►
From YouTube: 2022-01-20 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
A
B
A
D
A
A
A
Fine
cool
all
right
thanks
everyone
for
joining
as
usual.
Please
add
your
names
to
the
attendees
list.
We'll
have
a
couple
of
topics
to
go
through
today.
A
As
usual,
we
usually
like
to
start
off
with
like
a
couple
intros
for
the
new
joiners
of
the
stick.
Just
to
you
know,
add
some
context.
A
F
Hi
yeah
I
can,
I
can
go
real,
quick,
I'm
reese,
I'm
with
new
relic
and
yeah,
I'm
new
to
open,
telemetry
and
python
also,
so
I
just
wanted
to
see
what
the
hot
topics
are
and
see
just
kind
of
get
more
context
around.
What's
going
on
and
and
learn
more.
A
A
A
Okay,
cool,
I
don't
know
if
everyone
else
gave
their
introduction,
but
like
hi,
I'm
layton,
I'm
one
of
the
python
maintainers
I've
been
working
on
this
for
about
like
ever
since
the
inception,
actually
so
pretty
pretty
old,
but
I'm
not
very
old.
So
that's
like
yeah!
That's
what
it
is
nice
nice
to
meet!
You.
G
Yeah,
I'm
diego,
I
think
I
think
you
introduce
yourself.
G
A
D
A
On
slack
yeah,
okay,
cool,
so
right,
I
guess
we
could
just
get
started.
Let
me
share
my
screen:
real
quick.
A
All
right,
I
guess
first
topic,
diego
remove
support
for
3.6
yeah.
A
A
G
G
Yeah
3.6
is
reached
its
end
of
life
december
last
year.
Can
you
click
in
the
in
that
link
that
I
have
there
right?
No
there,
the
top
one,
no.
G
G
Okay,
yes
3.6!
Lately,
if
you
go
to
the
latest
points,
move
to
the
right
piece
yeah,
I
think
3.6
has
very.
D
G
Yeah
but
3010
is
it's
very
recent,
so
it's
very
normal
that
it's
not
being
used
still
so,
but
I
mean
it
is
not
being
used
that
much
and
it's
always
great
to
get
rid
of
stuff
to
support.
D
So
my
my
counter
argument
here
is
that
I
I
think
this
originally
stemmed
from
the
build
failures
which
were
not
tied
to
a
particular
version
number.
The
only
reason
it
was
tied
to
the
3.6
version
number
was
the
fact
that
3.6
is
the
only
version
that
we
run
the
contrib
tests
for
so
I
think
like.
If
I'm
not
mistaken,
that's
why
this
was
originally
brought
up
right.
G
I
am
motivated
enough
to
remove
three
or
six.
I
was
counting
the
days
until
it
reached
the
end
of
life.
I
mean
this
is
something
that
I
had
in
mind
some
time
so,
just
and
and.
F
I
think
if
we.
G
H
F
G
A
H
Even
it
is
either
it'd
be
nice
to
remove
it
to
just,
if
only
to
speed
up
ci,
but
but
we
removed
three
of
four
and
earlier
ones.
Before
we
had
one
dot
upward
facing,
I
mean.
D
H
Yeah,
so
I
guess
we
probably
need
to
come
up
with
like
some
some
policy
for
dropping
older
versions
and
maybe
also
compare
that
with
other
things,
probably
with
node
and
ruby
and
and
then
and
then
see
what
like.
If,
if,
according
that
policy,
3.6
is
dead,
remove
moment.
G
H
We
use
an
example
of
java,
but
so
I
I
don't
know
if
that
specific
case
is
more
like
python
2
versus
python
3.
like
if
we
supported
python
2
on
1.0,
then
we
would
never
drop
it.
I
guess,
but
maybe
for
python
3.
That's
not
that
big
of
a
deal
but
but
still
maybe
we
should
bring
it
up
in
the
internet's
meeting
and
try
to
see
what
the
general
feeling
is
about
that.
D
G
No,
no
sorry
support
for
3.7
and
in
one
year
five
months,
just.
