►
From YouTube: 2022-02-15 CNCF TAG Observability 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
A
B
C
Hey
we're
waiting
for
a
few
folks
to
join.
Today
we
have
a
presentation
by
hubble
and
perhaps
a
few
more
things.
D
C
You
know
I'll
just
I'll
just
open
the
meeting
formally
welcome
everyone.
This
is
a
meeting
of
the
cncf
tag
observability.
This
is
the
cncf
sponsored
meeting.
Please
don't
put
anything
in
the
chat
or
or
that
will
be
in
violation
of
the
code
of
conduct,
we're
waiting
on
thomas
graf,
who
I
understand,
we'll
be
giving
an
overview
of
psyllium
and
hubble
in
particular,
and
pinging
them
to
see
if
they
are
online
and
coming.
C
A
Hey
man,
hey
everyone,
while
we're
waiting
for
our
fearless
presenter
or
presenters
to.
C
A
Too,
because
I
actually
gave
a
presentation
internally
on
ebpf
last
week
and
was
talking
about
psyllium
and
pixi
and
stuff,
so
it
would
have
been
nice
to
see
it
in
action.
C
Yeah
I
had
also
lined
up
gibbs
colin
to
give
an
update
on
the
persona
work
that
they
had
been
doing
last
year.
Hi
richie.
B
C
Yeah,
I'm
not
so.
I
had
talked
with
liz
rice
and
thomas
graff
about
them
doing
the
slam
thing
and
they
are
not
showing
up
so
while
we're
waiting
is
there
anything
anyone
wants
to
talk
about.
Quite
literally
anything
is
fair
game.
There's
times
we
could
talk
about
whether
it's
new
work
place
where
folks
can
contribute,
but.
C
So
I
had
talked
to
liz
in
the
last
tlc
meeting.
You
know
I
had.
I
had
said
that
we
had
had
in
the
past.
Observability
projects
come
to
present
and
that
and
it
would
make
a
lot
of
sense
to
have
a
presentation
on
hubble
as
it's
an
incubating
project
and
whatnot,
and
so
I
had
assumed
well
not
assumed,
but
we
had
planned
and
put
on
the
agenda
for
the
meeting
today,
but
they
have
not
appeared
so
and
alolita
has
just
joined
us.
E
C
E
Hubble
to
be
presented
should
we
bring
him
on
slack.
Oh,
I
have.
C
Already
maybe.
C
Yeah,
maybe
maybe
something
came
up,
but
I
don't
have
much
more
on
the
engine
to
prep
for
today
other
than
that,
so.
A
No
worries
I
just
want
to
introduce
myself,
I'm
jackie.
I
actually
used
to
be
at
this
continuous
delivery
foundation.
I
was
their
director
of
ecosystem
and
community,
but
since
what
is
it
now,
probably
a
month,
I
transitioned
over
to
cubecost,
and
so
we
are
interested
in
getting
more
involved
in
the
community.
A
So
one
of
the
things
that
I
would
be
interested
in
figuring
out
is
also
like.
We
just
submitted
an
application
to
the
cncfs
sandbox
for
our
open
cost
model,
which
is
our
open
source
project.
So
I'd
love
to
be
able
to
come,
also
have
like
one
of
our
engineers
come
and
present
the
open
cost
model.
I
don't
I
I
don't
know
what
the
process
is,
so
I
just
thought
I,
since
we're
not
doing
anything.
I
think
sometimes
dumb
questions.
C
Thank
you
so
much
for
for
coming
and
welcome
the
process
is
pretty
open.
You
put
yourself
on
the
agenda
and
and
that's
it,
but
we
seem
to
have
some
time
that
apparently
has
freed
up
in
our
calendar.
For
today
you
are
welcome
to
chat
for
a
while
about
that.
If
you
like
or
come
back
yeah.
A
I
just
yeah,
like
I
said
I
I'm
pretty
new
to
the
company
so,
like
the
you
know,
fun
information,
you're,
probably
looking
for
I'm
not
the
right
person
to
to
give
it.
But
I'd
love
to
put
myself
in
the
agenda
then
and
then
just
bring
in,
like
I
said
one
of
our
engineers
to
to
do
a
presentation
on
an
open
cost
or
even
web.
