►
From YouTube: 2022-04-21 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
Not
sure
if
we've
ever
actually
met
too
bad
yeah,
I
don't
think
we
have
but
yeah
p.m.
Honeycomb,
doing
open,
telemetry
stuff,
nice
yeah
just
trying
to
like
contribute
upstream
wherever
I
can
influence
more
of
that
like
with
like
hey,
we
should
hire
people
who
like
contribute
upstream
and
and
then
like
trying
to
like
spread
it
more
around
the
company.
Like
hey
we're,
taking
a
bet
on
this
thing.
B
A
B
Coming
to
hotel
meetings,
oh
sometimes
so,
like
I
used
to
work
for
microsoft,
I
was
on
the
dotnet
team
and
worked
in
our
like
compilers
languages
group,
and
so
we
didn't
have
any
devrel
people.
B
Well,
we
kind
of
did,
but
that
was
a
weird
time
at
microsoft,
where
they
they
had
just
absolutely
ridiculous
amounts
of
money
they
threw
at
people,
gave
them
like
no
agenda
and
also
gave
them
no
agency
on
product
direction
and
they're
like
hey,
go,
be
deborah's
and
they're
like,
and
so
we
were
frustrated,
but
so
we
did
our
own
devrel
as
a
consequence
cool.
So
I
do
some
of
that.
Try
to
evangelize
the
hotel
to
other
people
like
at
some
of
the
honeycomb
stuff,
be
like
hey
guess
what
you
know.
A
I
used
to
to
wear
a
lot
of
hats
at
light
step,
but
running
kind
of
keeping
the
open
telemetry
project
going
seemed
to
be
like
a
unique
value
proposition
that
I
had
the
ability
to
do
and
was
interested
in,
and
so
I
kind
of
like
put
down
like
product
management
and
some
other
things,
and
I
help
helping
austin
parker
kind
of
grow,
a
devrel
team
over
here
right
now,
because
that
seemed
like
the
the
most
natural
home
for
doing
this
work
in
an
organization
that
doesn't
have
like
an
explicit
like
open
source.
A
You
know
engineering
effort
or
something
like
that,
so
so
yeah,
that's
kind
of
what
I'm
up
to
these
days
and
but
yeah
like
one
of
the
places.
I
I'm
sure
you
know
from
prior
prior
meetings
here,
one
of
the
places
we
we
really
need.
A
lot
of
help
is
like
getting
the
the
docs
and
the
website
like
into
a
good
spot,
and
then
I
feel
like
the
next
step
after
that
is
an
engineering
effort
to
improve
the
install
experience.
Like
I
think,
step
one.
A
The
thing
we
need
to
do
most
urgently
is
is
document
the
current
installation
process
better
for
people
in
all
the
languages
and
making
sure
like
getting
started
is
is
reasonable,
but
then,
like
the
step
back
after
that
is
like
how
do
we
actually
like
clean
that
up
into
a
better
process
which
involves
like
some
like
spec
work,
like
I'm
sure,
if
you're
familiar
with
open
telemetry,
you
know
how
it's
like
currently
just
like
it's
well
factored.
A
But
you
know
you're
not
required
to
install
open
telemetry
as
this
like
mass
unit,
and
you
can.
You
can
bring
your
own
implementations
and
there's
a
lot
of
like
flexibility
that
I
think
is
important.
If
you're
trying
to
make
something
like
a
standard,
you
got
to
give
people
all
of
these
escape
hatches,
but
what
we
haven't
defined
that
is,
haven't
defined,
hey
austin.
A
What
we
we
haven't
defined
is
like,
like
a
clean,
wrapper
around
all
of
that
stuff.
So
if
you
are
just
doing
the
normal
thing
of
just
like
starting
the
sdk
with
like
tracing
metrics
and
logs,
like
all
enabled
along
with
like
the
otlp
exporter
I'll
point
it
at,
like,
you
know
a
default,
you
know
local
port,
and
you
know
where
you
expect
a
collector
to
be
running
like
that.
Like
default,
experience
hasn't
been
like
defined
or
added
to
the
project,
and
we
haven't
defined
a
configuration
file
format
for
the
sdks
to
drive.
A
All
of
that.
So
I
actually
see
that
as
like
a
one-two
step
and
that's
like
figuring
out
what
the
definition
of
that
I
mean.
That's
not
strictly
like
the
comms
group
job
to
do
that.
