►
From YouTube: CNCF SIG Observability 2020-09-15
Description
CNCF SIG Observability 2020-09-15
A
C
D
C
C
C
Oh,
you
know
at
the
top
of
the
agenda.
We
should
make
a
an
announcement
about
our
technical
lead.
Now
that
that's
been
formalized.
B
B
B
Yes,
we
have
a
special
guest.
Today
we
have
cheryl
here
vp
ecosystem
of
cncf,
also
responsible
for
the
end
user
surveys.
As
most
of
you
probably
know,
there
was
an
end
user
observer,
end
user
survey
on
observability,
that's
a
tongue
breaker
quite
recently,.
B
And
as
a
result
of
this
question,
I
was
kind
of
wondering
why
is
stats
d
rated
as
highly
as
it
is,
and
it
feels
kind
of
unfair
picking
out
one
specific
project
but
as
prometheus
literally
started
in
part
to
to
replace
stetsd
everywhere,
it
feels
kind
of
fitting
to
code
it
out
and
to
try
and
get
a
better
connection
to
end
users
and
their
actual
needs
like
it's.
It's
fine
that
we
as
as
project
maintainers
and
sig
members
and
such
spearhead,
the
way
for
for
where
technology
can
go.
B
But
it's
also
very
much
about
enabling
end
users
to
actually
take
that
and
use
it
and
obviously
like
there
is
this
whole
discussion
about
many
people
see
see
observability
as
a
coast
center,
so
not
as
something
which
drives
innovation,
but
just
something
which
you
need
to
have
like.
I
don't
know,
toilets
or
or
elevators
like
it's
infrastructure,
it's
supposed
to
come
out
of
the
wall,
and
else
no
one
cares
and
we
can't
change
those
settings
at
least
not
immediately.
B
Yet
there
is
probably
some
some
space
into
which
we
can
into
which
we
can
try
and
move.
So,
for
example,
what
what
sure
and
we
talked
about
was
maybe
having
a
survey
sent
out
to
the
end
users
about
their
main
pain
points.
B
What
could
we
do
to
support
them
like
one
example
would
be
applied
from
here
of
and
then
just
give
them
like
five
or
ten
different
things,
maybe
more
five
things
side
which
the
end
users
can
choose,
what
what
gives
them
the
most
benefit
to
what
eases
the
most
pain,
deliberately,
not
free
from
for
the
users,
because
we
kind
of
need
to
spear
it
where
it
makes
sense
to
go,
but
then
basically
have
them
be
able
to
choose
what
they
want
and
that's
why
cheryl
is
here
to
talk
about
this.
D
I
thought
I
was
just
here
to
listen:
oh
okay,
but
yeah,
but
I'll
just
say
two
minutes
on
it:
the
reason
that
we
created
the
end
user
technology
radar
that
I
just
put
into
chat.
So
you
can
read
it.
I
guess
there
were
two
reasons
really
one
was
to
show:
okay,
what
do
end
users
really
use,
and
what
do
they
recommend?
D
D
G
The
result
hi
everyone
hi,
my
name-
is
I'm
using
the
opportunity
also
to
introduce
myself
I'm
from
I
work
at
logs
io,
so
that's
my
relation
to
the
domain
and
to
observability
and
specifically
to,
of
course,
for
cncf
related
projects
and
so
on.
On
that
respect,
I
think
that
this
is
something
that
we've
I
personally
so
have
experienced.
I
think
all
of
us
throughout
logs
io
this
you
know
maybe
lack
of
communication,
not
enough
communication
amongst
amongst
the
different
projects
in
the
the
different
sigs
and
the
different.
G
Let's
say
the
different
activities
and
different
pillars,
as
I
sometimes
call
them
so
definitely
welcome.
We
need
to
see
what's
the
effective
way
of
establishing
to
it.
So
this
is
sounds
like
a
a
good
start.
I
would
maybe
create
some
follow-ups
on
that
cheryl
if,
if
I
may
suggest
to
see
maybe
practical
means
of
creating
an
ongoing
facilitation
of
this,
these
interlinks,
but
definitely
on
the
one
in
the
industry
does
converge.
G
We
specifically
at
logs
io,
put
a
lot
of
effort,
putting
a
lot
of
effort
on
creating
these
bridges
between
the
disparate
project.
You
know
grafana
and
kibana
and
jager
and
prometheus
and
and
and
so
on
and
so
forth.
G
But
then
again
we
often
don't
find
the
counterparts
for
that,
and
we
definitely
would
like
to
see
that
as
a
community
activity
community
engagement
to
to
have
that
as
part
of
the
ongoing
discussion
and
maybe
also
standardize
the
way
that
these
interact,
like
classic
examples,
would
be
correlations.
You
know
log
trace
correlation
metric
to
logs
and
so
on.
These
are
classical
things
that
people
oftentimes
mention,
but
but
bringing
that
into
the
the
community
discussion.
The
sick
discussions
is
vital
in
my
opinion,
and
definitely
welcome
activity.
A
So
I
I
just
wanted
to
chime
in
as
as
someone
that
was
an
industry
analyst
for
many
years
during
this
market.
