►
From YouTube: 2023-01-10 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
A
A
Okay,
I
think
we
can
start
so
yeah
I
added
the
first
couple
items
there
in
the
agenda.
One
is
I
learned
that
Auto
operator
is
implementing
opamp
I,
see
a
Jacob
in
the
call.
Is
this
the
Jacob
who
is
doing
the
work
for
the
operator
or
someone
else.
A
Okay,
anyway,
I
linked
to
the
pr
I
think
it's
cool
to
see
one
more
implementation
of
an
old
pump.
That's
that's
great
I
think
it's
it's
good
right
also,
since
we're
we're
also
doing
this
for
the
collector.
We'll
have
two
essentially
two
independent
implementations
of
wall
pump,
so
it
will
be
great
to
have
some
feedback
from
these
implementations.
A
It's
a
great
great
chance
to
fix,
whatever
needs
Peak
fixing
in
the
spec
or
the
gold
Library.
So
that's
a
that's
a
good
thing
that
is
happening
and
and
in
preparation
for
that
I
posted
a
draft
proposal
of
what
what
it
means
to
declare
all
pumps
stable.
A
The
pr
is,
there
I,
primarily
posted
it
there
so
that
we
kind
of
we're
ready
for
that
declaration.
We
know
what
exactly
we
mean
by
saying
stable.
So
please
take
a
look
if
you
have
any
thoughts
on
on
on
the
product,
buff
also
what
what
needs
to
be
declared
stable
in
the
protobot,
but
also
outside
the
protographs
right,
something
I
don't
know
in
the
spec
needs
some
sort
of
wording
of
what
it
means
to
introduce
newer
versions
whatever
right.
A
C
Yeah,
so
there's
been
some
concern
around
us
making
sure
the
agent
of
servant
server
to
agent
websocket
connection
is
healthy
and
trying
to
do
so
in
the
most
kind
of
cost-effective
way
and
there's
some
questions
around
hey.
Can
we
just
use
a
websocket
ping?
We
can
do
that
from
the
server
to
agent,
because
the
ping
pong
frame
will
just
work
with
the
way
that
the
gorilla
client
is
constructed.
C
However,
if
I
wanted
to
do
anything
at
the
application,
Level
I
don't
have
access
directly
to
the
websocket
client
I.
Just
have
my
the
op
amp
client
and
I'm
I'm,
not
I,
don't
have
access
to
say,
do
application,
Level
ping,
pong
I,
don't
know
if
anyone
else
has
thought
about
this
or
done
anything
in
that
space,
or
are
you
just
using
op-amp
payloads
to
detect
right
failures.
A
C
Okay,
yeah
I
mean
it
would
be
the
least
one
like
the
the
funny
thing
is
like
if
we
just
expose
the
underlying
websocket
client
I
could
have
just
done
it,
but
then
it
feels
weird
just
because
we
also
have
the
HTTP
client
right.
There's
no
parity.
There
yeah.
A
So
with
the
one
of
the
options
was
that
we
make
it
an
option
in
the
settings
when
you
start
the
old
pump
Clan
right,
you
say
whether
you
want
it
to
do
the
Keeper,
Live
ping
pongs,
and
maybe
you
specify
a
duration
or
something
like
that.
Yes,
this
is
what
we
want
to
do.
I.
C
Okay,
this
is
good,
so
I'll
follow
up
with
on
this
issue
and
specifically
and.
C
C
A
lot
of
this
is
coming
from,
like
part
of
the
our
engineering
team
who's
implementing
our
our
back
and
op-amp
server.
Implementing
the
websocket
server
we're
all
coming
from
other
areas
where
we
found
this
absolutely
necessary.
So
in
Sensei
we
found
it
absolutely
necessary
and
then
Engineers
that
came
from
other
companies
deploying
large-scale
websockets
were
like.
We
just
need
a
cheap,
frequent
or
infrequent
ping
pong
frame
to
keep
this
thing
up
and
healthy
yeah.
C
So
it's
all
coming
from
previous
experience.
It's
not
coming
from
a
particular
deployment
of
op-amp.
A
C
So
that's
that's
all
I
really
had.
Oh
the
other
one
I
forgot
to
add
to
the
agenda
was
for
the
Hotel
collector
contrib
op
amp
extension
I
was
wondering
where
we're
at
with
the
effective
config
for
extensions.
B
The
the
pr
and
The
Collector
core
repo
is
open
for
that.
It's
undergoing
review
right
now,
sick,
okay,.
A
B
A
And
the
approach
looks
good
to
me
and
I.
Think
Morgan
was
fine
with
the
approach
itself
as
well.
He
was
just
asking
for
the
I
guess:
removal
of
the
non-essential
bits
from
the
pr
which
were
not
necessary
right,
that's
blocky
was
objecting.
It
is
some
small.