
in
the
repo.
So
I
think
we've
already
cut
out
php
from
kind
of
the
the
upcoming
release.
I
I
saw
you
had
a
draft
pr
juliana.
You
still
want
to
push
forward
with
that,
or
are
you
gonna
keep,
maybe
just
holding
off
on
it.
B
That's
what
I
wanted
to
ask
actually,
so
I
know
that
when
gary
opened
the
rewrite
on
rust,
he
mentioned
something
about
the
quote
that
that
was
just
like
multiplying
8.90,
for
how
many
products
we
have-
and
he
mentioned
something
hey.
We
could
add
some
logic.
So
I
was
thinking.
Maybe
we
could
add
some
logic
in
php
and
then
rust
calls
php.
B
And
then
we
have
php
in
the
symbol
because
just
doing
a
http
call
from
rust
to
the
to
the
ph
to
a
php,
simple
service,
wouldn't
be
that
hard
and
I
actually
got
php
running
generating
expense.
So
it
would
be
only
implementing
this
logic.
A
I
I
I'm
I'm,
I'm
just
going
to
play
a
little
bit
devil's
first
off.
I
think
we
should
do
this,
but
I'm
going
to
play
devil's
advocate
is:
do
we
plan
on
eventually
writing
a
php
service
to
administer
the
product
catalog
as
like
a
version
1.1
or
whatever
it
is,
and
what
would
that
mean
to
this
php
quote
service?
Should
we
do
that.
E
Like
who
would
just
replace
it
in
place
right
because
it
seems
like
the
juliano's
option
is
a
great
temporary
replacement,
but
I
still
think,
like
the
you
know,
northstar
vision
would
probably
have
an
actual
php
admin
service.
The
reality
is
like
when
you
say
we
I
mean
I'm
on
a
php
dev.
I
think
juliana
is
just
becoming
a
php
dev.
I
don't
think
I
don't
know
if
we
can
do
it,
but
if
we
give
it
time
we
can
find
a
job.
We
can
it's
kind
of
how
I
look
at
it.
A
I
I
guess
what
I'm
trying
to
say
is:
let's
say
in
december
january:
whatever
later
in
life,
we
decide,
we
get
a
php
admin
server
for
a
product
catalog.
What
happens
to
this
quote
service?
Do
we
keep
it?
E
E
You're
giving
me
like
vietnam
flashbacks
my
main
job
right
now.
I'm.
A
F
I
I
don't,
I
think,
that
it's
as
egregious
of
a
solution
to
like
just
multiply
a
couple
of
numbers
together,
as
maybe
it
feels
to
other
people
in
the
room
like.
I
don't
think
that
that's
that
bad
of
just
like
okay
yeah,
we
don't
have
this
great
service
that
can
get
those
numbers.
Yet
unless
product
catalog
does
it-
and
I
just
don't
know
it,
but
it
feels
like
it
would
be
more
useful
to
like
it.
F
D
A
Yeah,
it
is
what.
C
C
We
can
go
and
say
like
oh,
this
is
this:
has
all
the
hotel
languages
except
swift.
Like
that's
easy
to
say
if
we
said
this
is
all
telling
except
swift
and
php,
then
that
that
feels
more
agree
like
that.
That
feels
like
more
of
a
like,
because
you're
missing
an
import
like
an
important
back-end
language.
C
Let
me
add:
some
more
air
quotes
around
important
php.
There,
hey
happy.
C
Yeah
some
people,
as
that
was
to
say,
but
it
it
like
swift,
is
a
swift
is
like
an
emission
of,
I
don't
say
convenience,
but
certainly
of
practicality.
Php
emission
would
be,
I
think,
harder
to
like
paper
over.
C
D
C
E
G
A
C
Right
like
this,
it's
probably
better
to
have
the
quote
service
than
the
admin
service
and
we'll
just
like
mark
out
the
admin
service.
If
we
want
to
like
like
here's,
the
thing
too,
if
we
want
to
you
know
if
we
wanted
to
add
the
admin
stuff
back
in
at
some
point
like
we
can
throw
that
into
the
earling
one,
it's
not
hard
to
go
in
and.
E
C
So
that's
this
is
like.
