►
From YouTube: 2020-08-05 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
B
B
C
I
have
like
a
general
question
about
including
the
go
sum
files
and
the
churn
that
that
causes
prs
will
become
stale
pretty
shortly
after
being
opened
and
I'm
not
a
go
mod
expert
by
any
stretch,
I
just
kind
of
wonder
if
there
is
an
actual
benefit
to
including
them
if
they
do
change
so
frequently
on
their
own.
B
Oh
yeah,
I
think
goal
modules
do
recommend
to
include
gosum
files
in
the
source
controller
and
the
reason
is
the
bit
reproducibility.
B
C
I
guess
my
question
is
why
they
change
so
frequently
seemingly
unrelated
subprojects
or
like
receivers,
and
maybe
I
could
do
more
investigation.
That's.
B
D
B
B
B
B
B
B
D
D
Oh
man,
not
anytime
soon,
no
so
here
in
I'm
in
san
mateo
county
in
the
bay
area,
they
are
there
starting
to
shut
down
stuff
again,
I
I
wanted
to
go
and
get
my
beer
trimmed
and
I
had
an
appointment
and
because
they
had
allowed
them
to
be
open
like
for
the
last
month
or
so
barbershops
and
stuff,
like
that,
I
just
shut
it
back
down.
D
D
B
It's
we
are
calling
it
stage.
Three
most
places
are
open
with
some
restrictions.
I
would
say
it's
neither
neither
there
neither
there
like,
like
half,
half
open,
half
up
close,
but
still
doing,
I
guess
better
than
than
the
us
most
of
the
us.
D
So
I
am
going
to
try
to
put
the
you
know
the
proof
that
that
locks
over
a
few
thing.
A
B
If
you
had
any
any
comments
and
anything
that
you
wanted
to
be
changed,
you
can
still
file
an
issue
or
a
pr
against
it.
When
you
can't
fix
it.
B
B
Yeah,
so
there's
that
I
think
david
is
out
he's
on
pto,
so
I
don't
think
he's
joining
and
probably
I'm
not
sure
what
he's
up
to
right.
Now
he
was
working
on
the
logging
library
on
that.
So
let's
see
what
he
does
after
respect
from
the
from
the
following
yeah,
I
don't
have
anything
myself.
That
was
the
the
primary
thing,
the
logs
overview.
That's
a
good
thing
to
do
anything
from
you
or
from
others.
Maybe.
D
Nope
nope
nope
on
my
side,
it's
it's
mostly
about
you
know
trying
to
also
sort
of
find
somebody
in
our
end,
who
can
start
like
digging
into
the
code?
We
haven't
we're
not
quite
there.
Yet
our
guys
are
currently
you
know
digging
into
the
you
know.
You
know
tracing
aspect
of
open,
telemetry
and
engaged
there
and
you've
probably
seen
them
pop
up
there.
You
know
you
know
jimmy
can
and
those
guys
are
from.
We
have
a
bunch
of
guys
in
poland
that
are
working
on
that.
D
You
know
we're
certainly
also
going
to
be
super
interested
in
the
metrics
thing
and
then
obviously
this
one
right,
you
know,
but
the
sort
of
from
our
perspective
that,
like
in
terms
of
you,
know
prioritization
the
closest
one,
the
one
that
we
are
basically
the
closest
to
releasing
is
is,
is
related
to
tracing.
You
know,
okay,.
B
If
you
have
anybody
who
is
interested
in
metrics,
there
is
a
current
ongoing
active
discussion
around
the
metrics
protocol
in
the
protocol.
So
if
you
have
someone
who
wants
to
have
a
look
at
that,
I
think
that
would
be
useful.
F
Yeah,
we're
in
the
same
boat,
we're
building
exporters
for
metrics
and
logs
or
metrics
and
traces.
Sorry,
so.
B
Okay,
and
by
the
way
we
now
have
the
logs
as
a
publicly
supported
data
type
in
the
collective.