D
Great
one
one
question
I
have
then
is
like:
would
it
make
sense
at
least
support
at
least
support
3.6
until
we
have
like
the
metrics
released,
because
right
now
it
feels
like
if
we
were
to
drop
3.6
people
who
were
still
using
3.6
to
continue
using
open
telemetry
with
the
previous
versions,
but
they
would
only
get
tracing
like.
Would
it
would
it
make
sense,
or
at
least
give
them
something
that
works
like
for
both
signal,
as
opposed
to
kind
of
leaving
them
in
this
half
state?
D
C
I
was
also
going
to
say
that
when
was
the
end
of
life
for
3.6,
it
was
december
2021,
so
just
like
a
month
ago
right
should
we
maybe
wait
at
least
a
few
months.
A
Another
thing
I
wanted
to
point
out
like
so
like
at
least
for,
like
you
know
like
like
for
the
azure
side,
when
we
depend
on
like
open
census
like
2.7,
you
know
was
like
dropped
a
while
ago,
but
we
still
have
to
support
it,
because
we
do
have
a
lot
of
customers
who
still
use
it,
and
I
don't
know
like
what
the
motivation
like
if
there
is
a
big
motivation
to
have
to
drop
this.
A
But
like
I
always
thought
that
we
we
created
things,
you
know
for
the
end
user
right
and
like
dropping
3.6,
would
make
our
product
not,
as
you
know
like
like
usable
right
for
at
least
people
who
are
new,
or
you
know,
people
who
are
still
on
3.6
and
have
have
a
reason
to
stay
on
it.
G
Well,
it
I
mean
that's
true:
every
time
we
drop
a
version,
we
will
somehow
affect
the
users
us
dropping
3.6
is,
I
think,
not
any
different
from
what
we
have
done
before
when
we
drop
through
the
five
and
we
drop
three
or
four
so.
A
I
think
it's
different,
because
that
was
like
a
calculated
like
we.
We
saw
that
the
number
of
users
weren't
so
impactful
right,
like
it,
was
a
tactical
decision
to
do
that.
G
D
Can
we
can
we
take
a
an
action
mirror
to
look
at
the
delta
between
when
3.5
was
end
of
life
than
when
we
stopped
supporting
it
and
maybe
creating
like
we
could
create
a
policy
around
yeah,
at
least
having
some
kind
of
consistency?
Otherwise,
otherwise
it's
just
a
pain
for
users
to
not
really
know
when
we're
gonna
stop
supporting
the
thing.
G
Yeah,
I'm
not
opposed
to
that.
I
mean
I'm
fine
if,
if
we
decide
to
stop
supporting
three
or
six,
if
we
see
that
in
in
the
last
month
the
average
was
some
certain
arbitrary
percentage
or
something
like
that.
The
the
point
that
I'm
trying
to
make
here
is
that
this
is
not
any
different
from
what
we
have
done
for
yep.
D
A
These
only
track
like
the
downloads
of
the
the
package
right
yeah,
it
doesn't
say
anything
about
the
usage
of
people
who
are
actually
using
it
already
right.
Sorry,.
D
A
C
Of
course
one
other
question:
does
anybody
know
if
there
were
like
syntactic
changes
in
3.6,
like,
I
think
the
only
like
big
thing
I
could
think
of
is
the
regular
dictionaries
became
ordered,
but,
like
I
know,
for
3.4
and
3.5,
there
were
some
like
syntactic
changes,
so
those
would
like
really
break
people
who
are
using.
C
C
G
I
think
it's
well
is
definitely
possible
and
very
probable
that
there
are
syntactic
breaking
changes
from
3.6
to
307..
It
happens
almost
always.
The
minor
releases
of
python
are
not
guaranteed
to
be
backwards
compatible,
and
they
always
include
some,
and
they
usually
always
do
something
that
breaks.
C
G
C
D
I
just
added
a
note
to
the
doc
here,
but
we
dropped
the
3.5
supports
six
months
after
the
end-of-life
notice.