What's
the
point
or
I
feel
like
the
the
argument
for
the
admin
service
is
much
slimmer,
because
nothing
else
is
taught
because
the
admin
service,
if
I
recall,
like
only
talks
to
the
feature
flag
service
and
the
database,
that's
supposed
to
be
there
for
the
product
catalog
and
like
the
second
one
of
those
doesn't
exist,
and
we
have
db
calls
already
right
like
when
every
time
there's
a
feature
flag
evaluation,
that's
a
db
call.
C
A
Juliano,
do
you
have
any
plans
on
updating
the
proto
for
this?
Are
you
just
going
to
do
a
call
for.
B
Now
not
really
because
we
do
not
have
support
for
grpc
and
php,
so
the
php
can
only
have
a
client
in
jpc
so
yeah.
I
was
thinking
about
just
using
hp
request
from
rust
or
something
else
to
php
and
then
done
done
as
we
already
have
hp
for
to
ruby.
So
then
I
thought
yeah,
maybe
okay,.
A
C
C
E
A
I
noticed
this
as
well:
payment
needs
to
have
baggage
enabled
because
we
actually
do
check
pay
baggage
inside
the
payment
service,
so
that
should
be
100.
Okay,
I
think
the
rest
of
them.
We
should
probably
n
a
them
because
we're
not
going
to
be
looking
at
baggage
in
any
of
them,
at
least
not
yet
I
don't
know.
Maybe
I
think
we
can
metrics.
A
Yeah
span
links
will
only
be
on
the
front
end.
That's
the
only
thing
I
think
where
we,
we
could
actually
come
up
with
a
use
case
where
spam
links
make
sense.
I.
C
D
B
C
Yeah
this
is
this
is
not
me
like
our
work
here.
I
think
we've
actually
done
a
good
job,
I'm
just
saying
the
only
other,
like
the
only
like
the
use
case
I
hear
about
from
customers
or
whatever
is
kafka
right.
It's
people
putting
stuff
on
the
queue
and
we
don't
have
any
like
pub
sub
in
here.
We
don't
have
any
queues
in
here.
We
don't
have
another
like
natural
place
to
put
in
spam
links,
so
I
would
rather
us,
rather
than
coming
up
with
some
very
convoluted,
like
fake
reason
to
use
them.
A
If
you,
while
you're
still
here,
I
I
got
another
one
for
any
for
you.
I
don't
think
russ
has
any
instrumentation
libraries
yeah.
E
It's
happy,
I
was
gonna,
say:
let's
go
ahead
and
call
these
out
so
rust,
instrumentation
library,
I
know
the.net-
does
have
manual
span
creation,
so
I
don't
even
need
to
be
effective.
A
E
It
rpc
context
provocation.
This
is
probably
all
handled
by
instrumentation
libraries,
maybe
just
n
a
for
most
of
the
film.
A
E
Okay
span
data
enrichment.
Do
we,
I
think
here
you've
looked
at
the
future
flag
service
recently.
Do
we
do
any
of
those.
D
D
C
Did
you
cause
like
just
tell
him
austin
said
hi
yeah,
yeah,
okay,
I'll,
not
just
d
of
him
right
now
and
be
like
hey
pierce
gonna
ask
for
a
favor.
C
Sas
and
I'm
on
francais,
but
he
I
I
mean
yeah,
I
he
knows
early,
I
mean
he
literally.
A
A
A
I'll
see,
there's
one
more
thing
that
we're
we
identified.
We
wanted
to
do,
but
we
don't
have
anything
to
do
and
I
see
gary's
still
on
on
the
horn
here,
because
it
it
almost
requires
gary
to
write
the
code
since
he's
our
expert,
but
oh
boy,
we
said
we
were
gonna,
have
rust,
check
a
feature
flag
and
cause
just
an
unexpected
delay.
A
Should
the
shipping
address
be
canada
or
not
us,
I
should
say
okay
yeah,
so
we
need
rust
to
make
a
call
to
the
feature
flag
service
to
check
to
see
if
the
feature
flag's
enabled
and
if
it
is
or
actually
I
think
first,
you
should
check
to
see.
If
the
address
is
usa
or
not,
and
if
it's
not
usa,
then
you
check
the
feature
flag
service.