A
When
I
kind
of
look
at
the
result
of
what
was
recently
published
from
the
observability,
I
guess
what
I'll
call
the
radar
or
whatever
sorry
I
forgot
the
tech,
the
name
of
the
publication,
but
it
seemed
pretty
misleading
because
you
have
all
these
companies
kind
of
like
reading
this
out
and
then,
when
you
read
the
methodology
behind
it,
I
mean
it's
30
companies
with
a
couple
hundred
respondents
of
what
they're
using
for
observability,
and
I
don't
think
it's
a
very
good
reflect
of
the
reality
of
what
many
of
us
see
across
user
communities.
A
And
so
I
was
a
little
bit
disappointed.
I
guess
at
the
results
and
how
it's
being
constructed
by
the
mark
by
the
by
the
community.
I
guess,
but
I
understand
it
was
first
attempt
at
something
like
this
and
I
think
it's
an
important
thing
for
the
community
to
do
and
produce
it
needs
to
be
broader
in
terms
of
data
collection.
If
we're
going
to
make
some
type
of
recommendations
as
to
what
users
should
be
assessed
and
implementing,
and
using
that's
all
that
I
wanted
to
chime
in
yeah.
B
Thank
you,
I
I
think
there's
this
is
mixing
at
least
two
things,
because
we
with
my
sick
hat
on
it's
very
much
about
leading
the
way
and
about
enabling
end
users
and
others
to
to
use
certain
things
there
is.
The
survey
is
more
about
what
are
people?
What
is
the
current
state
of
the
world?
B
There's
some
more
about
the
background
of
the
mythology
in
the
poc
recording
which
just
finished
so
you
can
watch
this,
and
I
highly
suggest
this
course
cheryl
answered
quite
well
to
to
a
few
of
those
points
which
you
just
made.
So
I
highly
recommend
watching
that
recording
the
toc
recording
of
today's
date
just
one
hour
before
it's
somewhere
in
the
team.
C
Yeah,
thank
you
again
for
the
feedback.
I
think
at
least
one
thing
that
popped
out.
G
C
Statement
about
what
should
people
be
doing,
I
think
both
within
the
sig,
given
the
membership
here
and
and
the
charter
of
the
sig.
We
very
much
don't
want
to
be
prescriptive
about
what
what
specific
technologies
or
or
or
things
people
should
use.
I
think
it
was
more.
C
I
don't
want
to
speak
for
the
survey,
but
I
think
it's
more
about
capturing
what
people
have
reported
that
they
are
using
and
then
granted.
I
think
one
thing
that
we
we
talked
I
talked
about
last
hour
is:
what
could
we
do
in
this
sig
to
help
broaden
the
the
scope
of
the
the
the
survey
and
and
how
do
we
capture
people's
experiences
and
what
drives
their
decisions?
C
What
you
know
what
the
requirements
are
and
such,
and
so
that
was
one
topic
I
wanted
to
bring
here
today
to
brainstorm
around,
and
perhaps
in
our
next
meeting
have
some
proposals
around
you
know:
how
can
we
better
engage
with
industry
and
with
folks
that
are
using
perhaps
cncf
tech,
but
likely
lots
of
other
things
as
well?
C
So
if
you
have
any
suggestions
about
how
to
how
to
provide
a
broader
sample
set,
more
data
is
generally
better.
H
H
But
generally
I
mentioned
great
work
per
se,
but
having
a
sample
set
of
23
is
definitely
something
that
needs
to
be
worked
on,
and
I
think
like
putting
my
vendor
hat
on
for
for
a
minute.
The
cncf
is
very
probably
broadly
publishing.
This
is
the
state
of
observability
and
then
you're,
basing
it
on
23
companies
and
30
respondents
within
those
23
companies
is
a
bit
of
a
bold
statement.
Let's
put
it
that
way.
H
It's
it's
totally!
Fine
that
the
toy
like
approach,
it's
end,
user,
driven,
not
vendor,
driven
like
a
lot
of
other
activities
out
there,
but
you
mentioned
before
in
the
future,
there
will
be
more
possibilities
to
engage
a
a
wider
audience,
and
I
get
just
a
mismatch
of
when
I
look
at
on
on
twitter,
how
it
is
published,
what
it's
representing
versus,
what
it's
actually
representing
and
that
that
might
give
you
some
hiccups,
especially
from
from
from
the
broader
tooling
community
out
there.
H
That's
maybe
the
only
attack
point,
but
still,
I
think
it's
also
up
to
this
group
to
find
a
way
to
to
broader
which
reach
out
here-
and
you
mentioned
before
that
with
some
one
month's
head
up
heads
up.
I
think
it's
fair
that
also
other
tools
and
other
providers
in
that
space
are
engaging
more
with
their
communities
and
invite
them,
as
they
are
a
ctf
member
to
to
push
this
in.
E
I
B
I'm
going
to
make
a
point
of
order.
I
I
understand,
concerns
around
mythology
and
mythology.
I
agree
that
more
data
is
better
than
less
data.
I
understand
why,
especially
many
vendors
would
like
to
to
see
better
representation
of
what
they
themselves
are
selling
and
why
there
are
strong
incentives
to
get
into
those
things.