Okay,
so
maybe
we
maybe
we
review
this
in.
I
guess
five
months
now.
D
G
E
D
A
D
Item
is
to
add
a
note
in
our
in
our
documents
somewhere
about
the
policy
of
supporting
end
of
life,
python
versions,
sure.
A
After
that,
do
we
need
to
have
a
discussion
of
how
we
handle
this
in
terms
of
like
versioning
and
everything.
A
F
D
A
Like
do
we
need
to
have
like
a
like
a
supportability
path
for
for
customers,
that's
what
I'm
asking
like
if
they
were
on
3.6
before,
or
is
that
like
out
of
our
scope,.
G
G
G
So
so
do
you
mean
if,
in
six
months,
someone
is
still
using
three
or
six
yeah
and
we
introduce
a
change
and
that
change
does
not
work
in
three
or
six
yeah?
Okay,
and
what
are
you
suggesting?
We
do
in
the
situation.
G
A
G
G
A
E
A
All
right
moving
right
along,
so
I
believe
last
week
during
the
sig
there
was
a
issue
that
brought
up.
You
know
someone
wanting
to
see
when
the
earliest
release
was.
I
believe
our
last
release
was
over
a
month
ago,
so
we
should
probably
plan
on
having
one
around
next
week.
A
D
Can
we
is
it
possible
to
get
like
an
end
to
end
metrics
sdk
working
by
the
next
release?
That
would
be
pretty
awesome
with
yeah
with
ltfp
right
yeah.
It.
A
I
think
I
think
we're
waiting
on
the
metric
reader.
Where
is
that
this
thing
and.
C
Do
it
without
without
views,
but
then
there's
like
two
more
pieces.
I
need
to
be
wired
up,
I
would
say
unless
we
want
to
just
like
make
it
work
in
the
short
term
and
then.
A
We
usually
do
it
on
every
tuesday
of
every
month,
so
targeted
25th.
Probably
that
would
be
a
bit
tight
to
be
honest
to
get
in
those
two
prs
and
have
the
other
two
components
in
what
do
you
guys
think
yeah.
A
This
is
nice
to
have,
but
again
this
is
all
like
like
no
one's
using
this
so
or
sorry,
no
one,
no
one
should
know
about
it.
So.
A
I
don't
think
I
personally
don't
think
it's
a
big
deal
like
we
can
just
release
normally
next
week
to
keep
up
with
their
cadence
and
then
we'll
just
have
hopefully
a
end
to
end
by
next
month.
C
A
Yes,
I
think
that'd
be
a
good
idea.
If
we
have
time
and
I'm
thinking,
we
do
okay,
cool.
So
any
other
topics
you
guys
want
to
talk
about
that
are
not
pr
issue
related.
A
All
right,
so
this
is
the
periodic
metric
reader
spicer
conf.
You
just
want
to
talk
about
a
little
bit.
I
think
most
of
us
are
on
the
same
page.
C
B
Here
we
go
there.
There
was
one
issue
that
you
assigned
yourself
to
implement
the
metric
reader,
not
the
periodic,
like
the
issue
that
implements
the
color.
Is
it
there?
I
can't
find
it.
G
Yeah
there
was
a
metric
reader
issue
I
had
assigned
it
to
myself,
but
at
the
moment
I
was
thinking
about
pretty
much
taking
what
was
in
the
prototype
and
putting
it
into
a
metrobuilder
class.
I'm
not
sure
if
that
issue
is
relevant
now
after
or
if
this.
If
this
other
issue
of
a
pair
of
yours
fixes
it
all
together.
B
Uh-Huh,
okay,
so
it's
better!
This
gets
get
satisfaction.
G
Yeah,
maybe
you
can
mark
mark
your
prs.
That's
also
fixing
this
issue.
G
Yeah
when
we
had
these
these
prototypes
remember-
and
I
was
pretty
much
splitting
it
into
several
prs
and
creating
an
issue
for
this.