If
it's
enabled
delay
like
10
seconds,
I
don't
know
some
really
really
long
delay
that
feels
unnatural
and
then
we
return.
A
Anytime
they're
doing,
I
think
it's
shipping,
when
they're
they're,
calling
to
get
the
tracking
number
that's
what
I
think
it
is
get
tracking
number
that
call
there.
If
the
address
is
outside
usa
and
the
feature
flag
service
is
enabled
or
the
feature
flag
for
this
is
enabled.
Then
we
delay
like.
F
Whatever
some
yeah
10
seconds,
I
get
it
it's
asleep
or
something
that
sounds
doable.
I
can
check
that
in
an
issue
and
get
started
on
it
pretty
soon.
I'm
caught
up
in
a
project
here
for
the
next
couple
of
days,
but
once
I
get
that
through
it's
my
bachelor
party
this
weekend
and
then
after
that.
A
F
Can
I
can
work
on
it
probably
starting
yesterday.
A
F
A
F
A
F
Just
want
to
check
is
there
proto
for
checking
this
feature
flag?
Yes,.
A
F
E
D
Is
it
okay?
Is
it
easy
for
this
sleep
to
be
made
non-blocking
in
the
the
rest
there.
D
F
Doesn't
matter
because
tokyo
does
do
the
asynchronous
runtime
in
a
way
that
I'm
pretty
sure
like
once
you're
in
the
method,
unless
you're
saying
this
part
is
also
async,
and
you
await
for
that,
like
I'm,
pretty
confident
that
it'll
sleep
and
then
it'll
try
to
do
the
asynchronous
call
so
long
as
you
sequence
it
properly,
but
that's
part
of
the
the
fun
I'll
play
around
with
it
and
see
what
happens.
You'll
see
on
the
trace
right.
E
Well,
yeah
congrats
scary,
for
that
too,
and
definitely
enjoy
yourself
helm,
sharks
next
next
item
on
the
agenda,
so
I
didn't
put
this
one.
So
if
whoever
did
could
you
speak
to
it
potentially.
H
Yeah,
it
was
me
so.
On
friday
I
was
working
with
someone
from
the
traditions
team
to
deploy
the
hotel
demo
to
our
kubernetes
cluster,
and
so
we
were
going
through
the
different
things
that
are
part
of
the
health
shark
and
we
had
to
create
some
resources
manual
resources,
one
of
them
it
was
the
english
rules
for
both
the
other
web
store
and
the
collector.
H
So
I
guess
those
things
is
something
that
we
we
can
consider
adding
them
to
the
to
the
home
charts.
C
Yeah-
and
it
would
be,
you
need
fancy
logic
for
local
self-hosted
cloud.
G
We
could
probably
make
that
work
for
this.
I
think
there's
more
important
things
for
this
particular
chart
at
the
moment.
That
would
be
a
higher
priority
in
my
mind,
there's
several
defects
with
it
right
now
that
are
causing
problems
that
I'd
like
to
see
fixed
first
and
then
there's
some
other
more
basic
configuration
options.
That
would
probably
be
more
beneficial
than
english,
although
I
don't
know
for
sure.
C
I
I
feel
like
we
should
just
follow
tyler's
lead
here
and
button
up
what
we've
got
and
then
look
into
ingress
and
things
like
that
for
a
just
thinking
forward,
a
little
bit
after
v1.
I
would
like
to
have
like
a
an
ideation
around
like
what
would
this
look
like
if
it
was
like
for
aws
for
gcp
for
sure
right
and
then
doing
things
like
that
as
part
of
that
work,
where
we
can
come
up
with,
like
a
you
know:
hey
here's
what
this
looks
like
on
cloudx
and
focus
more
on
the
self.
A
A
So
right,
where
we
say
hey
you
run
cube,
can
cuddle
proxy
or
deploy
an
ingress
route,
and
that
would
probably
allow
us
to
satisfy
those
who,
for
whatever
reason,
corporate
doesn't
allow
you
to
run
cube
cuddle
proxies.
But
you
have
the
ability
to
create
ingress
routes
instead,.
G
A
G
It's
definitely
not
up
to
date.
Yeah!
That's!