Our
ceo.
So
I
know
there's
a
a
big
commitment
from
our
side
to
just
get
more
involved
in
in
the
kubernetes
community
and
cncf.
A
So
that's
why
I'm
attending,
because
there's
no
one
else
right
now
that
has
the
bandwidth
but
I'd
love
to
just
yeah,
get
more
folks
in.
C
Well,
that's
in
scope
for
the
tag
you
know
the
tag
is
not
just
for
cncf
projects.
The
tag
exists
to
in
part,
inform
the
toc
on
gaps
in
the
ecosystem
and
opportunities
open
source
projects
that
are
not
yet
part
of
the
of
the
cncf
as
well
as
established
projects.
So
you
know
if
it's
if
it's
about
the
observation
of
cloud
native
systems,
which
includes
observing
how
much
money
is
being
evaporated
by
the
minute,
that's
absolutely
within
scope.
A
C
For
sure
there
are
a
couple
pieces
of
adventist
trivia.
We
could
fill
a
little
bit
of
time
with,
but
they
are
fair.
Few
one
I
believe
at
the
beginning
of
may
richie,
maybe
straight
on
the
date,
but
we're
coming
up
on
two
years.
Richie
and
I
being
co-chairs
and
so
elections
will
nominations
and
elections
will
be
happening
in
the
com
in
the
coming
months.
C
C
C
Rather,
when
it
sounds
silly,
but
when
we
do
agree
on
a
logo
for
the
tag
that
unlocks
a
skill
tree
in
the
form
of
community
sites,
discussion
boards
and
some
some
some
group
management
that
the
cncf
has
for
groups
like
the
technical
advisor
groupon
observability.
So
it's
a
logo
but
there's
a
lot
behind
it,
and
so
that
will
be
next
meeting
in
two
weeks
as
well
for
the
first
tuesday
of
the
next
month,
a
lolita
or
richie.
Is
there
anything
you
wanted
to
chat
about?
C
I
I
feel
a
little
sheepish
that
we
have
an
empty
agenda,
but
it's.
B
So
maybe
for
jacqueline
the
process
to
get
this
on
is
write
it
into
an
agenda
like
into
a
fugit
future
agenda
and
and
then
you're
on
the
list.
So
it's
really
simple,
the
main
ask
is
bring
technical
people,
bring
technical
content.
B
D
B
A
B
The
the
other
thing
is,
if
you
have
a
timeline
of
of
intention,
like
you
mentioned,
submitting
to
to
toc
for
for
submitting
the
project
blah
blah
blah
like
if
there
is
any
timeline
or
if
any
of
those
steps
already
happened
or
anything.
That's
also
relevant
information
for
the
group.
Yeah.
C
I
I
had
a
question
too
right.
I
guess
it's
separate
from
that.
I
don't
know
if
there's
any
other
thoughts
on
that
firewire
have
we
talked
so
I've
missed
the
past
like
couple
and
I
know
a
while
back,
we
were
talking
about
like
the
like
working
group
like
observe
kubernetes
working
group
stuff,
and
I
was
wondering
if
there's
been
any,
I
guess
like
progress
on
that
or
if
we've
gotten
any
closer
to,
I
guess
kind
of
yeah
like
moving
forward
with
that.
C
Yes,
so
last
week
at
the
tlc
meeting,
I
kind
of
gave
a
brief
overview
of
that.
I
can
provide
a
link
to
the
youtube
and
whatnot
it's
it's
on
their
channel
to
give
them
a
heads
up
that
we
had
formed
this
working
group.
C
Our
plan,
our
record,
is
that
q1
of
this
year
is
sort
of
planning
and
preparation
and
design,
and
things
like
that
and
q2
and
q3
are
when
we're
hoping
to
hack
actually
do
implementation
with
some
contributors
on
board
by
acute
by
the
beginning
of
q2.
We
had
originally
slated
this
to
be.
You
know
two,
you
know
q1
and
q2
of
this
year,
where
we
did
most
of
the
implementation,
but
between
holidays,
coveted
for
some
of
us
and
our
families
and
some
other
things.
C
C
And
so
is
there
like
a
meeting
time
for
the
working
group?