So
if
I
remember
correctly,
this
was
those
those
cases,
but
that
that
was
time
ago,
and
if,
if
this
pr,
the
sequence
is
going
to
include
the
abc
class
and
implement
the
final
collide
method
and
everything
else
is
pretty
much
fixed
like
that.
A
Kidding,
okay,
cool
yeah,
so
yeah
everyone
who
you
know
it
has
been
working
on
metrics.
I
think
this
would
be
a
good
pr
to
look
at.
I
have
I'm
still
looking
at
it
as
well.
So
awesome,
let's
just
try
to
get
this
in
real,
quick.
Any
other
comments
on
this
pr
that
people
want
to
address.
D
I
I
have
an
issue
that
I
could
use
helpful,
so
I
created
three
issues.
I
could
maybe
I'm
guessing
at
the
issue
number
based
on
when
I
created
them,
but
if
you
can
just
put.
D
Jeez
alex
this
is
just
breaking
up
the
it's
nice,
never
mind,
it's
not
that
complicated,
yeah
yeah.
No.
This
is
just
breaking
up
the
different
instruments
that
are
needed
to
support
otlp.
D
So
I
added
some-
and
I
just
I
was
hoping-
maybe
I
could
split
the
work
off
so
that
other
people
can
also
work
on
the
other
metrics.
I
don't
have
to
take
it
all
on
myself.
D
I
think
that's
all
the
instrument
we
need
to
care
about,
because
all
the
other
instruments
I
think
have
been
deprecated
like
in
in
histogram
and
in
some
and
all
those
other
ones.
We
don't
care
about
for
hlp.
So.
D
Yeah
yeah
sorry,
I
should
have
new
market
as
then.
D
Add
it
so
yeah,
so
if
people
are
looking
for
something
to
do
that's
fairly,
it's
fairly
scoped,
adding
adding
support
for
these
might
be
a
good,
a
good
task.
C
D
Exactly
it's
it's
literally
just
taking,
what's
already
in
the
instruments
and
yeah
making
them
available
through
the
exporter,.
A
Alex,
if
for
anyone
who
creates
metrics
issues
in
the
future,
if
you
can
mark
it
like
this,
that'd
be
great,
just
add
it
to
the
project
and
add
the
labels
as
well.
You
can,
you
can
add
it
to
the
project.
C
D
E
A
All
right
question:
diego
for
this
measurement
mechanism:
select
measurement
consumer;
anything
of
contention
that
you
want
to
bring
up.
G
It's
it's
aaron.
I
don't
have
a
comment
in
the
issue
that
it's
to
be
fixed
by
that
okay,
you
can
take
a
look
at
it.
Sorry
yep
corresponding,
so
I
think
pretty
much
whatever
means
is
that
we
can
do
this
later.
C
Yeah,
and
not
not
just
like
as
a
way
to
optimize
our
time
just
more
like,
I
don't
think
we
should
expose
it
until
we're
pretty
happy
that
we
want
to
have
it
around
forever.
G
You
know
yeah
also.
This
is
also
the
reason
why
I
did
so.
I
think
we
can
remove
the
the
the
1
0
10
r
c
10
label
from
that
issue.
I
think
it's.
A
Right
now,
right,
okay
sounds
good.
Did
you
want
to
either
mark
this
as
draft
or
like
comment
on
this
pr
that
that's
the
stance
that
we're
taking,
possibly
because
we
might
not
have
measurement
consumer
in
the
future?
I
don't
know.
G
I
mean
what
happens
is
that
the
design
as
it
is,
is
based
on
the
idea
that
the
machinery
consumer
is
a
component
that
can
be
swapped
yeah,
so
so
it
it
needs
a
way
of
to
make
it
select
a
little
bit
better.
Yeah.
A
Maybe
maybe
I
misspoke
it's
just
a
way
so
that
anyone
who
sees
this
pr
will
will
know
our
stance
that,
like
you
know
it
might
not
need
to
be
addressed
immediately.
I
think.