B
Yet,
as
the
person
who
invited
cheryl,
I
don't
want
this
to
be
a
sick
call,
turning
into
into
complaining
about
a
specific
survey,
because
there
are
channels
for
for
giving
feedback
about
about
these
things
and
such
and
like
especially
about
the
amount
of
data.
But
again
I
highly
recommend
anyone
who
didn't.
I
know.
Otherwise.
You
were
there
to
watch
that
recording,
but
I
don't
want
this
call
to
turn
into
cheryl
bashing,
because
that's
not
why
I
invited
her
and
I
feel
like
a
really
really
bad
host.
B
If
50
of
what
we
now
talk
about
is
basically
people
complaining,
so
maybe
for
for
the
context
of
this
call,
can
be
pivot
away
from
from
all
of
this
more
towards
how
to
better,
engage
with
end
users
and
and
how
to
make
the
ecosystem
wider
and
broader
and
and
positive
things.
H
Yeah,
I
just
have
to
intercept
it.
This
was
not
bashing.
This
was
just
sharing
opinions
and
you
have
been
on
this
call.
I
said
it's
a
great
move
in
the
great
direction
so,
but
I'm
totally
supporting
on
the
point
how
we
can.
B
And
again,
I'm
deliberately
engaging
with
this
one.
There
is
always
the
difference
between
sender
and
receiver.
I
understand
that
you
phrased
this
as
an
opinion.
Yet
for
me
as
the
one
who
invited
cheryl
it
very
much
came
across
us
and
you
were
not
the
only
one
making
comments
in
this
regards
anyway.
C
B
C
In
the
interest
of
moving
moving
forward,
I'm
actually
curious
not
to
call
you
out
jonah,
but
I
think
it's
probably
safe
to
say
that
the
the
bulk
of
the
folks
on
the
call
are
of
an
engineering
sort
of
bent
or
discipline
and
soliciting
feedback
and
doing
market
research
and
or
you
know,
a
survey
that
would
be
much
more
broad
that
we
as
a
sig
might
undertake
that's
a
discipline
just
like
engineering
or
product
management
or
anything
else.
And
I'm
an
engineer
enough
to
know
that.
C
A
So
I
think
it'd
be
nice
to
give
users
an
understanding
of
maybe
where
they
are
and
and
where
they
should
be
or
could
be
going
with.
With
regards
to
their
open
source
adoption.
And
then
the
question
is:
do
we
include
vendor
technologies
in
there
or
not?
And
that's
where
it
becomes
a
bit
more
interesting
questionable?
A
B
A
B
For
making
the
phone
call
of
order
again,
it's
like
we're
not
here
to
to
debate
specific
end
user
service
in
their
mythology.
I
understand
the
desire
and
the
need
to.
I
absolutely
do,
but
this
is
outside
the
scope
of
the
sick
call
within
the
scope
of
the
sick
call
and
the
reason
why
I
again
personally
invited
charles
to
be
part
of
this
and
which,
which
is
why
I'm
feeling
bad
about
basically
inviting
her
to
a
place
where
people
complain
and
again.
Please
watch
that
recording
because
part
of
your
questions
were
answered
there.
B
So,
anyway,
looking
forward
in
a
positive
manner
what
specific
things
should
or
could
we
be
engaging
with
with
the
end
user
community,
so
I
do
have
a
which
are
well
phrased
and
which,
which
like
don't
don't
steer
users
in
a
certain
direction,
but
still
show
what
we
as
as
people
who
hopefully
know
where
this
is
going,
think
they
might
be
interested
to
go.
But
what
other
things
could
or
should
we
maybe
be
doing.
C
I
have
a
proposal
or
a
concrete
idea
and-
and
it
might
seem
silly
but
to
your
point,
both
jonah
and
others.
This
whole
space
is
evolving
extremely
rapidly.
In
the
last
two
weeks,
I've
had
a
number
of
calls
with
commercial
vendors
because
again,
at
least
within
my
own
company,
we
have
you
know
a
lot
of
teams
just
use.
We
are
polyglot
to
an
extreme
and
we're
working
on
consolidating
things,
but
I've
had
discussions
with
a
number
of
what
you
would
call
like.
C
More
traditional
vendors
who
have
within
their
company's
open
source
contributors
to
cncf
projects
like
open,
telemetry
and
jager,
and
things
like
that,
but
and
others
and
I've
found
that
even
when
discussing
with
them,
they
actually
are
a
little
bit
behind
in
some
cases
of
knowing
what's
new
and
what's
happened
in
the
last
quarter
or
the
last
half
year.
So
what
would
folks
think
of
the
sig
curating
sort
of
like
a
an
activity
stream
of
sorts?
C
I
want
to
be
careful,
though,
that
we
don't
apply
a
big
filter
on
it
or
it
doesn't
turn
into
advertising.
But
you
know
I
don't
I
don't
know
what
the
right
format
is.
Maybe
like
a
a
cadenced
newsletter
or
or
updates
from
from
the
sig
or
or
if
it's
more
just
like
a
like,
I
don't
want
to
say
a
twitter
feed
because
that's
the
wrong
term,
but
but
some
way
to
to
keep
the
community
abreast
of
like
what
new
projects
are
there.