Yet
there
is
not,
but
we,
like,
I
said,
the
chairs
abuelita
and
I
and
pastor
chair
meeting
on
thursday
to
kind
of
finalize
all
of
that
stuff
and
put
the
package
together
and
launch
it
formally,
but
it's
for
all
intents
and
purposes
it's
been
ken
ken
finnegan
myself,
daniel
kahn
and
michael
hassenblast.
We've
had
a
series
of
meetings
over
over
the
course
of
december
and
the
beginning
of
january.
C
If
anyone
is
interested,
please
reach
out
or
put
your
name
in
the
pr
or
just
in
some
other
way,
indicate
your
interest
and
also
on
the
mailing
list.
When
we
do
a
doodle
will
go
out
to
find
you
know
a
time
that
works
for
those
that
that
wish
to
be
involved,
cool
thanks,
yeah.
C
C
So
they
are
excited
to
come
and
give
us
a
more
in-depth
overview
of
what's
happened
in
that
project
over
the
last
year,
the
advances
they've
made
the
new
things
that
are
there
for
for
all
to
see
and
use.
So
that
is
also
on
deck.
C
They
were
at
our
last
meeting
and
spoke
briefly
around
some
of
the
challenges
of
using
pixi
at
scale,
not
pixie,
but
using
ebpf
at
scale.
They
had
informed
a
discussion
a
few
weeks
ago
when
bumblebee
came
to
talk
about
the
ebpf
work,
that
that
project,
which
is
nascent
and
new,
has
been
doing,
and
so
we
had
subsequent
conversations
and
they
have
asked
to
come
and
give
a
more
in-depth
overview
of
the
last
year
of
progress.
C
That's
about
all
I've
got
for
today,
I'm
going
to
paint
them
one
last
time.
I
don't
want
to
have
people
sitting
here
with
that
air,
but
does
anyone
else
want
to
talk
about
anything
today,
or
should
we
unfortunately
call
it.
E
Early,
I
think
again,
matt
and
everyone.
I
know
we
were
planning
this
session
out
today
going
forward.
I
think
we'll
get
a
bit
more
organized.
We
are,
as
matt
said,
we've
you
know,
kind
of
discussed
a
calendar
of
events
and,
and
you
know
again
different
groups
and
projects
presenting
richard.
I
I
think
it
would
be
nice
to
actually
have
an
update
from
the
prometheus
and
community.
Also,
perhaps
that
would
be
something
that
the
you
know
you
could
pull
in
one
of
the
engineers
to
present
on
and
I'll.
E
Definitely
you
know
get
a
open,
telemetry
update,
ready.
You
know
also
for
covering
you
know,
maybe
first
week
of
march,
but
we
can
plan
that
out
and
again.
I
think
we
have
done
pixie
before
the
idea
really
is
you
know?
How
can
we
actually
invite
some
of
the
our
customers
also
who
are
using?
E
You
know
a
lot
of
the
projects
that
we
talk
about
here
to
also
come
and
present.
But
again
I
think
that
moving
forward,
let's
be
a
bit
more
organized
it'll,
just
help
us
in
making
this
a
good,
productive
discussion
session.
D
And
this
board
suggests
I'll
be
up
to
the
right.
This
would
be
great
if
we
can
then
write
premieres
folks,
as
I
remember
they
run,
this
show
every
monday
right.
It
would
be
nice
to
know
on
what
station
they
are
right
now,
yeah.
E
D
E
I
mean
again,
you
know
richard
and
I
both
worked
on
the
prometheus
full
interoperability
of
prometheus
in
in
open
telemetry.
So
the
great
the
good
news
is
that
you
know
we
are
actually
fully
compatible
for
metrics
and
can
fully
interoperate
and,
as
you
know,
jager
and
zipkin
is
also
fully
supported
in
both
you
know,
with
both
pipelines,
but
specifically
hotel.
E
You
know
definitely
can
guarantees
an
end-to-end
trace,
tracing
support.
D
B
So,
first
of
all,
it
is
open
source
absolutely
same
as
same
as
everything
I
I'm
personally
biased.
Of
course
I
work
at
grafana
labs,
so
I
would
recuse
myself
from
any.