G
Well,
if
you
want,
we
can
close
this
vr,
because
I
mean
who
knows
which
implementation
we're
gonna
decide
on
the
future.
We
just
leave
the
issue.
A
Okay
cool,
so
you
kind
of
also
want
to
get
this
vr
in
all
right.
This
one
was
pretty
straightforward.
E
A
G
G
All
the
known
I
mean
I,
I
made
the
collect,
always
return
appointment:
okay,
yeah,
no.
A
Yeah
thanks
guys
yeah
we're
making
pretty
good
headway
with
all
these
metrics
prs,
guys
getting
merged.
G
A
Yeah
this
this,
this
stand-up
idea
that
aaron
had
is
really
lighting
a
fire
under
people,
because
we
can't
have
nothing
to
say
right,
yeah,
okay,
cool
all
right,
so
any
other
issues.
Pr
is
anything
you
want
to
talk
about.
G
Yeah,
okay,
so
that's
fairly
new.
I
think
I
opened
it
last
night.
I
think
it
is
yeah
it
pretty
much
as
that's
the
view
class
and
the
selection
mechanism
for
instruments.
So
please
take
a
look.
G
G
G
C
No,
no
just
please
take
a
look
at
that.
Okay,
were
you
planning
to
also
do
I
guess,
like
histogram
as
well.
G
Can
you
click
on
the
on
the
corresponding
issue?
Please
later?
Okay,
yeah.
I
marked
that
as
draft,
because
in
this
issue
you
mentioned
two
things:
first
create
a
component
to
handle
okay
scroll
up
yeah.
I
think
that
pr
addresses
the
second
point.
I
am
not
that
sure
what
you
meant
with
the
first
point.
C
I'm
trying
to
remember
so
so
like
the
way
it
works
in
java
is
they
have,
I
think
one
there's
called
like
temporal
metric
storage
or
something
like
that
and
that's
the
thing
that
holds
the
previous
metric
point,
the
cumulative
previous
metric
point
and
it
does
the
temporality
conversion.
So
it's
just
more
of
a
suggestion
like
I'm,
I'm
fine
to
do
it.
However,
I,
like
the
just
writing
a
function,
is
a
component
as
well.
I
guess
so.
Okay,.
G
Okay,
okay,
okay
and
then
we
can
mark
it
as
that's
ready
for
review,
not
as
rough
because
I
had
it
as
draft
only
because
I
wasn't
sure
about
the
first
point.
So,
but
if
this
is
enough,
that's
later,
can
you
mark
that
pr
is
ready
for
release
sure.
C
Yeah,
I
think,
can
you
also
add
the
two
pr's
to
the
review
in
progress
and
then.
G
Oh
yeah
I'll
do
I'll
I'll
go
through
all
my
pr's
and
make
sure
that
they
the
issues
and
they
are
up
to
date
with
the
dashboard.
Sorry,
I
I
didn't
know
about
this
wonderful
thing.
I
know
it
was.
A
G
A
Hey
aaron
just
curious:
if,
if
we
don't
have
a
component
to
handle
this,
where
would
the
storage
occur
for
keeping
track
of
like
previous
stuff.
C
A
G
G
G
Yeah
I
I
approve
that
and
commented
there.
Oh
thanks.
I
just
added
a
verification
of
what
the
error
handling
mechanism
will
do.
I
just
want
to
make
it
clear
that
your
handling
mechanism
will
not
implement
these
checks.
Yeah.
A
Nice
thanks
yeah,
it's
it's
very
simple
pr.
If
someone
can
just
like
take
a
look
at
that,
that'd
be
great.
A
Back
to
projects,
okay,
we
have
like
a
little
bit
under
10
minutes
to
do
kind
of
a
standard
b
kind
of
thing.
I
guess
we
just
go
around
talk
about.
We've
worked
on
for
me.
I
I'm
I'm
just
tackling
the
unassigned
api
sdk
stuff
going
through
them,
trying
to
like
separate
myself
from
the
any
conflicting
prs
that
are
out
right
now.
So
that's
pretty
much
what
I'm
doing.