You
know
this.
H
Maybe
share
a
little
what
what
the
methodology
was
for
engaging
end
users
and
what
the
challenges
usually
are.
I
know
myself,
it's
like,
like
you,
get
like
surveys,
you're,
not
necessarily
feeling
massively
inclined
to
to
answer
them
and
maybe
just
something
that
how
maybe
some
other
work
or
some
other
activities
can
help
to
engage
more
people,
because
I
think
that's
the
part
we
agreed
on
and
richie.
H
You
can
tell
me
to
stop
if
you
think
it's
out
of
of
the
context
here,
but
how
we
can
engage
the
broader
end
user
community
to
provide
their
feedback.
I
understand
it
is
like
super
hard
and
like
from
getting
services
surveys
myself.
How
many
do
you
answer,
but
maybe
if
you
can
share
some
of
the
challenges
that
you
had
and
where
people
might
be
able
to
help.
D
So
I
do
appreciate
all
of
the
feedback
that
people
are
giving
and
richie.
I
also
appreciate
you
standing
up
and
saying
this
is
not
why
it's
not
cheryl
bashing,
I.
What
I
want
to
say
is
that
I,
my
role
in
this
is
to
set
and
run
the
survey,
but
I
think
to
get
the
questions
get
that
information
that
you're
looking
for.
D
I
think
you
need
to
be
building
the
personal
connections
with
end
users,
because
it's
not
something
that
I
can.
I
can
report
things
that
I've
heard,
but
I
want
you
to
actually
meet
those
people
and
dig
in
a
lot
deeper.
I
definitely
couldn't
give
you
like.
Okay,
here's
five
issues
that
people
want
fixed
like
that's,
not
something
that
is
my
it's
my
position
to
say
like
on
behalf
of
all
end
users
everywhere
in
the
world.
D
This
is
what
they
want
fixed
with
regards
to
observability,
so
I
do
want
to
figure
out
like,
and
I
think
this
comes
down
to
how
much
work
you
are
willing
to
put
in
not
you.
Personally,
I
mean
the
group
is
willing
to
put
in
to
connect
with
people
and
build
the
relationships,
and
that's
why
I
don't
think
that
there's
a
standardized
methodology
where
I
can
say
like
just
do
this
and
end
users
will
give
you
this.
B
Please
your
end
users
tell
us
what
you
want
to
hear
about
is
precisely
this
thing
to
offer
something
which
they
actually
care
about.
Instead
of
just
hey,
here's
another
10
000
question
survey
have
fun.
Please
ignore
us,
that's
because
like
unless
you
give
them
an
incentive
to
to
answer,
they
will
not,
of
course
like
why?
B
Wouldn't
they
because
they
don't
have
an
incentive
like
it's
almost
autologic,
so
yeah
it's
on
us
as
someone
as
a
group
who
has
been
given
a
chance
by
cheryl,
to
ask
questions
and
to
to
get
something
channeled
towards
this
community
to
come
up
with
some
value
prop
which
they
find
valuable,
easy
and
super
hard.
D
I
I
wonder
if
it
would
help
to
borrow
an
idea
from
economics
where
these
surveys
are
done
as
part
of
a
panel,
which
kind
of
repeats
every
so
often
and
part
of
the
value
proposition
could
be
that
these
people,
you
know
you-
can
publish
who's
part
of
the
panel,
and
maybe
some
people
would
like
that.
That
might
also
help
us
develop
relationships
with
these
people
so
that
it's
not
just
another
survey.
They
have
to
fill
out.
But
this
is
a
panel.
K
Yeah,
I'm
an
end-user
right.
I
work
for
a
bank
and
we've
been
trying
to
engage
more
with
cncf
and
other
open
source
communities,
and
it
takes
a
lot
of
work
right
in
order
to
on
the
main
facilitator
right
for
cncf
and
trying
to
find
the
people
inside
of
the
organization,
and
they
are
all
interested
in
participating.
But
the
day-to-day
and
the
operation
really
restrained
them
from
joining.
K
But
it's
an
effort
that
I've
been
putting
and
being
able
to
engage
some
people.
But
what
happens
in
the
day
to
day
is
that
normally
we
rely
on
vendors
right
and
the
vendors.
They
are
biased
right.
They
want
if
they
support
a
specific
solution,
and
they
would
normally
recommend
that
way
and
there's
some
sort
of
entropy,
because
we
as
end
users-
sometimes
we
don't
know,
what's
happening
in
cncf
and
maybe
many
people
that
use
kubernetes
and
some
of
these
solutions
for
observability
or
service
mesh.
They
don't
even
know
that
about
cncf.
K
So
I
think
there
are
different
things
like
there's:
a
lack
of
awareness
from
the
end
user
that
there's
a
a
whole
set
of
other
alternatives,
there's
a
community
that
are
other
solutions
out
there
other
than
the
ones
that
they
are
using
or
being
supported
by
the
the
vendors.
K
So
these
are
some
of
the
thoughts
that
I
have
that
as
an
enterprise
trying
to
get
more
value,
trying
to
be
not
to
be
locked
in
in
a
specific
solution,
but
trying
to
find
alternatives
and
still
struggling
getting
the
right
resources
within
the
company
to
join
the
community.