That's
like
a
specification
role
but,
like
I
feel
like
we
are
the
group
of
people
who
will
probably
have
good
ideas
about
what
that
should
actually
look
like.
Well,
I.
B
C
B
Like
I
mean
we
like,
for
example,
in
the
the
pollinator
slack
at
honeycomb
there's,
you
know,
discuss
open
telemetry
thing,
that's
usually
pretty
active
and,
like
a
lot
of
the
same
stuff
keeps
coming
up
there
and-
and
it's
it's
like
on,
you
know,
hey.
I
think
I
can
get
something
going,
but
it's
kind
of
complicated
yeah
or,
like
you
know
I
I
have
in
my
head.
B
I
know
what
I
want
to
do,
but,
like
the
api
doesn't
really
tell
me,
like
you
know,
which
direction
I
should
take
right
now
and
by
the
way
this
is
my
first
time
tracing
so
like
I'm,
not
even
really
familiar
with
this
concept.
To
begin
with,
yeah
or
or
like
you
know,
oh,
when
should
I
use
the
like,
like
at
what
stage
in
my
telemetry
processing
pipeline
like
I
know,
I
want
to
drop
health
checks
like
do
I
do
that
in
the
code?
Do
I
do
that
in
the
collector?
B
How
do
I
go
about?
Do
that
like
there's
totally
and
then
like
enterprisy
kind
of
concerns,
that
kind
of
pop
up
as
well.
D
So
yeah
some
color
on
that,
because
I
was
just
talking
to
some
enterprise
people
using
hotel
and
their
question.
I
think
a
lot
of
it
is
there's
just
not
enough
documentation
about
like
how
this
fits
into
the
stuff
they're
doing
so
like,
for
example,
they're
using
spring
an
older
version
of
spring,
but
they
have
like
other
custom
middlewares
right
and
they
have
like
weird
proxies
between
services
and
like
there's,
this
kind
of
whole
class
of
it's
under
it.
D
Concepts,
and
in
like
you,
know,
middleware
injection
and
also
like
to
be
honest.
I'm
not
I'm
not
an
expert
in
anything,
but
I
do
look
at
what
like
the
code,
that
is
there
and
a
lot
of
it
is
like
engineered
in
such
a
way
and
the
architect
in
such
a
way.
That
is
extremely
difficult
for
me
to
follow,
without
having
a
lot
of
like
domain
knowledge
about
not
just
java
or
not
just
tracing
but
like
how
the
java
sig
has
structured
their
artifacts
and
their
like
code
paths
right
and
what
libraries
are
where
and.
A
B
Oh
yeah,
like,
if
you
want
to
like
look
at
how
do
I
start
a
span
in
javascript,
you
don't
go
to
the
hotel,
js
repo.
You
go
to
the
hotel,
js
api,
repo
yeah.
A
D
A
A
Whole
java
agent
thing
that
you
just
don't
have
in
most
other
languages,
that's
really
valuable
in
js.
They
currently
need
to
share
as
much
of
an
implementation
across
two
very
different
environments,
like
the
browser
and
and
node
with
like
some
like
key
core
components
being
like
working
radically
different
like
context
propagation
yeah
like
it's
and
you
know,
golang
has
like
different
rules
around
what
constitutes
breaking
change
from
other
languages
that
that
further
kind
of
emphasize
how
you
break
the
packages
down
there
and
yeah
yeah.
C
A
But
but
that's
all
like
yeah,
that's
where,
like
like
a
documentation
effort,
I
think,
can
really
help
focus
on
installing
and
setting
all
of
this
stuff
up.
In
addition
to
like,
like
there's
like
the
basics
of
like
how
do
I
use
the
api
like
creative
span
or
whatever,
but
that
actually
like
doesn't
quite
get
you
to
where
you
need
to
be,
if
you're
gonna
like
instrument
your
code
right
like
understanding
how
to
instantiate
a
tracer
and
like
what
to
name
all
that
and
like
what?
What
should
you
wrap
in
a
span
versus
not
and
yeah,.
B
B
It
tells
me
something
is
wrong,
but
it
tells
me
it
doesn't
tell
me
anything
else
is
like
not
sufficient
for
my
needs,
so
I'm
bought
into
that
and
I'm
bought
into
like
tracing
is
like
one
way
to
like
help
get
out
of
that
that
that
world,
but
I
don't
even
really
know
like
they're
like
I
can
install
some
automatic
stuff
but
like
if
I
see
that
a
given
service
is
like
slow.