I
think
the
biggest
thing
is
that
we
just
need
to
get
the
environment
variable
updated.
Also,
the
environment
variables
are
not
easy
to
update,
so
I'm
actually
working
on
something
right
now
to
like
clean
up
how
we
use
environment
variables
throughout
the
chart,
because,
okay,
it's
kind
of
messy,
it's
like
all,
hacked
together
right
now,
I'm.
D
G
We
can
and
there's
actually
some
default
environment
variables
that
are
already
doing
that.
That's
the
first
thing
I'm
refactoring
right
now
is
how
we
use
those
default
values:
okay,
yeah.
A
G
That
that
the
issue
would
definitely
be
very
much
appreciated,
there's
actually
also
an
open
pr
right.
Now,
that's
a
higher
priority
than
me
doing.
This
refactor
that'll
actually
fix
the
feature
feature
flag
service.
G
Hopefully,
when
that
one's
resolved,
like
things
will
be
a
little
bit
more
stable,
although
it
sounds
like
there's
some
other
environment
variables
that
need
fixed
up.
So
if
you
go
and
open
an
issue
and
just
like
list
out
these
or
all
the
environment
variables
that
need
updated,
that
would
be
super
helpful.
A
Which
one's
that
one
is
that
it's
like
330.
G
There's
a
pr
open
up
right
now,
344
that
fixes
issue
343.,
it's
very
specific
to
the
feature
flag
service.
But
you
know
that's
the
that's
the
one!
That's
failing,
ci
right
now,
so.
A
He
then
that's
actually
what,
on
my
side,
I
had
the
same
one.
It
was
a
feature
flag
service
portal,
okay,
cool
yeah,
specifically
the
one
that
goes
through
right.
There,
free
flag,
server,
support
grpc
service
port
which,
by
the
way,
I'm
not
a
fan
of
how
their
names,
because
they
break
other
and
you're,
helpless.
A
G
A
G
Health
chart
will
always
reflect
the
needs
of
the
demo,
so,
like
I
would
say,
the
demo
sig
you
guys
should
do
whatever
you
need
to
do
in
order
to
like
make
your
environment
variables
and
all
of
your
code
work
as
expected,
the
helm
chart
will
always
adjust
to
fix,
will
always
adjust
to
meet
like
those
needs.
A
The
issue
that
tyler's
talking
about
is
the
feature
flag,
defines
a
couple
environment
variables,
called
feature
flag,
service,
port
and
feature
flag,
grpc
service
port,
ideally
they're
called
feature
flag
service,
port
and
feature
flag
service,
grpc
port
instead
or
something
along
those
lines
we're
basically,
instead
of
grpc
being
before
the
word
service
should
be
after
the
word
service
in
the
environment,
variable
name
that
would
require
us
to
change.
I
think
product,
catalog
and
gary
would
need
to
know
this
for
for
what
he
does
with
the
rust
service.
A
I
think
it
makes
sense.
We
rename
it
to
be
consistent
across
the
board
on
our
side
and
I'll
coordinate
that
one
with
you
tyler.
G
So
yeah
that
sounds
good
and
then
the
other
thing
that
we'll
you
know
me
and
dimitri
and
alex,
will
continue
to
take
care
of
maintenance
for
the
charts
and
all
the
ci
stuff
and
proper
patterns
and
examples
in
this
thing
and
keeping
it
in
line
with
the
other
charts.
G
I
performed
the
upgrade
this
last
time,
although
I
think
I
I
messed
some
stuff
up
missed
some
of
the
environment
variables
and
that
just
happens
because
I'm
not
super
familiar
with
the
actual
code.
That's
getting
deployed.
G
So
if
we
could
have
someone
from
the
demo
sig
take
point
on
the
actual
updates
like
when
the
image
changes
and
making
changes
to
the
chart,
or
at
least
making
it
clear
what
changes
need
to
happen
in
the
chart
in
like
an
open
issue
or
something
then
either
me
or
whoever
wants
can
can
handle
the
upgrade,
because
I
think
what
happened
on
the
three
the
three
four
upgrade
was.
I
wasn't
aware
of
all
the
changes
that
happened
and
I
missed
some
of
the
environment
variables
yeah.
That.
