►
From YouTube: 2022-07-05 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
B
Hi
youtube
seems
like
it's
to
be
a
quiet
meeting.
A
Yeah
and
cjo
said
he
wouldn't
be
available
for
the
next
like
three
weeks.
B
A
Like
I
mean
he
would
he
might
be
available,
but
he
would
probably
he
doesn't
like
he's
not
sure
of
his
schedule.
So
he
has
asked
me
to
host
the
meeting
for
like
the
coming
three
weeks,
including
yeah.
That
makes
sense.
C
B
B
B
Sure,
absolutely
so,
I
think
I
capture
everything
here
on
this
page,
but
the
gist
of
it
is.
I
want
to
work
on
the
code
coverage
for
this
repo
and
just
to
increase
the
overall
code
coverage.
I
spoke
to
riley
about
this.
He
said
that
he
thinks
a
healthy
repo
would
be
around
the
92
to
95
coverage
mark
okay
and
well.
We
got
plenty
of
areas
for
improvement,
so
just
code
coverage
by
itself.
B
I
wanted
to
target
the
the
one
a
there
is
to
identify.
First,
what
are
the
most
important
libraries
I
think
open
telemetry
makes
sense.
That's
the
you
know
the
one
library
everything
else
was
built
off
of
and
siegel
proofread
this
and
he
suggested
the
iltel
hotel,
otlp
exporter.
In
addition
to
that,
if
there's
anything
else,
people
think
would
be
important
very
eager
to
get
some
input
from
the
community.
B
Then
these
would
be
trying
to
get
it
as
close
to
100
as
possible
like
if
these
are
like
the
most
important
prioritize,
just
getting
these
as
much
covered,
and
then
everything
else
would
just
try
to
get
to
the
90
mark
and
then
reassess
when
we
get
to
there.
The
one
thing
to
bring
to
attention
is
this
may
require
some
tests
that
are
rather
trivial.
B
What
I
mean
by
that
is
I've
already
gone
through
the
code
coverage
and
also
the
visual
studio
has
a
similar
tool.
Show
code
coverage
and
there's
things
that
are
missing
is
like.
If
there's
an
if
else,
block
well
the
if
is
fully
tested,
but
then
else
would
might
just
return
a
default
value
and
there's
no
test
cases
around
that
or
if
there's
a
try
catch
somewhere
and
then
that
catch
throws
an
exception
or
something
extra
happens.
B
B
And
the
second
point
to
discuss
is
just
the
reliability.
Riley
had
brought
these
to
my
attention.
He
had
created
two
issues,
that's
what
you
see
in
blue
at
the
bottom,
the
prometheus
exporter
and
then
the
incoming
requests.
So
they
were
flaky
some
time
ago
they
were
failing.
I
guess
a
lot.
B
B
A
B
And
I
think
I've
seen
some
other
ones
fail,
but
then
I
didn't
write
them
down,
so
that
would
be
sort
of
an
ask.
The
community
is
to
help
like
hey.
If
you
see
something,
if
it's
you
know
impacting
your
ability
to
get
prs,
you
know
merged
open,
an
issue
report
it
somewhere,
so
that
we
can
do
some
more
digging
into
how
to
fix
it
or
how
to
replace
the
test.
A
Yeah,
and
also
I
I
don't
think
it
like-
we
should
focus
much
on
fixing
api
combat
because
we
would
want
to
move
to
the
new
package
validation
tool
anyway.
So.
B
Right,
I
think
that's
what
was
decided
on
that
on
that
thread.
B
So
but
this
is
it
if
anybody
has
any
feedback
or
any
comments,
and
the
result
of
this
is,
I
start
just
opening
more
prs
to
improve
code
coverage.
I
think
you've
seen
a
couple
I
know
like
alan
has
commented
on
the
one
that
I've
got
open.
Thank
you,
alan
for
helping
me
with
that.net
format,
thing.
C
I
don't
have
anything
that
I
think
this
looks
great.
In
my
opinion,
my
code
coverage
is
great
and
yeah
the
reliability,
especially
that
prometheus
one.
