►
From YouTube: 2022-05-27 meeting
Description
Instrumentation: Messaging
A
C
B
C
A
Hello
tyler
before
we
started,
can
I
ask
you
something
about
transform
processor
yup,
so
we
have
an
issue
with
a
pretty
big
customer.
A
C
C
C
A
So
my
question
is:
how
far
away
we
are
from
introducing
like
metric
specific
functions
to
the
transform
processor.
So
essentially,
we
need
aggregation
at
least
and
removing
some
metrics
based
on
the
actually
removing
attributes
stuff
like
that.
So
we.
E
C
A
C
Think
hold
on
you:
can
we
don't
have
a
drop
attributes
function?
We
have
a
set
which
you
could
use
to
make
an
attribute
blank
and
you
can
limit
the
number
of
attributes
and
you
can
truncate
attributes
so
that
they
don't
go
over
a
limit,
but
yeah
we
do.
We
don't
have
a
drop
function.
Yet
there's
a
lot
of
uses
for
a
drop
function.
We
should
probably
make
one
soon.
I
see
that
you.
C
One
yet
either
it's.
C
The
it's
pretty
unstable,
I
would
say.
C
Alpha
because,
okay,
the
the
issue
that
we
talked
about
on
the
sig
with
the
enums,
is
a
pretty
important
one
like
some
of
the
some
of
the
fields
that
you
might
want
to
use,
just
like
don't
like
they're,
not
actually
usable,
like
you,
couldn't
reference
them
in
a
function
correctly,
and
then
there's
some
also
just
some
defects
with
some
fields
that
they're
being
returned
and
data
types
that
don't
mean
anything
in
the
parser.
So
like
you
can't
actually
compare
anything
or
if
you
tried
to
set
a
data
type
it
like
wouldn't
work.
C
Okay,
so
I'm
that
one's
a
p1,
that's
the
one
I'm
working
on
right
now
so
yeah
like
it's
definitely
alpha
like
you
use
at
your
own
risk,
like
things
may
change
like,
but
you
could
definitely
add
a
function
and
like
it'll
work
and
like
the
functions
are
pretty
isolated
so
like
as
long
as
no
one
screws
with
your
function,
your
function
itself
should
be
pretty
stable
but
like
there
are
some
things
that
customers
might
want,
like
the
conditions
like
there's
no
less
than
there's
no.
B
B
A
C
So
in
in
a
condition
you
could
say
where
attribute
like
attribute
a
is
equal
to
something
and
you
could,
but
you
could
only
do
it
one
time
and
you
can
do
equals
two.
B
D
A
C
Production
use:
okay,
yeah
for
production
use
for
sure
it
sucks
to
add
things
to
the
other
processors
that
like
we
want
to
get
rid
of
but
like
at
the
same
time.
The
transform
processor
like
alpha
is
the
right
state
for
it,
but
for
like
a
large
customer
that
maybe
wants
some
stability.
Transform
is
like
probably
not
there
yet.
D
A
A
I'll
probably
be
contributing
to
like
replace
the
functionality.
C
A
little
bit
but
yeah,
I
like
it's
my
primary
focus
right
now,
I'm
trying
to
focus
on
getting
the
thing
stable
and
not
so
much
on
the
on
like
adding
new
functions
myself
yet,
but
I
think
when
we
do
start
trying
to
replace
existing
processors,
we
should
have
like
large
overarching
issues
that
track
that,
but
I
haven't
created
any
of
them.
Yet
alex
was
asking
me
about
that
yesterday
and
I
was
like
we
want
to
like,
but
I
just
haven't.
I
haven't
made
it
that
far.
A
C
Especially
but
anyway,
it's
like
the
same
problem
with
the
logs
transformer
log
transform
processor.
It's
like
we
don't
want
to
rewrite
that
entire
thing
and
then
also
do
it
again
in
the
transform
processor
like
when
that
thing
emerges,
we'll
just
do
it
one
time
at
the
transform
processor,
I
think
that's
kind
of
how
it
is.
A
C
A
Down
and
I
believe
boxing
will
be
gone
like
a
limited
availability
in
june
as
well,
so
we're
like
we're
a
bit
on
the
pause
there.
Are
you
going
to
document?
A
I
am
not
I'm
moving
to
yes
back
to
yes
and
it's
like
I'm
splunk
pulled
me
into
some
other
stuff.
I
mean
because
of
some
like
creepy.
I
don't
have
enough
resources
in
the
world
and
one
part
of
the
splunk
only
yeah,
I
hate
it,
but
I
have
to
do
it.
D
Temporary
okay,
let
me
share
my
screen.
B
A
B
B
Should
be
the
last
linting
related
thing
for
package
tanza,
nice,
nice.
D
B
It's
anthony
mirabella,
aneurysm,
okay,.
D
D
It
doesn't
seem
critical,
I'll
put
priority
2
at
this.
One
should
be
fine
and
do
we
have
parameters
export
there.
C
A
C
C
C
D
I
like
what
fake
fast,
let's,
let's
ask
for
help
here,
the
first
issue
I
believe,
and
the
priority,
probably
I
don't
know,
flakey
test.