So
yeah,
I
didn't
say
anything
like
being
pointed
any
specific
problem,
but
just
sharing
some
thoughts.
From
my
perspective.
B
Now,
I
think,
that's
very
valuable
and
part
of
of
the
solution
is
for
more
for
more
end
users
to
to
become
part
of
those
efforts.
B
Of
course,
if
I
just
had
a
quick
look
at
at
the
attendance
list,
I
mean
matt
is
also
an
end
user
and
everyone
else
is
basically
coming
from
a
vendor
and
myself
ever
since
I
joined
grafana,
I'm
also
not
not
working
on
something
else
at
the
day
job
and
doing
this
in
my
free
time,
but
now
I'm
also
part
of
a
venue.
So
the
the
inherent
issue
you
have
is
most
of
us
are
working
for
winter.
B
Somehow,
and
this
like,
I,
I'm
convinced
that
everyone
here
is
trying
to
do
their
best
to
be
impartial,
but
yet
we
we
all
have
our
biases
and
such
so
unless
more
truly
truly
neutral
people
start
joining
and
helping.
We
can't
break
this
up
on
the
plus
side.
What
we
can
repay
you
with
is
more
in-depth
knowledge
and
and
being
able
to
actually
shape
this
in
the
way
which
you
want,
as
as
an
end
user.
B
H
One
one
thought
that
I
have
here
and
now
what
I
should
thinking
about:
putting
this
forward
in
in-app
delivery
as
well.
We
had
usually
a
lot
of
project
presentations
in
in
the
past.
The
project
reviews
what
like
talking
to
people
they
usually
find
interesting,
is
more
or
less
hearing
it
from
other
end
users
so
actively
inviting
end
users
on
how
they
address
certain
topics
or
what
they
do
internally
and
share
this
with
other
people.
H
That's
definite
value,
it's
not
coming
from
a
vendor,
it's
coming
from
somebody
who
actually
implements
and
people
who
have
done
it
there's
also
a
record
that
I
think
a
lot
of
people
would
watch
it's
totally
unbiased.
This
is
how
we
run
stuff.
This
is
the
problems
that
we
have.
This
is
what
how
we're
solving
them
and
as
you're,
building
building
this
up
into
like
regular,
sick
type
of
presentations.
H
This
could
definitely
be
helpful
and
some
people
might
be
willing
to
speak
and,
from
my
experience
from
like
other
kind
of
user
groups
or
whatever
you
want
to
call
them
once
somebody
starts
to
talk
about
their
personal
experiences.
Other
people
will
chime
in
and
will
join
the
discussion.
C
H
C
Yeah,
that's
that's
a
great
point
and
actually
I'm
I'm
actually
working
on
a
blog
post
or
two
now
for
the
coming
quarter
about
what
we're
doing
within
my
company
as
an
end
user,
we're
running
a
largely
cncf
based
stack
and
I'm
hoping
that
we
can
build
up
a
library
of
case
studies
and
or
blog
posts
that
that
we
can
curate
as
a
sig
and
not
say
this
is
what
you
know
the
sig
says
you
should
do,
but
here
are
some
n
examples
of
different
of
ways
that
that
actual
people
in
the
wild
are
assembling
these
various
building
blocks
to
meet
their
own
requirements.
C
So
I
think
that's
a
that's
well
put
and-
and
I
think
we
can
walk
the
line
you
know
between
between
providing
examples
and
or
quick
quick
starts
to
to
play
with
some
of
these
things.
For
for
new
for
new
folks,
and
maybe
we'll
have
you
know
a
snowball
and
then
we'll
start
beating
people
away
with
a
stick
for
lack
of
time.
Are
there?
Are
there
more
things
on
this
topic
or
should
we
move
on?
I
think
we've
got
a
couple.
Other
things
there
is
a
nice
segue
here
I've.
C
Actually
I
wrote
the
next
one
on
the
agenda,
I'd
like
to
solicit
recommendations
for
webinars.
Now
that
we've
got,
I
think
each
and
every
meeting
we've
had
more
attendees,
slowly
growing
from
more
from
more
companies
than
than
the
the
previous
ones.
But
I
know
litmus
is
one
that
I
personally
haven't
haven't
used
their
project,
but
are
there
other
recommendations
from
from
folks
on
the
call
about
who
we
might
want
to
reach
out
to
to
set
up
some
webinars
or
show
and
tells
or
something
like.
F
That
and
webinars
would
be
about
like
what
like
they're
they're,
presenting,
for
example,
new
project
that
we
yeah.
We
hear
a
little
about,
for
example,.
G
C
I
would
love
to
get
like
you
know
what
15
20
minutes
half
an
hour,
just
whatever
the
right
level
of
depth
is
about
what
the
project
is,
what
its
goals
are,
how
it
works
and
how
you
can
engage
or
something
like
that,
something
very
focused,
not
necessarily
on
technical,
deep
dives.
We
could
do
that
too,
but
I
think
given
where
we
are
just
gelling
the
community
around
you
know,
or
maybe
we
we
just
look
at
the
landscape
and
just
you
know,
do
do
round
robin
from
from
all
of
the
projects.