So
if
there
is
a
wheel
to
do
to
implement
the
logs
exporter,
it's
not
possible
to
do
that
as
well
or.
B
And
and
and
the
document
that
I
was
just
talking
about
the
overview
says
why
not
right
and
what
we,
what
we
suggested
instead
there's
another
agent
now,
let
me
actually
post
a
link
to
that
overview
document.
I
think
it's
it's
a
useful
thing
to
have
interesting.
B
G
No
I'm
just
as
soon
as
I
get
done
with
my
other
work.
I've
got
I'm
going
to
do
one
final,
like
kind
of
performance
testing
of
the
flip
bit
stuff
and
do
kind
of
the
comparison
of
see
exactly
what
the
overhead
is
of
using
going
everything
going
through
the
collector
versus
just
fluent
bit,
but
yeah.
That's
I
haven't
actually
been.
I
haven't
done
anything
additional
with
that
actual
implementation,
work.
F
A
B
B
We
have,
I
think
we
have
a
few
open
issues
in
the
collector
repository.
We
have
probably
a
couple
in
the
specification,
but
not
much.
The
things
that
we
want
to
have
is
basically
an
sdk
for
implementing
logging,
the
open
telemetry
way
in
logging,
libraries,
which
is
what
david
works
on,
and
he
he
also
has
an
open
pr
for
java
implementation.
B
Now,
I
think
for
us
to
ga
we
probably
need
to
at
the
minimum
have
that
at
least
one
complete
implementation
of
an
offer
for
logging
sdk,
that
that
also
supports
some
popular
loading.
Libraries
suffer
from
java.
That
would
be
probably
like
forgery
or
or
yeah,
or
some
yeah,
some
others.
Maybe
two,
I
think,
that's
bare
meaning
for
us
to
say
that,
so
we
need
to
have
a
specification
that
tells
how
to
use
the
loads.
B
We
need
to
have
at
least
one
major
language
supported,
and
I
would
aim
for
maybe
more
than
one
language
and
and
in
the
collective
vlogs
to
be
supported
as
a
data
type
that
you
can
receive
and
send
so
a
few
a
few
experts
at
least
to.
B
I
would
like
to
see
some
engagement
from
vendors
as
well
right
to
support
exports
to
them.
I
believe,
if
I'm
not
wrong,
we
from
splunk
will
be
adding
an
exporter
for
splunk's
protocol
and
at
least
we
want
to
see
a
bit
more
right
and
some
others.
F
F
B
Log4J
it's
what
is
it
called
tender
or
something
like
that
right
and
I
think
that's
what
david's
er
also
does.
I
am
not
completely
sure
so
he
probably
he
probably
knows
he
knows
more
about
that
thing.
So
yeah.
B
I
can
ask
him
about
the
state.
Okay
for
other
languages.
I
I
don't
know
because
there
is,
let's
say,
probably
the
next
most
popular
language
at
open,
inflammatory,
I'm
guessing
is
goal
and
for
goal.
The
situation
is
somewhat
different.
There's
no
single
most
popular
library
there,
and
there
is
there's
this.
I
think
the
problem
with
passing
the
context
in
go
is
also
less.
I
guess
it's
unclear
how
we
do
that
in
go
in
java.
You
have.
A
B
You
have
the
you,
can
read
the
context
from
thread
local
storage
and
go.
You
can't
do
that
and
logging
libraries
in
go.
They
don't
usually
take
context
as
a
parameter
to
logging
statements.
So
you
have
a
problem
that
if
you
want
to
extend.
B
You
can't
do
that
easily
what
that
means.
I
don't
know
we
need
to
think
through
that
whether
that
means
we
need
to
introduce
a
new
api,
possibly
for
go,
but
that's
what
we
were
trying
to
avoid
at
least
initially,
some
thought
is
needed
here.
Whoever
is
interested
in
that
part
of
the
project
and
the
libraries
and
the
apis.
B
Any
help
here
is
appreciated.