G
C
G
Minimize,
though,
I
understand
right
now
that
the
chart's
in
like
a
fluctuating
state,
yeah.
C
Okay,
can
we
oscar
you're
on
the
call
right
so
for
306?
It
feels
like
it's
still
pretty
flaky.
H
Yeah
yeah
I've
been
working
on
that
I
don't
know,
what's
going
on,
there's
something
around
one
of
the
calls
that
is
sometimes
not
triggering.
Sometimes
it
is
discussing
this
issue.
We
are
going
to
take
a
look
at
today
to
see
if
we
can
have
it
fixed.
D
C
H
Yep
yeah
I'll
take
a
look
to
see
what's
going
on,
there's
there's
something
that
needs
to
be
have
changed
on
the
international
tests,
because
the
catalog
changed
in
the
previous
year.
C
Just
I
mean
the
the
tests
need
to
be
something
that
either
like
when
things
change
like.
If
the
data
source
changes
behind
the
scenes
right,
then
I
don't
think
the
like.
If
the
items
in
the
catalog
change,
I
don't
necessarily
want
to
have
to
have
people
go
and
like
remember,
to
update
the
test
or
they're
breaking
the
build
like.
H
Yeah
those
can
be
pulled
from
the
products
source
of
truth.
C
H
C
Yes,
it
does
but
didn't
296.,
oh
well.
We.
D
Oh,
it
was
setting
up
the
volume
of
how
to
include
the
the
binary
file
right
that
I
have
to
do.
C
I
do
think
we
need,
I
mean,
because
we
do
need
the
health
checks
on
the
containers
for
helm
right
like
we
need
for
kubernetes
in
general,
because
we
need
to
expose
health
checks
to
well.
If
we
don't
expose
the
health,
if
we
don't
suppose
health
checks,
then,
like
cube
state
metrics,
doesn't
report
health
checks.
I.
F
D
I
think
kubernetes
might
do
health
checks
in
a
different
way
and
not
through
this,
like
grpc
probe
thing.
C
B
I
think
everything
depends
on
the
way
we
implement
in
kubernetes.
We
can
just
say
like
we
can
implement
some
help,
endpoint
and
then
just
do
http
requests
and
serving
okay
done,
and
that
would
work.
The
jrpc
is
easy
because
you
add
the
binary
and
then
just
specify
that
inc
in
the
yaml
file
and
then.
D
C
D
Okay,
so
do
we
want
to
make
kubernetes
like
a
required
thing
for
this
project?
Are
we
just
doing
docker
compose.
C
Right,
like
we
don't
need
like
docker
and
kubernetes,
do
not
need
to
be
100
feature.
Parity
like
docker
can
expose
a
more
limited
set
of
like
stuff,
like
that's
fine,
and
then
we
just
say
like
hey,
if
you're
using
kubernetes,
you
should
be
using
this
grpc
health
checking
thing
that
it
does
for
you
and
that
would
simplify
our
builds.
That
would
simplify
our
dependency
tree.
That
would
reduce
moving
parts
which.
A
D
H
A
A
All
for
fewer
things
to
do.
C
Yes,
I
think
less
moving
parts
is
good.
I
also
want
to
be
respectful
of
time.
Does
anyone
have
anything
else
urgent?
Because
I
do
have
a
heart
out
at
the
top
of
the
hour.
E
Yeah,
I
think
we
just
have
two
minutes
left,
I'm
not
sure
if
tyler,
if
you're
still
on
the
call,
if
you
have
a
list
of
like
best
practices,
we
need
to
implement
the
home
chart.
That's
probably
a
good
use
of
our
time
in
september.
So
you
can
either
just
send
me
a
quick
note
or
add
it
in
some
sort
of
issue
that
we
could
track.
E
I'd
really
appreciate
that
we
don't
have
a
meeting
next
week,
I'm
assuming
because
of
labor
day
so
most
of
the
us-based
people
will
not
be
present
and
then
we
can
potentially
spend
the
meeting
after
that
talking
more
in
depth
about
metrics.
I
think
that's
probably
the
biggest
thing
that
will
slip,
but
maybe
we
can
do
a
couple
of
implementations
before
the
v1
release,
depending
on
how
we're
feeling.