D
But
in
general,
let's
say
anything
that
would
use
primitives
format
where
we
can
meet
all
these
different
pillars.
So,
for
example,
if
traceability
can
use
primitives
format
is
a
lock
collection,
use
parameters,
format
and
parameters.
Naturally,
we
can
just
release
this
events
on
the
flight,
so
this
is
where
it
would,
but
it
would
be
better
to
to
see
how
and
open
telemetry
entries
would
go
there
and
convert
to
permit
this
format.
D
E
D
Tell
you
even
more
because
usually
we
sit
on
the
level
when
we
just
you
know
everyone
talk
about
these
three
pillars,
which
is
a
matrix
logs
and
traces,
but
nobody
paying
for
next
layer,
which
is
alerts
itself
and
and
big
view
for
observability,
and
this
is
what
actually
trying
to
do
by
using
this
open
source,
which
is
trying
to
bring
alerts
back.
D
C
Work
together,
so
I
want
to
make
sure
I'm
understanding
the
request
correctly.
Are
you
looking
to
have
someone
present
an
overview
of
sort
of
prometheus,
open
protocols
like
remote,
read,
remote,
right
and
open
metrics
and
things
like
alert
manager
and
other
configurations,
and
and
provide
some
examples
of
how
different
projects
within
the
cncf
can
be
brought
together?
To
achieve
a
consistent
platform?
D
D
You
cannot,
for
example,
easily
transfer
this
data
for
future
review
for
another
open
sources,
for
example
for
elasticsearch,
so
if
there
is
any
attempts
to
just
reformat
this
data
later
and
get
other
open
sources
that
do
not
use
primitives
format
to
use
this
data
example,
I
collect
primitive
for
90
days
on
one
minute
interval
and
later
elasticsearch
people
will
have
to
do
long-term
analysis.
They
can't
because
it's
a
different
format.
We
have
to
find
a
way
move
it
back.
D
There
is
some
solutions,
but
there
is
not
unified
solution
that
it's
the
usual
problem
with
any
with
any
stuff.
You
are
more
than
welcome
to
bring
your
data
and
it's
highly
difficult
to
expose
this
data
back
to
anyone
else.
So
that's
what
I'm
trying
to
say
if
we
can
discuss
how
parameters
can
help
to
bring
this
artifact
to
any
open
source.
B
We
can
absolutely
try
and
and
cover
these
kind
of
use
cases
within
within
the
tag
like
as
content
or
teaching
or
something.
I
do
believe
that,
in
particular
in
the
prometheus
and
also
10
point
loki
case
with
the
remote
right
protocols,
there
is
a
defined
way
to
get
everything
out
without
any
barrier,
because
that's
just
part
of
the
overall
design
to
to
enable
both
getting
in
and
out
and
open
telemetry
is
also
defining
protocols
like
the
otlp
switch,
which
seem
like
a
good
choice
for
for
a
lingua
franca
completely
detached
from
using
a
library.
B
We
have
always
tried
as
tech
to
get
more
end
users
involved
and
and
like
figure
out
what
what
actual
benefit
we
can
give
to
the
end
users
and
and
found
it
hard
to
find
out
what
that
might
be.
So
this
is
good
signal
that
we
can,
that
I
don't
know,
maybe
create
the
list,
have
links
to
things
or
or
republish
existing
content
or
something.
C
There's
also
an
opportunity
here,
I
think,
to
kind
of
I
don't
know
if
you
were
boris.
I
know
if
you
had
attended
last
meeting
but
catherine
paganini
from
the
business
values
subcommittee
and
the
glossary
project
team.
C
There's
another
effort
within
the
business
value
subcommittee
to
generate
sort
of
a
matrix
subsection
by
subjection
of
the
landscape
and
provide
some
context
in
higher
level
summaries
of
the
different
projects
that
projects
can
put
in
and
and
then
use
that
as
a
springboard
to
kind
of
produce
additional
content,
be
it
short
videos
on
youtube
or
blogs
or
presentations.
C
In
the
tag
meeting
or
some
combination.
I
will
say
that
you
know
the
cncf
through
the
use
of
protocols
like
open,
metrics,
remote,
read
remote
right,
prometheus.