Where
do
I
put
my
first
span,
yeah.
A
B
Do
I
think
about
that?
Generally
speaking,
like
I
know,
it's
going
to
be
unique
to
my
code
base
but
like
if
I'm
not
even
necessarily
familiar
with
my
code
base,
let
alone
like.
Oh,
what
are
so
like.
If
I
compare
that
with
like
there's
a
great
debugging
technique
and
ides
that
you
can
use
where
you're
like
okay
well,
I
know
that
the
code
is
going
to
flow
from
this
point
in
the
code.
B
So,
like
this
point
of
the
code,
I'm
going
to
set
breakpoints
there,
I'm
going
to
set
one
in
the
middle
and
I'm
just
going
to
let
it
execute
and
stop
at
each
one
and
look
at
the
state
at
each
point
and
then
binary
search
my
way
through
there
and
like
that's
like
a
fairly
like
you,
don't
need
to
know
somebody
else's
code
base
to
like
teach
them
that
debugging
technique.
Is
there
an
equivalent
technique
for
that
with
instrumentation?
B
D
I
think
the
challenge
is
that
conceptually
that's
very
easy
to
explain
to
someone
that
hey
you
have,
you
should
add
a
span
around
the
entire
request
and
then
you
should
add
a
span
around
like
it's
like
you
start
at
the
cert,
like
you
start
the
service
level
right
and
then
you
go
down
and
down
and
down
and
down
and
down
you're
pressing
the
search
space.
Every
new
span
you
add
the
problem
is:
is
that
if
it
was
as
simple
as
like
writing
one
line
of
code
that
said
span
dot
here
right,
easy
yeah.
A
D
Yeah
we
need
someone
to
like
chinese
firewall,
a
implementation
of
dev
mode
in
the
collector
yeah,
and
that
would
solve
like
a
lot
of
these
problems,
because
then
we
could
just
tell
people
hey
you
set
this
up.
You
run
you're
pointing
things
at
the
collector.
You
turn
on
dev
mode
on
the
collector
and
you
go
to
this
page,
and
it
just
shows
you
a
stream
of
all
the
spans
that
are
being
printed,
and
it
shows
you
like
where
the
context
is
breaking
right.
A
D
It
also
sent
metadata
about
the
live
site
like
this
was
using
our
tracer
implementation,
so
we
could
hack
other
stuff
in
there,
and
one
of
the
cool
things
was
is
that
we
basically
added
in
meta
spay.
It
likes
meta
spans
that
corresponded
to
like
life
cycle
events
for
the
tracer
itself.
So
when
you
start.
D
D
Sorry,
by
the
way,
I'm
really
sorry
for
being
late,
I'm
in
san
francisco,
I
was
here
to
do
a
training
and
then
an
unfortunate
thing
happened,
and
so
now
I
am
trapped
in
the
light
step
offices
for
the
next
12
hours.
A
But
yeah
so
some
something
like
like
that,
like
basically
debugging
dev
tools
would
be
like,
and
this
is
a
request
we
get
a
lot
like
at
the
last
hotel
community
day.
People
usually
phrase
it
in
terms
of
like
I
want
to
see.
What's
in
the
collector,
basically
is
like
one
way
of
thinking
about
it.
It's
just
like.
Can
we
build
a
ui
that
attaches
to
the
collector?
C
So
to
me
this
sounds
all
like
observability
for
observability,
exactly
yes,
and
that
that's
something
I
to
to
to
add
to
what
you
said
in
the
beginning,
documentation
getting
started
having
something
like
this
live
preview.
I
think
there
is
some
something
missing
in
all
of
a
lot
of
open.
Telemetry
is
all
this
yeah?
How
can
you
troubleshoot
your
tool?
You
want
to
use
for
troubleshooting
and
that's
yeah.
Open
telemetry
is
not
yet
to
put
it
that
way.
Setting
a
prime
example.
C
I
shared
a
reddit
post
somewhere,
where
someone
said
like
yeah,
installing
open,
telemetries,
quite
difficult,
yeah
yeah.
I
think
they
used
to.
C
For
that
so,
and
I
think.
C
Yeah
and
and
then
that's
what
what
I
talk
a
lot
about
also
here
at
appd
and
cisco
in
general,
with
people
like
hey,
if
you
want
to
help
the
open
telemetry
community
and
go
there
and
make
things
easier
and
and
really
see
like
hey,
because
I
think
that's
at
the
end
how
people
are
going
to
adopt
something.