A
Do
I
choose
someone
or
is
that
how
it
works
to
go
next?
Okay,
there
you
go.
G
All
right,
so
I
I
I've
pretty
much
been
reviewing
matrix
pr,
so
I've
been
trying
to
do
that
as
fast
as
possible.
Please
ping
me
directly
if
you,
if,
if
I'm
not
reviewing
any
pr
of
yours,
I'm
trying
to
do
that
for
every
metric.
C
G
I
opened
the
uspr
recently
and
now
I
am
planning
on
working
on
the
views.
Storage,
pr,
that's
pretty
much.
What
I'm
what
I'm
doing
right
now.
I
think
I
already
assigned
myself
to
the
issue.
C
Sure
so
I
did
the
I
I
did.
I
think.
C
Where
there
was
the
infinity
value
at
the
end,
which
we,
which
isn't
how
I
should
look
into
tlp
and
then
I
also
did
the
I
implemented
the
synchronous
measurement
consumer
so
and
I
added
a
stub
for
the
metric
reader
storage,
which
is
the
next
thing
I'm
going
to
work
on,
because
in
that
one
and
then
I
think
that
and
diego's
the
issue.
Diego
just
said
he
picked
up
with
the
view.
Storage
should
complete
the
the
whole
pipeline.
I
A
G
Okay,
I
just
want
to
sign
myself.
I
I
had
assigned
myself,
but
I
was
just
a
bit:
no,
no,
that's
fine!
I
haven't
done
any
work
on
that.
I
was
just
a
bit
confused
because
the
the
view,
the
views
and
new
storage
things
were
kind
of
very
related
to
that.
But
but.
A
G
I
mean
it's
not
that
critical,
I
I
I
prefer
view
instrument
match
because
it
we
create
one
of
these
objects.
Every
time
there
is
a
match
between
an
instrument
and
a
view,
so
we
at
the
end
we
have,
as
many
view
instrument
match
objects
as
much
matches
the
views
and
instruments
there.
I
don't
prefer
to
to
use
storage,
yeah.
G
A
E
A
A
E
G
Yeah,
you
can
rename
it
view.
That's
I
I
opened
it
the
time
ago
with
the
name
I
had
in.
G
C
No
worries
the
only
other
thing
I
was
going
to
do,
and
let
me
make
sure
it's
actually
in
here
is
the
make
necessary
metric
symbols.
Private.
C
C
Okay,
cool
sounds
good
cool
and
then
I
guess
alex.
D
I
haven't
done
anything
else
and
the
one
otlp
metric
that
I've
set
up
and
added
those
issues,
so
I'll
probably
be
tackling
the
next
metric
point
data
point
type
on
the
list
that
I
have
and
then
kind
of
move
from
there
as
I
go.
I
also
I
created
a
doc
for
examples.
Last
I
don't
know
if
it
was
last
week
or
something
oh
yeah,
that
was
it
would
be.
D
It
would
be
awesome
if
we
could,
like,
as
we
continue
to
add
features
to
the
metrics
sdk,
continue
to
update
that
thing
to
allow
us
to
have
like
a
working
example,
since
all
of
us
need
the
working
example.
So.
D
Yeah
yeah,
that's
that's
the
one
which
actually
allowed
me
to
find
a
bug
in
the
logs
example.
At
the
same
time,
so
that
was
cool
who
hasn't
gone
to
chican
did.
F
B
Yeah
I
had
been
focusing
on
the
metric
reader
lately.
Besides
reviewing,
I
can
pick
something
new
once
I
get
this
done
this
one
change
that
is
left.
A
A
All
right,
I
think
it's
pretty
much
everyone
that
was
involved.
Anyone
else
yeah
feel
free
to
pick
up
any
any
tasks,
we're
pretty
much
heads
down
on
trying
to
get
metrics
out
as
fast
as
possible.
So
yep
looks
like
that's
all
the
time
we
have.
Anyone
have
any
other
pressing
matters
before
we
part
for
next
week.