►
From YouTube: 2021-11-04 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).
B
C
How
cold
is
crazy,
cold,
like
seven
degrees.
C
C
C
C
I
think
this
week
or
like
the
past
week,
I
think
logs
and
both
logs
and
metrics
were
merged
into
main
properly.
So
that's
pretty
awesome,
and
now
we
don't
have
to
go
through
the
that
headache
of
future
branches
anymore.
So
nice
yeah,
definitely
it's
exciting
yeah
so
and
it
looks
like
aaron
you've
been
pretty
active
on
the
metrics
stuff.
Awesome
work
on
that
with
diego,
oh
yeah.
I'm
not
sure
my
screen.
D
C
Okay,
all
right
topics
for
today
next
release.
C
Yeah,
that's
a
good
question.
We
probably
should
release
one
more
time
before
the
holidays.
I
guess
especially
if
it
will
contain
part
of
the
alpha
packages
or
the
experimental
packages.
A
So
so
the
api
like
in
theory,
you
could
probably
instrument
stuff
with
it,
but
the
sdk
is
not
usable,
there's
nothing
to
correct.
Actually
with
those
that
said,
it's
still,
it
is
all
on
the
experimental
package,
with
the
underscore
so
yep
is.
Is
that
a
problem?
Do
you
think
or.
C
C
Right
would
this
be
like
I'm
thinking,
I'm
trying
to
think
of
like
the
message
and
we
send
towards
the
people
that
were
already
waiting
on
metrics
if
we
release
this
package,
but
it's
only
like
the
api.
Would
this
confuse
people
even
more?
You
know
just
be
like
hey,
like
they've,
been
on
the
lookout
for
this,
and
if
we
do
any
kind
of
announcement
or
not
even
announcement
like
just
even
in
the
release,
notes
right
pointing
to
the
fact
that
we
released
some
metrics
thing.
C
E
I
guess
if
you,
if
you
were
to
message
just
like
in
the
release,
notes
that
the
api
can
be
used,
but
that
there's
no
way
of
actually
generating
metrics
with
the
with
the
current
package,
as
is,
and
maybe
that's
clear
enough
for
people
and
just
to
give
them
a
heads
up
that
that's,
I
guess
coming
up
soon.
C
Yeah,
I
think
it'll
be
clear,
it'll
be
just
a
kind
of
a
bummer.
You
know
like
a
lot
of
people
are
looking
like
following
that
thread
and
stuff,
but
I
guess.
C
E
A
Guys,
sorry,
I
was
just
gonna
say
I
can
update
the
documentation
to
say
to
say
that
as
well
and
if,
if
we
really
want,
we
could
like
like
what's
the
alternative,
would
that
be
waiting
to
release
or
just
not
releasing
it
at
all?
What
do
you
think.
C
E
A
C
Right:
okay,.
C
Logs
is
definitely
usable
yeah.
I
feel
like
that.
We
should
at
least
announce
so
people
can
start
trying
out
yeah
I'm
for
releasing
it
and
like
not
telling
I'm
sorry
I'm
for
releasing
metrics
and
not
telling
people
about
it.
But
that's
just
like
my
opinion.
A
C
Okay,
do
we
want
to
aim
for
a
specific
date,
maybe
sometime
next
week,
or
do
we
plan
on
releasing
we
plan
on
releasing
once
or
twice
before
the
year
ends?
Do
you
guys
think.
C
B
C
Okay,
cool
any
other
topics
that
people
want
to
bring
up.
A
Or
one
last
thing:
that's
slightly
related
on
those.
Do
we
want
to
delete
those
future
branches
or
just
like?
Let
them
sit
around.
A
C
Oh,
is
there
any
well
what's
the
repercussions
of
either
side
like
if
we
leave
them
versus
the
frequently.
A
I
think
it's
just
a
bit
confusing,
especially
since
we
have
metrics
and
metrics
new.
A
C
A
C
Okay
kind
of
funny,
I
can't.
E
E
C
Okay,
cool,
I
guess,
moving
right
along,
so
you
can't
put
up
a
pr
up
here.
Let's
take
a
look:
oh
yeah,
this
one
no
ice.
I
like
this.
Do
you
come
here
today.
F
Yeah,
it
seems
like
I
don't
know
it
already,
so
please
take
a
look
at
it.