Why
should
someone
go
from
a
vendor-specific
agent,
for
I
don't
know
java
javascript
whatever
to
open
telemetry,
if
it's
just
so
much
more
complicated
to
install
and
maintain.
A
C
A
Yeah
and
the
reality
is,
it
actually
doesn't
need
to
be
more
complicated
to
install
there's
the
extra
burden
of
a
tracing
system,
regardless
of
where
you
get
it
from,
is
going
to
have
more
trickiness
with
installing,
because
it's
a
cross-cutting
concern.
So
just
you
know,
there's
this
extra
burden
of
understanding
context.
A
Propagation,
that's,
I
think,
but
I
think
that's
like
a
temporary
issue
in
the
sense
that
open
telemetry,
just
in
general,
the
concept
of
distributed
tracing
is
becoming
more
popular,
and
so
I
think,
as
time
goes
on,
this
will
become
more
of
a
a
normal
concept
for
developers.
This
concept
of
like
there's
a
context.
That's
like
you,
have
access
to,
and
you
know
spans
and
like
traces
and
how
that
works.
Half
the
problem.
C
D
But
I
want
to
point
out
like
that,
because
this
is
something
we've
talked
about
before,
just
to
echo
it
like,
when
you
get
down
to
it
to
brass
tacks.
The
first
step
in
using
up
the
telemetry
is
actually
using
context
and
understanding
context,
and
when
you
lead
off
with
someone
by
saying
okay,
you
want
to
yeah.
It
feels
very
much
like
to
bacon
apple
pie.
You
must
first
invent
the
universe.
A
I
mean
the
the
first
step
for
people
and
what
should
hopefully
be
sufficient.
Is
you
install
the
sdk
and
you
install
all
the
available
instrumentation
that
matches
your
system
and,
if
your
sys,
if
your
app
is
built
out
of
stock
parts,
then
that
should
be
enough
right,
like
which
is
the
equivalent
of
like
the
apm
vendor
solution
right
now,
it's
kind
of
the
same
deal
there,
which
is
like
it's
kind
of
based
on
the
idea
that
your
system
is
built
out
of
enough
stock
components
that
that
we
will.
D
A
D
C
D
See
my
traces
in
light
stuff
for
honeycomb
or
fd
or
wherever
there's
nothing
that
actually
tells
you
what
any
of
it
means
like.
There
is
a
modeling,
slash,
comprehension
gap
as
well,
where
that
I
think
can
be
solved
documentation,
but
we
don't
do
a
great
job
of
taking
someone
through
like
the
gilded
path
of
from
nothing.
Here's
what
you
need
to
get
in
here's,
what
it
mean!
Here's,
what
your
first
span
or
your
first
trace,
actually
means,
and
now
we
can
start
to
go
into
more
advanced
stuff
like
yeah.
I
feel
like.
D
A
I
do
think
tools
is
kind
of
the
next
step,
because,
if
shit's
not
working,
then
your
first
step
is
going
to
be
where
in
this
like
pile
of
stuff.
Is
it
not
working
and
that's
also
the
place
where,
as
like
the
open,
telemetry
community,
we
could
inject
some
like
metadata
comprehension
around
what
this
data
is
supposed
to
be
it
with
without
devtools,
then
that
we're
providing
as
a
community
that
it's
basically
on
the
vendor
backend
system
to
provide
all
of
that
comprehension
which
they
should.
A
But,
like
you
know,
that's
all
gonna
have
to
go
through
the
extra
lens
of
like
how
that
back-end
system
wants
to
see
the
world
which
is
not
necessarily
gonna,
be
like
open,
telemetry
specific.
Some
of
that
is
still.
C
B
I
I
know
that,
like
summer
of
last
year,
one
of
our
main
big
accounts
was
like
throwing
a
major
fit
over
like
the
collector
causing
problems
and
like
the
collector,
didn't
even
report
at
the
time
like
what
version
it
was
in
like
the
header
and
we're
like
okay.
Well,
we
literally
cannot
help
you
unless
we
just
like
parachute
an
engineer
into
your
organization
and
like
work
with
you
directly
and
so
like.
B
We
contributed
a
thing
to
like,
add
that,
but
I'm
sure
there's
lots
of
other
things
that
could
be
added
where,
like
even
just
like
base
level
metadata
about
components
that
are
being
used,
could
then
just
be
surfaced
in
all
the
other
visualization
tools
so
like
if
you're,
at
least
already
using
you
know
one.