F
So,
let's,
let's
pick
some
project
if
you
picked
litmus,
let's
start
with
with
that
and
and
just
ask
them
directly
if
they
are
happy
to
do
so.
I
think
that's
a
great
idea.
C
Yeah
and
and
to
be
completely
open
and
transparent.
I
just
joined
their
slack
channel
about
20
minutes
before
the
meeting,
and
it's
just
the
one
that
I
haven't
personally
gotten
my
hands
dirty
with.
So
that's
completely
arbitrary.
That's
why?
If
there's
a
better
first
couple,
aura
folks
here
are
from
projects
and
and
and
would
like
to
put
their
put
their
hat
in
the
ring.
Please
please
do.
C
All
right:
well,
I
can
take
an
action
to
reach
out
to
the
litmus
project
administrators
are
leaders
and
and
circle
back,
either
in
channel
or
in
the
next
meeting.
If
it
takes
that
long.
F
C
Oh
yeah,
I'd
like
to
curate
a
road
map
of
them
and
just
have
them
on
a
cadence.
That
would
be
great
and
I
think
it
would
also
help
build
trust
within
the
community
and
just
you
know,
humans
like
novelty
right.
If,
in
terms
of
community
building
like
you
know,
people
will
be
more
likely
to
join.
If
we,
if
we
have
a
roadmap,
and
we
have
a
reason
for
people
to
come
to
learn
new
things.
F
Yes,
I
I
would
be
also
maybe
before
we
start
actually
asking
people.
Maybe
we
can
agree
what
we
can
share
in
terms
of
what
agenda
we
are
expecting
right,
because
I
don't
think
we
want
to
spend
you
know
maybe
20
minutes
in
how
to
deploy
it.
What
amazing
feature
it
gives.
F
What
problems
we
they
having,
what
we
can
help
with
or
what
they
expect
for
sick
observability
to
help
them
with
right.
Maybe
something
on
this
tom
as
well.
C
Of
it's
not
too
prescriptive,
but
it
provides
some
guidance
because
again
we
don't
want
just
marketing.
We
don't
we
don't
want.
You
know
you
know,
architecture
type
stuff,
but
okay.
Well,
maybe
that's
a
better
goal
for
the
next
meeting,
for
so
maybe
over
the
next
two
weeks,
I'm
happy
to
chime
in
but
well
I
can
make
a
github
issue
for
it
or
and
we
can
just
track
it.
C
People
can
contribute
that
way
and
again
you
know
if
anyone
on
the
call
here
has
experience
doing
this
in
the
past
and
has
guidance,
please
feel
free
to
chime
in
to
the
github
issue.
I'll
put
it
in
the
notes.
A
little
later
does
anyone
else
have
other
feedback
before
we
move
on.
C
Okay,
I
guess
the
next
one's
all
you,
okay,.
F
So
yeah,
I
don't
have
too
much
to
do
too
much
of
new
status.
I
was
on
holidays,
but
then
we
definitely
move
on
our
little
project
of
slitix.
F
F
So
it
will
allow
for
materials
to
park
that
file,
and
this
was
kind
of
needed
internally
but
like
we
are.
We
want
to
extend
that
and
explore
and
experiment
with
different
products,
well
different
inputs,
maybe
maybe
in
3db,
maybe
cortex,
but
also
different
output
files,
so
it
needs
to
come
up
like
this.
That
will
be
kind
of
efficient.
F
Ideally,
please
turn
the
customer
pro.
C
C
Oh
wow,
can
you
be
a
part-time?
I
think
we're
getting.
C
Anyway,
so
what
we've
built
so
at
every
quote,
we
have
our
own
anomaly
detection
back
end.
C
That's
worked
off
some
custom
time
series
database
and
in
the
last
two
weeks
my
update
from
there
is
we've.
We've
now
got
an
mvp
working
of
using
prometheus
data.
C
We
basically
just
start
with
a
prom
ql
query
and
hit
our
cortex
deployments
in
our
case,
but
it's
just
a
prom
ql
end
point
and,
and
the
actual
analysis
is
still
done
on
a
copy
of
the
data-
that's
kept
in
a
mysql
database,
so
you
know
we're
we've
added
prometheus
as
a
time
series
and
the
folks
that
have
built
this
service
have
been
surprised
by
prometheus
and
and
particularly
some
of
the
things
that
can
be
done
prior
to
query
time
around
setting
up
recording
rules
to
to
pre.
C
You
know
the
the
not
to
go
too
deep,
but
we
we've
got
sort
of
the
the
legacy
pre.
I
don't.
I
never
know
how
to
say
it's
absolutely
obsolete
your
new
project,
and
so
we
would
hope
that
in
the
next
two
to
four
weeks,
once
we've
got
more
at
scale
versions.
Working
of
of
you
know
not
using
this.
We
could
then
have
a
baseline
with
which
to
compare
using
this.
L
I
can
jump
in
from
this
analytics
case
and
offset
6
2.
bartek
started
but
dropped.
I
think
he
will
try
to
say
there
is
a
plan
for
the
much
much
more
performant
way
to
do
this
kind
of
file,
format,
commerce,
convert
conversion.