I'm
kind
of
more
focused
on
the
you
know
the
overview
and
the
collector
side
of
things
at
the
moment.
D
No,
I
said,
sounds
good
go
ahead
so
guys,
because
I
think
I
kind
of
get
a
sense
of
where
you
are
tonight
is
coming
from
there
and
and
like
a
similar
sort
of
you
know,
concerns
on
our
end
as
well
and
in
terms
of
you
know
like
how
do
we
make
it
real?
We
want
to
make
it
real.
How
real
is
it?
You
know
where's
the
chicken
where's
the
egg.
You
know,
what's
the
pork
et
cetera
so
should
we
maybe
work
towards
having
a
kind
of
punch
list
of
sorts?
D
That
would,
you
know
basically
start
with
what
you
just
kind
of
said
to
basically
define
what
a
ga
would
look
like,
and
then
you
know,
maybe
we
can
also,
to
some
degree,
maybe
parcel
out
some
of
these
tasks
among
the
folks
that
we
have.
D
You
know
I
don't
know
so
that
not
not
every
vendor
has
to
redo
everything.
You
know.
Maybe
we
can
you
know,
I
don't
I
I
personally.
Actually,
I'm
just
realizing
as
part
of
this
conversation,
I
have
the
desire
to
sort
of
understand
that,
but
I
don't
really
have
a
very
good
view
on
that,
like
what
we
would
consider
would
be
needed,
and
I
think
when
you
have,
you
have
a
pretty
solid
idea.
You
know:
do
you
think
it
makes
sense
to
kind
of
work
that
into
something
a
little
bit
more
official.
B
Yeah,
I
think
it
makes
sense
in
a
way
so
this
overview
document-
I
guess
in
a
way
says
that
but
in
indirect
right,
so
you're
saying
have
a
more
specific
list
of
things
that
we
would
want
to
have
sort
of
a
list.
Yeah.
D
D
Yeah
the
overview
is
it's
really
nice,
but,
like
you
know,
there's
going
to
be
a
lot
of
tltr
issues
and
people
are
going
to
look
at
this
and
like
what
do
you
want
to
hit
me
for
I
can
people
can
barely
focus
on
the
tweet
right
so
and
I
don't
mean
to
sort
of
just
push
all
the
work
back
to
you,
and
this
is
also
something
that
you
know
like
somebody
else
could
pick
up.
You
know
and
and
see
whether
we
can
create
it.
D
You
know
off
the
of
the
overview,
but,
like
you
just
rattled
down
about
like
you
just
rattled
off
a
bunch
of
things
now
I
wish
I
would
have
had
the
chance
to
write
them
down,
but
so
any
which
way
you
want
to
do
that.
You
know
I'm
happy
to
describe
if
needed.
You
know,
others
can
do
it.
We
can
start
from
somewhere
and
then.
A
B
Okay,
yeah,
I
think
that's
that's
a
good
point.
I
wouldn't
want
to
just
go
and
write
that
that
list
on
my
own
only
I
would
want
more
participation
from
the
community
as
well
from
you
guys
and
from
others,
so
that
it's
it's
I
I
don't
want
it
to
be
limited
only
to
what
we
think
is
necessary
for
for
us,
but
maybe
that's
not
really
the
most
important
thing
right
now,
despite
us
for
being
important
for
us
right.
B
So
let
me
maybe
maybe
I
guess
I
can
try
to
come
up
with
a
draft,
but
then
invite
some
more
discussion
more
more
to
to
to
give
input
on
that
yeah.
I
think
it's
a
good
point
to
have
sort
of
a
more
specific
target
towards
which
we're
working
right.
B
All
right
yeah,
that
was
good.
Could
anybody
else
have
any
thoughts.
H
Hi
first
time
here
I
just
wanted
to
say
that
the
the
issues
of
how
do
you
get
the
context
to
the
right
place
in
the
log,
I
think,
is
very
important.