B
You
know
you're
either
you're
already
using
datadog
lightstep
honeycomb
all
that
stuff
and
if
you're
at
least
somewhat
familiar
with
how
you
get
that
kind
of
info,
you
could
actually
get
it,
whereas
I
think
like
there's,
I
think,
ironically,
a
lot
yeah
hotel
components
are
not
very
observable.
Today.
Yeah.
A
And
we
do
actually
report
a
lot
of
data,
but
again
it's
this
like
self-fulfilling
cycle
of,
like
you
know
like
every
tracer,
like
has
like
a
tracer
name
that
describes.
You
know
the
the
package
that
you
installed
to
get
that
instrumentation
and
a
version
number
and
hopefully
in
most
back
ends.
You
could
have
access
to
that
information.
A
Yeah,
it's
it's
it's
this
gap
between,
I
mean
it's
all
the
problem.
All
auto
magical
systems
happen
have
which
is
like
when
the
magic
works
then
great,
like
you're,
fine,
and
you
don't
need
to
understand
it,
but
like
the
moment
the
magic
doesn't
work.
A
If
you
then
have
to
drop
all
the
way
down
to
like
the
seventh
layer
of
hell
and
understand,
literally
like
every
detail
about
this
stack
of
stuff
that
the
the
magic
was
was
rapping
over,
then
that's
like
really
rough
yeah,
and
so
like
the
next
step
in
documentation
after
installation
is
like
yeah.
How
can
you
describe
what
layer
of
stuff
do
we
need
to
be
able
to
like
describe
to
people
so
when,
when
the
magic
just
doesn't
work
like?
How
could
we
like
guide
them
through?
A
It's
like
well,
okay,
here
are
like
the
basics
that
you
need
to
know
now
to
maybe
get
yourself
debugged
right
like
these
are
instrumentation
packages.
You
can
look
at
them
here
like
spans
when
your
context,
propagation
context,
propagation
primer,
when
it
breaks
you'll,
see
like
these
broken
things
that
look
like
this.
A
C
D
We
all
know
we're,
we
all
know
how
to
find
each
other
yeah.
I
think
this
would
be
a
good
conversation
to
continue
having
in
a
way
that
we
can
solicit
like
public
or
at
least
other
offended
people
with
an
affinity
for
this
right
and
try
to
drive
something
yeah
in
the
meantime.
You
know,
I
think,
let's.
A
Do
we
got
a
person
starting?
I
don't
know
if
austin
mentioned
this
before,
but
we
have
a
new
person
starting
and
we're
going
to
be
contributing
that
the
docs
that
we
have
such
as
they
exist,
the
information
we
have
written
down
in
various
places
for,
like
light
step,
customers
we're
going
to
just
try
to
port
all
of
that
stuff
over.
A
B
Yeah
we
actually,
we,
we
hired
a
technical
docs
contractor
to
like
help
fix
up
our
docs
in
general,
but
then
one
of
the
sub
projects
they
had
was
just
like
write
docs
about
like
language
agnostic
concepts
for
hotel.
Just.
B
B
In
the
like
the
the
picture,
the
problem
is
that
that
was
turns
out.
This
is
a
really
hard
topic
for
a
technical
writer
who's,
not
already
you
know,
yeah
like
we
can
contribute
them
today,
and
I
think
we'll
probably
do
that
pretty
soon
now
that
a
death
march
of
a
project
we
were
on
is
now
done,
but
that
also
required
a
major
docs
revamp
but
like
they're,
probably
gonna
need
more
work
yeah.
B
D
B
A
Bring
this
time
yeah,
it
sounds
good,
but
yeah,
I
think
step.
One
is
just
any
information
that
you
are
finding
that
you
think
is
useful,
like
just
starting
to
get
it
all
pasted
into
google
docs
or
somewhere,
where
we
could
then
shovel
it
into
the
website
and
or
like
like
hand
it
over
to
a
technical
writer
where
it's
like
here
is
like
the
information.
Can
you
use
your
technical
writing
skills
to
to
like
clean
up
how
we
are
presenting
this?
A
B
A
B
C
B
I'll
just
like
put
that
stuff
on
there,
so
that,
like
there's
a
link
where
you
can
like
see
it
and
potentially
review
it,
nice.