C
You
know
things
like
that
has
fostered
an
ecosystem
of
vendors
and
projects
that
all
address
sort
of
a
different
ways
to
to
solve
some
of
the
challenges
you're
talking
about,
and
so
you
know
we
do
need
to
be
careful
that
in
the
tag
there
you
know
we
can't
say
like
this.
Is
the
one
unified
way
to
do
it?
C
It's
this
and
this,
and
this
and
this
right,
what
we
can
say
is
we
have
open
protocols
and
open
formats
that
allow
end
users
to
interchange,
different
pieces
out
and
let
those
pieces
and
those
vendors
compete
on
the
merits
of
their
solution,
which
are
indexed
for
different
use
cases
and
have
different
trade-offs,
and
so
I
I
think
we
can
absolutely
suggest,
but
we,
but
we
do
just.
C
I
just
need
to
to
say
at
a
high
level
that
you
know
we
are
not
king
makers
here
in
the
tag
and
we
do
need
to
have
you
know
we
can't.
We
can't
declare
one
as
the
way
as
as
you
were
suggesting,
but
we
can
say
this
is
an
example
of
it
and
in
fact
the
project
that
we
talked
about
earlier
really
is
a
good
vehicle
for
this,
because
it's
a
it's
a
curated
end-user
driven
catalog
or
it
hopes
to
be
where
the
exact
sac
you
talked
about.
C
You
know
tempo
low-key,
cortex,
I'm
sorry
tempo
loki
prometheus.
I
believe
that's
an
example
of
putting
legos
together
in
a
way
that
fits
a
use
case,
but
there's
a
bunch
of
others
too.
D
D
C
I
will
certainly
do
that.
It's
on
it's
it's!
The
video
has
been
posted
for
a
while,
but
later
on
today.
Actually,
I'm
posting
something
to
linkedin
and
then
to
slack
with
a
summary
of
the
last
meeting
and
a
link
to
the
video
I
need
to
search
like
last
and
then
what
we
just
did
and
then
a
preview
to
what's
to
come.
So
we
want
to
get
more
to
a
cadenced
outward
messaging.
C
C
The
last
thing
I'll
say
is
last
year,
I
think.
Actually,
ironically,
when
pixie
presented,
someone
came
and
demonstrated
a
tool
that
is
specifically
for
bulk
grade
bulk
right
of
historical
data
from
prometheus
in
very
large
chunks.
There
was
that
there
was
a
tool.
I
can't
remember
the
name
of
it.
In
addition,
bartek
our
tech
lead
or
one
of
our
tech
leads
also
has
a
different
tool.
I
think
that
he
was
working
on
last
year.
I
don't
know
if
it's
done,
but
we
can.
C
We
can
ask
him
that
uses
grpc
to
do
the
grpc
approach
that
uses
bits
of
prometheus
and
other
things
to
actually
both
move
things.
In
the
same
way,
however,
I
don't
speak
for
the
previous
community.
I
think
richie
richie
is
an
authority
there.
So
he'll
have
to
keep
doing
this.
B
I
think
bartik
has
his
tool
in
the
thanos
repository
for
those
who
don't
know.
Thanos
is
basically
part
of
the
prometheus
family,
so
you
have
prometheus
as
the
single
binary
single
instance
blah
blah
blah
through
federation
and
that
you
have
cortex
and
thanos
as
a
horizontally
scalable
implementations
which
reuse
a
ton
of
prometheus
code,
and
there
is
also
huge
overlap
between
the
maintainers
of
the
three
projects
and
I
think
that
tool
lives
there.
D
Are
questionable
because,
yes,
there
is
a
was
a
case
for
primitives
that
would
dump
this
data
analytical
format
for
any
other
tools
and
protocol,
but
I
couldn't
find
things
like
this
on
cortex
side,
besides,
if
you're
involved
for
reading
right
and
cortex
you're
going
to
pay
resources
price
to
do
this
frozen
by
query,
creating
this
stuff.
So
that's
a
question.
So
if
we're
on
already
push
data
to
promote
this
format,
it
is
possible
to
bring
to
another
format
of
vice
versa.
D
B
For
I
mean
everything
which
you
get
out
of
also,
we
are
very
deep
into
the
specifics
of
prometheus.