Having
more
eyes
would
be
good.
One
like
there
were
a
couple
of
important
questions
that
were
brought
up.
B
Yeah,
but
I
guess
only
concern-
is
that
it
right
now
the
solution
would
only
work
on
linux
and
not
on
windows,
so
we'll
need
to
we'll
need
to
have
the
same
documentation
we
have
today,
but
make
it
conditional
to
windows
right.
F
Yeah,
that's
a
good
point,
so
these
these
libraries
that
you
know
that
we
are
targeting
right.
They
they
don't
already
like
support
windows,
they're,
also
like
mostly
3.7
or
later,
I
think,
like
yeah.
We
still
should
have
you
know
documentation,
but
I
don't
think
there
will
be
many
users
who
will
be
affected
on
the
window,
issues
who
will
be.
B
A
Yeah
yeah,
I
think
so
so
like
if
you
can't
fork
on
windows,
is
that
the
issue,
then,
how
does
like
g
unicorn
and
those?
How
do
they
make
new
processes
to
run.
F
C
Anyone
know
like
what
production
solutions
people
use
on
windows,
then,
if
not
celery
or
junior
corn
like
to
create
processes
and
stuff,
I
personally
haven't
seen
any
yeah.
Does
anyone.
C
I
work
on
github
dude.
Imagine
jeez
also
away
yeah
this.
This
reunit
solution
like
this
is
like
kind
of
like
the
work
around
that
like
customers,
who
come
to
me
like
I
just
tell
them
to
do
this.
Was
there
like
discussion
on
this
like?
Was
that
something
that
we're
not
trying
to
pursue?
C
Instead
of
doing
this,
of
attaching
them
to
this
hook?.
B
I
haven't
looked
at
this
pr
in
the
last
couple
of
days,
so
maybe
things
have
not,
but
in
the
beginning
the
the
two
solutions
we
were
comparing
was
the
reunit
thing
and
the
other
thing
was
delaying
creating
the
worker
so
lazily
creating
the
thread
which
had
some
downsides
that
yeah
there
had
some
downsides,
mainly
because
mainly
the
main
issue
was.
B
F
You
know
try
to
hack
something
already
existing
so
that
it
works
for
most
of
the
user,
but
then
you
know
we
we
we
thought
of
a
better
solution,
so
this
this
right
now
works
for
you
know
all
the
cases
I
have
like
a
repo.
You
know
attached
with
you
know
these
libraries
and
then
manual
this.
This
works
for
everybody
and
the
other
question
that
was
brought
up
was
what,
if
you
know,
people
are
already
using
the
hooks
provided
by
the
library.
So
will
there
be
a
double
instrumentation?
F
F
I
I
mean
I'm
I'm
not
sure
like.
If
other
you
know
reference
implementation
has
like
you
know.
Our
python
like
there
are
many
other
different
implementations
right.
G
A
Yes,
I
think
I
think
it's
kind
of
orthogonal,
because
this
problem
is
specifically
for
unix
fork,
because
the
any
any
of
the
processes
sorry
any
of
the
threads
that
are
not
the
the
thread
that
calls
fork
are
stopped,
so
it
doesn't
really
matter
so
much
the
threading,
it's
more
just
the
forking
problem
right.
Oh.
C
C
B
C
B
A
Actually,
I
think
what
we've
done
here
is:
we've
we've
replaced
that
with
just
reinstantiating
the
condition
with
a
new
lock.
So
you
can
see
it
right
there
like
in
230,
because
yeah
yeah,
so
we
can't
get
rid
of
that
code
because,
like
I
think
that
the
value
of
that
one
is
mostly
if
many
people
hold
the
condition,
you
can
just
reset
that
the
internal
state
once
but
since
we
are
the
only
one
who
has
this
condition
variable
it's
it's
okay
to
just
to
just
reinstantiate
it.
We
don't
need
to
change
the
internal
state.
B
D
F
A
Cool
yeah
one
other
thing
I
was
going
to
bring
up
here.
I
think
I
think
we
talked
about
this
before
at
some
point.
Is
this
kind
of
solution
won't
work
well
for
metrics?
If
you
think
of
something
like
a
prometheus
right,
if
you
fork
and
then
all
the
prometheus
endpoints
are
trying
to
host
on
the
same
exact
port,
find
at
the
same
port,
it's
not
going
to
work
right.