What's
the
priority
for
like
sp3
or
p2
is.
D
D
C
You
know
what
dan
I
saw
the
other
day
on
one
of
my
pr's,
a
load
test
failed
again.
So,
despite
all
that
stuff,
you
just
did
for
load
test.
I
don't
know
if
it's
merged
or
not
yet.
Hopefully
it's
not
merged
yet
because
I
just
saw
some
failures.
B
A
To
see
it,
it's
just
it's
looks
like
it's
issue
with
the
like
destination,
where
the
locks
have
already
been
sent.
Oh,
I
see.
D
B
B
C
D
Okay
yeah
spock
exporter,
but
they
have
jagger
as
well.
Okay,
yeah
I
get
to
jugger
and
traces
traces
together
and
looks
to
splunk,
say
yeah,
also
showing
up
in
splunk
server,
failed
server,
transport
connection,
transport.
C
I
can't
troubleshoot
the
jaeger
receiver
or
exporter
that
fast.
I
don't
actually
know
what
it's
config
looks
like,
but
this
is
in
grpc
the
such
weird
things
in
kubernetes.
D
C
Does
the
does
the
jaeger
exporter
need
the
http
protocol
in
the
endpoint,
or
does
it
know
how
to
add
it.
D
Stuff
like
that
yeah,
let
him
do
more
work
instead
of
us
yeah
and
what
is
this
jagger
company?
I
guess
I
don't
know
question.
D
B
B
E
C
The
way
there's
there's
definitely
one
issue
that
we
we
absolutely
have
to
get
to,
and
it's
like,
a
very,
very
detailed
write-up
on
a
huge
architecture,
change
in
processors
that
I
read
through
and
then
was
like
wow.
This
is
like
a
big
deal
if,
like
this
issue,.
C
A
C
E
A
A
E
A
C
This
person
has
not
given
us
a
lot
of
help
here;
they
need
to
post
their
config
and
then
it
sounds
like
it's
a
question,
not
a
bug.
Yeah.
A
C
Someone
someone
someone
throws
something
out
on
the
helm
charts,
I
think,
or
maybe
it
was
in
the
collector
core
around
like
not
being
able
to
find
answers
on
stack
overflow
either,
which
is
true.
There's
like
nothing
on
open
telemetry
on
stack
overflow,
but
I
think
it's
funny.
C
D
D
D
C
It
looks
like
I
think
that
person
is
the
owner
of
the
of
the
processor
sounds
like
he's
taking
notice
already.
E
D
C
D
C
A
B
A
B
A
B
C
B
B
B
D
B
C
I
noticed
dimitri,
you
don't
have
hundreds
of
pushes
for
windows
tests
anymore.
Did
you
get
all
those
done.
A
I
mean
windows
tests
are
running
now,
they're
fixed.
Finally,
but
we
still
have
all
of
this
issue,
but
fixed
so
yeah
I
fixed
like
pretty
much
most
of
them.
I
believe,
but
those
are
just
too
much
to
dig
and
do
so.
I
just
created
issues
for
them.
They
are
sticking
here.
Unfortunately,
when
those
tests
are
still
the
slowest.
How
yeah
is
that
to
you
guys.
B
A
C
D
C
No,
it
isn't
core,
it
isn't
core.
Maybe
it
got
moved
before
I
swear
it
wasn't
contrib,
because
that's
where
I've
got
notifications,
for
I
think
this
is
worth
talking
about
at
the
end
of
this
meeting,
because
it's
which
ones
it's
called
proposal
for
pipeline
branching
in
the
open,
telemetry
collector.
C
C
B
Yeah
I'll
I'll
take
a
look
at
it.
Try
to
comment
on
it,
but
I
think
what
I'll
try
to
do
is
just
throw
them
link
them
to
my
pr
or
my
proposal
and
ask
them
to
explain
why
that
wouldn't
work
good
idea,
see
we'll
see
if
they
they.
E
B
C
But
yeah:
well,
they
put
a
lot
of
thought
into
this.
There's
different,
there's
different
they've
suggested
different
solutions,
one
that's
like
a
full
rewrite
one!
That's
like
use
the
routing
processor
as
like
a
hook
in
to
like
do
this
kind
of
thing.
Yeah
there's
some
thought
into
this.
C
D
E
D
E
C
C
I
my
team
was
talking
the
other
day
around
maturity
of
open,
telemetry
and
like
how
many
people
need
to
be
contributing
it
to
like.
It's
not
like
a
as
like
a
good,
a
good
change
rate
and
like
we're
getting
things
done
in
a
reasonable
amount
of
time,
and
we
were
comparing
dev
stats
between.
C
E
A
E
C
A
C
Still
I
got
one
more
dimitri,
I
couldn't
move
it
in
the
helm,
repo,
there's
a
issue
that
needs
to
get
moved
to
the
operator
repo
and
I
couldn't
transfer
to
the
operator.
I
could
only
transfer
it
to
contribute
or
release.