I
think
that
that's
definitely
the
one
that
I've
seen
the
most
pretty
sure
so.
Okay,.
D
We
do
mothra
have
some
integration
tests
that
we
run,
those
probably
aren't
making
their
way
into
the
code
coverage,
but
you
might
be
able
to
turn
on
code
coverage
generate
whatever
the
output
is
like
json
or
xml,
or
something
take
it
out
of
the
container
and
then
merge
it
with
the
results
that
might
help
you
get
where
you're
trying
to
go
and
it
might
yield.
You
know
more
useful
tests
than
just
trying
to
like
mock
all
that
stuff,
because
integration
tests
are
actually
executing
logic.
B
And
yeah,
that's
that's
it
for
me
for
today.
Thank
you.
C
B
C
I
find
the
I'll
plop
it
in
the
agenda,
so
everybody
has
a
link.
It's
an
issue.
C
I'll
just
summarize
so,
oh
this
is
yeah
this
one.
So
basically
our
otlp
exporter
package
has
it's
compiled
a
number
of
different
ways.
The
net
462
version
relies
on
the
old
grpc
library,
which
I
think
is
end
of
lifeing,
maybe
not
immediately,
but
soonish,
maybe
and
anyways.
Also
it
has
like
these
these
native
dependencies
that
are
large
and
they
get
copied
around
and
so
on.
C
People
aren't
really
that
excited
about
them.
So
what
this
person
is
suggesting
is
like
hey.
You
know
now
that
it
supports
not
standard
2
or
sorry,
the
the
the
the
grpc
net
client
package,
which
is
fully
implemented
in
c
sharp.
It
doesn't
have
those
native
dependencies
and
so
on.
That
is
supported
by
netstandard2o
cool,
though
there's
I
think,
there's
some
caveats
there
that
I
I
don't.
I
don't
fully
understand
like
it,
needs
to
be
running
on
a
certain
version
of
windows
and
blah
blah.
C
Something
or
another,
so
so
that
said,
I'm
not
I'm
not
ready
to
just
say
like
well,
let's
just
remove
the
net
462
compilation,
but
I
was
about
to
suggest
to
this
guy
a
workaround.
C
Maybe
so
you
know
if
you
haven't,
if
you
haven't,
if
you
have
a
net
framework,
targeted
library
or
application,
and
you
reference
a
nougat
package
and
that
nuget
package
has
say
like
a
net
462
and
a
net
standard
2o,
it's
going
to
choose
the
net
462.,
but
I
think
a
workaround
would
be
to
like
if
they
can
somehow
like
refactor
their
code
and
like
push
like
their
otlp
exporter,
configuration
into
a
library
that
targets
netstandard,
2o
and
then
reference
the
our
nougat
package
from
there
they're
going
to
get
the
right
one
or
the
one
that
they
want.
C
A
Yeah
so
like
for
net
framework,
it
will
pick
up
the
net
462
compile
code,
so
we
don't
want
like
we
would
pull
in
the
older
grpc
client.
Is
that
right,
yeah
yeah?
Exactly
I
I
do
remember
like
like
we
had
a
similar
issue
long
time
ago
with
sql,
instrumentation
or
asp.net,
and
I
think
we,
where,
like
yeah,
something
similar
where
there's
net
462
uses
a
different
net
framework,
was
using
a
different
kind
of
library
and
net
standard.
2.0
and
data
were
using
something
else
but
yeah.
I
think.
D
D
D
D
C
If
that's
would
be
a
great
thing,
I'd
want
to
try
it
out.
First,
I
guess
before
suggesting
it
just
make
sure
I'll
see
if
I
can
find
it.
C
But
yeah
that
that
would
be
cool.
But
what
do
you
think
about
the
other
suggestion
of
just
like
I
mean
it's?
It
may
not
work
for
everybody,
but
if
he
can
split
up
his
code
a
bit,
you
know
maybe
somehow
just
just
put
his
otlp
configuration
code
into
a
separate
library
that
he
then
separately
references.
I
think
that
that
would
work
as
well
or
do
you
think
that
that
would
run
into
problems?
A
A
All
right,
then,
I
think
we
can
end
early.