A
So
we
have
a
similar
problem
for
google
cloud
where
you
don't
want
to
push
metrics
like
that
at
the
same
time,
but
in
that
case
I
think
we're
going
to
need
to
also
maybe
think
of
something
even
more
like
like
either
a
single
process
is
collecting
all
the
metrics
and
and
telemetry
for
the
other
forks
or
we
force
people
to
use
a
collector
or
something
isn't
that,
like.
C
A
limitation
on
prometheus
like
that.
That
would
be
a
problem
regardless
of
like,
if
they're,
collecting,
telemetry
right.
Sorry,
regardless,
if
they're
doing
this
like
they
can't,
they
can't
use
multi-processing
with
prometheus
anyways
right.
C
A
Post
for
your
post
for
cook
could
could
use
a
different
port,
so
you
could
reinstantiate
the
exporter
to
use
a
different
port
or
something
like
that.
But
then
yeah,
like
the
thing
that
that
scraping
still
needs
to
know
where
to
read
it
from
I.
I
guess
I
I
see
your
point,
but
I
think
like
maybe
well
sorry,
what
were
you
going
to
say?
Shigan.
F
I
I
want
to
say
like
this:
this
was
you
know,
brought
up
in
the
collector's
seat
for
and
a
long
time
back
that
you
know
it's
not
just
python.
Like
other
six.
Also,
have
you
know
this
problem?
F
Maybe
maybe
there's
something
at
the
collector
level
that
they
can
do
so
that
you
know
the
language
kind
implementation
won't
have
to.
You
know
do
this
like
they
have
to
deal
with
this
thing
matrix
specifically.
C
C
That's
kind
of
a
it's
kind
of
a
weird
message
to
users,
though
right
like
like
someone
who
wants
to
like,
let's
say
even
for
python
right
like
someone
who
has
like
an
app
right,
that's
instrumented
with
like
prometheus
and
everything
like
it
works
standalone
and
they
deploy
it
but
like
then,
then
they
host
it
on
like
gene
unicorn,
and
then
they
have
this
like
multi-process
issue
right
and
then
our
recommendation
would
be
like.
Oh
now,
you
have
the
instrument
with
collector
right
kind
of
like
a
broken
story.
A
Yeah,
I
I
think
you're
right
that
it's
like
a
general
problem
to
prometheus
and
does
anybody
have
experience
using
prometheus
with
do
unicorn
and
and
how
like,
what's
like,
the
typical
fix
for
that.
C
A
C
C
A
C
A
C
C
So
I
think,
alex
just
linked
the
a
document
explaining
exactly.
C
A
That
yeah
alex
did
you
get
like
a
tldr
from
this
nope
I'm
reading
it.
C
Okay,
anyways:
we
we
don't
have
to
take
up
time
from
the
meeting
to
do
this,
but
yeah
interesting,
read.
Maybe.
C
Cool
all
right
doesn't
look
like
there's
any
other
topics
or
pr's
or
issues.
Does
anyone,
oh?
Does
anyone
else
want
to
bring
up
anything?
Do
you
want
to
talk
about.
G
All
right,
hi
yeah,
I
just
wanted
to
introduce
myself.
I
joined
hello
for
the
first
time.
G
Yeah
yeah
I'm
around
from
epsilon.
We
we're
acquired
by
cisco
lately
and
we're
getting
more
into
open,
telemetry,
so
just
curious
to
see
what
you
guys
talking
about
and
probably
will
get
into
the
code
in
the
following
weeks
and
that's
it
so
nice
to
meet
you
all.
G
C
Yeah
nice
to
see
you
feel
free
to,
like
you
know,
ping
us
on
the
slack
channel.
If
you
need
help
on
anything,
you
know
to
get
started
on
like
some
issues,
you
know
to
get
your
feet
wet
in
the
community.
A
C
Some
damn
sorry,
I
totally
didn't
see
you
so
usually
we
do.
G
Like
a
round
of
introductions,
yeah
yeah,
I
just
got
in
a
a
little
bit
late.
C
Yeah,
usually
we
have
more
stuff
to
talk
about,
but
today
I
guess
it's
a
just
a
short
meeting.
So
if
that's
it,
I
will
see
you
guys
next
week.