H
I'm
mostly
coming
in
from
a
developer
point
of
view
of
using
the
logs
and
trying
to
get
other
developers
on
our
teams
to
get
the
log
and
context
data
into
our
observability
platforms,
so
yeah.
I
just.
I
just
want
to
reiterate
that.
H
I
think
it's
very
important
to
be
able
to
find
the
right
solution
that
causes
least
amount
of
friction
in
adopting
the
the
libraries
and
plug-ins
and
stuff,
because,
like
right
now,
I'm
looking
at
node.js
and
people
are
saying
you
know
just
pass
your
logger
object
into
every
function
and
pass
your
or
pass
your
request
id
into
every
function,
and
that
just
doesn't
seem
very
realistic
to
me.
It
might
be
the
only
way
to
do
it,
but
it
doesn't
feel
right
right
to
be
having
a
dependency
on
my
logger
everywhere.
H
There's
other
ways
of
people
are
mentioning
dependency,
injection
tools
and
all
these
other
fancy
things-
and
I
don't
know
it's
a
tough
nut
to
crack.
It
definitely
needs.
B
B
Likely
going
to
be
language
dependent
right,
there's,
no
one
single
answer
that
works
for
every
language
which
to
me
means
that
the
specific
language
sigs
need
to
think
about
how
they
do
that,
and-
and
they
have
this
for
this
very
similar
problem
for
traces
and
metrics
as
well.
So
we
assuming
they
solved
it
for
traces.
Whatever
the
solution
is,
it
should
be,
should
work
for
the
logs
as
well.
H
Yeah
for
traces,
it's
a
bit
easier
because
you
always
have
the
request
context
on
a
on
a
trace,
so
you're,
always
within
the
context
of
a
network
request
right.
It's
very
rare
that
you
have
a
trace
that
goes
between
function
calls,
but
for
logging.
It's
very
often
that
you
just
want
to
log
something.
That's
deep
in
your
function!
Call
that
isn't
necessarily
related
to
a
a
trace.
Specifically
right.
You
can
correlate
it.
There
is
going
to
be
some
trace,
that's
related
to
it,
but
often
like.
H
If
you
have
a
sub
process
or
or
a
worker
on
the
side
right,
you
may
not
always
be
able
to
correlate
that
to
a
very
specific
request
and
that's
what
and
that's
exactly
why
you
want
the
log
and
not
the
trace
but
yeah.
I
I
agree
that
it
could
be
done
in
the
six,
because
I
I
also
agree
it's
gonna
be
language
dependent,
but
if,
if
we
have
patterns
of
how
to
do
it,
right
maybe
go
is
not,
may
not
be
the
best
first
language,
because
it's
harder
to
do
it.
H
It
might
be
better
to
start
with,
like
java
and
c
sharp,
where
it's
very
easy
to
have
a
a
thread
thread
variables
and
you
know
yeah.
You
have
a
context
very
easily
grabbed
right.
Those
have
lots
of
log
libraries,
it's
very
clear
how
to
do
it
there
and
more
modern
languages.
Maybe
you
could
do
those
after.
B
B
It
may
be
just
the
portion
of
the
implementation,
but
the
idea
was
the
intent
was
at
least
that
there
is
going
to
be
an
appender
which
also
takes
care
of
the
context
of
the
request
context
and
includes
the
the
trace
of
this
panel.
The
the
staffing.
H
It
I
I
think,
if
we
could
reach
out
to
oh,
I
forgot
the
name
of
the
company
honeycomb,
so
I
think
their
b-line
library
does
a
lot
of
that
context
grabbing
in
a
log
context
as
well.
So
maybe
there's
some
things
you
can
learn
from
there.
H
Just
just
in
general,
because
they
have
they
have
their
library
for
a
lot
of
different
languages
and
seeing
how
they
combine
the
metrics
and
the
and
the
logs
together
and
the
events
right.
They
focus
all
around
events,
but
seeing
how
they
think
about
it
might
help
us
also,
because
I
I
don't
know
how
they
do
it.
I
just
I
just
know
that
they
do.