So
if
just
on
a
note,
but
anything
you
get
out
of
prometheus
through
the
web
interface
or
as
an
like
any
query,
you
run
against
prometheus
is
coming
back
as
json,
so
you
already
have
that
for
most
other
things
like
datadog,
what
have
you
they?
They
proclaim
compatibility
on
prometheus
remote
right,
so
you
should
be
able
to
move
the
state
at
least
towards
those
how
to
move
it
out.
B
I
don't
know
with
my
prometheus
head
on
during
kubecon.
We
expect
to
launch
the
prometheus
conformance
program
for
real.
We
all
we
already
did
a
pre-launch
and
and
published
some
initial
results
last
q
con,
but
in
may
we
expect
to
actually
have
something
like
with
a
logo
and
contracts
in
place
and
everything
where
you
can
rely
on.
Anyone
who
claims
this
actually
supports
it.
C
Yeah,
boris,
please,
if
you,
if
you
would
put
your
email
either
just
like
me
or
in
the
in
the
meeting
notes.
C
It's
really
okay.
The
other
thing
I
was
going
to
say
is
that
one
thing
that
I
did
with
our
our
prometheus
and
metrics
collection,
stuff
is
remote
right
and
particular
can
have
multiple
endpoints
right,
so
you
can
send
you
know
your
metrics
wherever
they're
going
and
then
in
our
case
we
had
like
a
data
science
and
machine
learning
analytics
team
that
wanted
to
do
some
historical
analysis,
so
we
send
things
to
there.
C
Also,
so
you
know
you
can
do
a
one-to-many
fan
out
of
of
remote
right
for
what
it's
worth
that
doesn't
cover
your
historical
use
case
necessarily
directly,
but
in
terms
of
data
from
now
moving
forward.
C
We
have
very
large
long,
bulk
reads,
for
example,
and
things
like
that
we
wanted
to
decouple
those
two
scenarios
to
two
different
back
ends,
but
because
remote
right
is
so
flexible,
we
were
able
to
do
so
without
much
drama
and
without
impact
to
production
workloads
which
would
just
keep
on
doing
what
they
were
doing,
sending
it
to
whatever
I
can
get
sent.
So
I'm
happy
to
hire.
I'm
sure
any
other
prometheus
team
could
could
point
you
in
the
right
direction.
It
was
just
a
simple
configuration:
okay,
okay,
no,
no
you're,
absolutely
right!
C
This
is
a
way
to
do
this.
Just
as
interesting.
This
fork
solution
sent
to
different
and
also
let
us
have
multiple
back
ends.
You
know
one
for
short
duration,
short
short.
C
Not
keeping
a
lot
of
historical
data,
but
doing
so
in
a
way,
that's
best
to
query
for
dashboards,
but
then
sort
of
like
for
the
historical
data,
something
that's
a
little
less
expensive
to
run.
That's
queried
intermittently
on
weekends,
with
batches
and
things
like
that.
So
so
all
of
these
sorts
of
use
cases
are
all
in
scope
and
are
supported
by.
D
D
Your
analytical
team,
coming
with
different
tools,
which
is
not
part
of
primitives,
for
example,
and
they
told
me,
can
we
use
it?
You
say
absolutely.
This
is
03..
They
try
to
do
this.
They
can't
because
it's
not
the
json.
This
is
primitives.
This
is
just
one
example,
so
it
should
be
possible
to
bring
to
primitives
format
and
it
should
be
possible
to
transfer
back
as
a
any
other
formats.
The
same
piece
of
that
that
is
not
here.
As
you
mentioned,
there
is
a
way
to
do
this.
D
C
Sorry,
but
I
took
the
too
much
of
your
time,
so
no
I
we
we
had
unexpectedly
lots
of
it.
So
does
anyone
else
have
anything
they
want
to
raise,
or
should
we
return
time
to
folks.
C
Okay,
okay.
Well
again,
I
apologize
that
we
didn't
have
the
presentation
that
we
advertised
we
will.
We
will
do
that
as
soon
as
we
can
when
we
figure
out
what's
up,
but
if
that's
all,
I
wish
everyone
a
good
week
and
I'll
see
you
on
the
first.
Hopefully
thank
you
see.