L
F
No,
I
think
azan
is
saying
that
actually
we
are
thinking
that
initially,
the
problem
is
that
this
memory
model,
like
it's
really
language,
dependent
right
like
if
you
are
using
python,
you
are
winning
because
arrow
is
in
python.
However,
in
golang
those
things
are
very,
very
initial.
Let's
say
there
is
some
basic
support
for
it,
but
the
problem
is
it's
not
really
portable
between
you
know
kind
of
processes
and
stuff
like
that,
so
and
that's
why
this
arrow
flight
was
proposed,
and
this
is
essentially
a
grpc
api.
F
However,
it's
not
well
introduced
yet
and
the
main
point
behind
you
know
our
try
with
this
obsoletics
is
that
first
we
need
to
establish
what
apis
people
want
and
once
they
start
using
it,
then
we
can
think
how
can
we
can
optimize
and
scale
it
out?
But
you
know
the
nothing
was
there
even
to
start
with
to
use,
promote
use
data
in
order
to
to
solve
some
analytic
queries.
So
once
we
have
this
connection,
the
plan
is
to
you
know,
actually
find
the
pinpoints
pain
points
and
and
improve
the
scalability
there.
F
So,
let's
start
with
something,
but
yes,
there
is
definitely
a
way
forward
to
make
it
more
optimized
if
needed.
C
So,
if
folks
want
to
use
other,
if
folks
want
to
use,
I
don't
know
where
the
echo
is
coming
from.
Is
it
me
anyway,
if
folks
wanted
to
use
other
back
ends
like
time,
scale
db
or
other
things
that
support
remote
read
remote
right?
Is
it
possible
to
to
to
do
that,
or
is
it
still?
Is
the
current
state
that
you
really
need
to
be
using
thanos
in
order
to
to
experiment
with
the
api.
L
Project
in
apache
site,
there
is
a
some
libraries
working,
for
example,
opposite
spark
merged
that
project
to
better
support,
pandas
data
frame
conversion.
L
F
Yeah,
I'm
not
sure
if
you
are
talking
about
the
same
thing,
but
the
essentially
what
what
you're
asking
here
right
now
thanos
and
pro
materials
are
supported,
because
this
is
what
we
what
input
we
need.
However
yeah
I
mean
it's
very
it's
just
adapters,
so
we
are
essentially
having
all
those
formats
from
tanos
frontiers
into
the
data
frame
in
memory
and
then
we
we
just
output,
parket
or
flight
or
whatever.
That's
that's
the
current
kind
of
idea.
F
So
whatever
arrow
will
improve
apache
arrow,
we
can
definitely
use
in
order
to
move
to
panda
and
maybe
to
other
spark
and
so
on.
So
definitely
we
are
looking
into
those
areas
as
well.
F
Yeah,
we
are
just
experimenting
so
please,
if
you
are
in
heavy
need,
for
example,
in
some
specific
conversion-
maybe
I
don't
know
like
analytics
on
log
lines
or
tracing
or
things
like
that.
We
are
definitely
want
to
hear
your
feedback
if
you
are
even
looking
into
reviews
like
this.
C
Cool,
I
think
that's
all
we
had
on
the
pre-done
agenda.
We've
got
a
couple
of
minutes
left
open
floor.
Is
there
anything
folks
would
like
to
announce
raise.
E
Discuss
hi,
this
is
sungku,
so
one
of
the
topics
in
the
or
our
chat
in
the
slack
was
about
discussing
the
link
I
posted
with
the
opnfv
barometer.
G
C
E
Thank
you
yeah,
so
pretty
much
work
in
the
open
v,
open
platform
for
network
function,
virtualization
community,
and
so
one
of
the
project
in
that
community
is
parameters
which
focuses
on
monitoring
and
exporting
the
metrics
via
prometheus
or
influx,
and
provide
a
better
monitoring
reference
solutions
to
rest
of
the
nfv
infrastructure.
E
So
within
nfv
infrastructure
we
focused
on
openstack
and
kubernetes
based
deployments
so
within
parameter.
We
provided
a
reference
example
of
how
you
could
deploy
something
like
collect
d,
for
example,
with
prometheus
and
grafana,
how
you
could
visualize
various
hardware
metrics,
and
so
we
also
our
team,
consists
of
like
red
hat
and
intel
folks,
mostly
I
I
work
for
intel.
Essentially,
so
we
contribute
to
collect
the
telegraph
and
looking
into
node
exporter
communities,
contributing
various
hardware,
plugins
and
integration
points
within
these
communities.
E
So
the
one-click
install
link
that
I
have
posted.
There
is
just
a
set
of
ansible
scripts
that
you
can
deploy
to
have
easy.
You
know,
collect
the
influx
and
grafana
stack
and
now
we're
looking
to
provide
a
similar
deployment
with
prometheus,
so
yeah
open
to
questions.
If
there
is
any
discussion
here.
F
F
So
that's
our
goal
and
definitely
we
are
looking
into
adding
more
projects
as
well,
not
only
for
me
to
use
so
the
question
here
is
that
what's
your
goal
behind
all
of
this,
do
you
care
about
like
one
click,
installation
of
production,
deployment
or
just
demo
or
and
what
exactly
you
know,
projects
you
want
to
have
involved.
I
guess
promise
only
right.
E
Yeah
yeah,
so
I
mean
looking
at
more
or
less
the
observability
stack.
Overall,
I
guess
the
main
goal
for
us
is
twofold:
one
is
from
opnv
community
perspective.
E
The
overall
community
is
looking
to
provide
a
reference
stack
for
telco
providers,
service
providers
and
operators
so
that
they
can
deploy
openstack
or
kubernetes
have
methodology
to
do
lifecycle,
management
and
observability,
so
essentially
to
integrate
our
reference
stack
into
that
reference
solution,
and
then
telco
service
providers
would
use
that
reference
solution
to
have
it
production
ready
within
their
environments.
So
that's
openv
side
of
things
and
other
side
at
least
from
intel.
E
You
know
the
customers
we
work
with
is
provide
this
reference
stack
with
a
set
of
metrics
and
configurations
where
they
can
easily
understand
what
is
prometheus,
what
kind
of
metrics
they
would
get
with
collect
d,
for
example,
or
telegraph,
and
understand
how
would
it
would
impact
their
production
environments
so
that
they
can
move
on
to
something
like
prometheus
or
collect
the
that
type
of
solutions?
E
So
so
one
is
so
essentially
providing
examples
to
the
folks
we
work
with
as
to
what
can
be
possible
using
these
software.
Using
this.
F
Software
yeah
amazing.
No,
I
think
this
is
definitely
a
similar
similar
goal.
I
would
yeah.
So
that's.
The
key
part
is
to
make
sure
you
connect
with
people,
and
you
announce
this
in
the
I
guess
from
to
use
community.
You
know
some
way
of
telling
hey.
This
is
how
you
can
contribute
this,
and
this
is
how
we
I
don't
know
we
are
looking
for
tips
depending
if
you
want
to
do
it
internally
or
or
so.
F
What
but
you'll
be
amazing
to
have
essentially
like
teamwork,
on
this,
to
enable
reference
architecture
with
prometheus
and
then
maybe
others
other
systems
as
well
but
looks
like
we
are
essentially
looking
on
the
metric
side
for
now
so
definitely
yeah,
and-
and
if
you,
if
you
have
that,
then
this
can
be
used
for
for
our
demo
purposes
as
well.
That
makes
sense
what
exactly
you
are
using
for
this.
F
I
don't
know
one
click
installation.
Is
there
any
pattern
technology
behind
that
any
language
or
how
do
you
construct
those
things
how
you
deploy
those
things.
E
Sorry
good
here
it
go
yeah,
it's
uncivil
scripts,
where
you
can.
You
know,
building
docker
containers
with
collective,
prometheus
and
grafana
and
with
the
preset
of
configurations
pre-set
of
metrics
plugins,
so
that
you
can
deploy
and
visualize
in
grafana.
What's
going
on
on
top
of
this,
of
course,
I
also
run
the
closed
loop
automation,
working
group
within
opmv,
so
we're
building
solutions
there
you
know
so,
okay,
how
can
you
leverage
alert
manager?
E
How
can
you
leverage,
kubernetes
extensions
like
telemetry
over
scheduler
integrated
with
prometheus,
and
how
can
you
leverage
this
for
your
vnfs
or
your
container
workloads,
low,
latency,
backup,
processing
workloads,
so
that's
again
we're
building
some
reference
solutions
there.
E
So
when
you
mention
provide
this
info
in
the
prometheus
community,
of
course
I
posted
this
in
slack.
But
how
else
would
you
suggest
I
share
some
of
the
work
going
on.
F
Yeah
sure,
first,
first
of
all,
I
think
there
is
already
ansible.
You
know
how
it's
called
play,
playbooks
or
something
for
for
promoters,
and
we
are
using
that
for
our
demo.
F
Promote
use
examples.
So
maybe
it
will
be
just
easy
as
talking
to
to
the
right
people
and
they
can
share
what
they
have
and
it
will
be
much
easier
to
start
with.
So
what
I
would
do,
I
would
go
to
prometheus
website
and
there
is
community
section
and
then
just
choose
the
main
entries
for
me
to
use
def,
probably
and
just
describe
the
same
thing
you
described
on
the
slack
channel,
and
this
is
the
mailing.
F
E
Yeah
for
sure,
I'm
hoping
to
be
a
regular
to
this
meeting
so
as
we
have
so
a
couple
of
demos
ready,
so
we
could
share
it
here.
C
Yeah
thanks
thanks
very
much
for
your
feedback
and
your
contribution
to
the
discussion.
I
think
we're
just
about
out
of
time
at
which
point
the
zoom
from
the
cncf
will
arbitrarily
cut
us
all
off.
So
we
can.
We
can
leave
the
way
we
want
to,
or
in
about
100
seconds
we
can
leave
the
way
zoom
will
have
us
leave,
but
is
there
anything
else
that
that
folks
want
to
give
a
shout
out
about
before
we're
cut.
C
Off
all
right,
well
with
that,
I
hope
everyone
has
a
safe
couple
of
weeks.
I
guess
I'll
see
you
online
and
in
two
weeks
and
thank
you
again
for.