►
From YouTube: 2022-02-10 Governance Committee private 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
B
We,
the
only
ones
here
today
I
know
today-
was
the
the
monthly
community
meeting.
So
I
don't
know,
usually
we
don't
do
both.
C
A
B
B
E
Calendar
stuff's
driving
me
nuts
sergey
thinks
he
so
sergey
was
talking
to
some
people
in
kubernetes.
He
thinks
some
of
the
issues
we're
having
not
the
zoom
ones.
Those
are
unrelated,
but
but
some
of
them
are
related
to
like
the
number
of
people
who
subscribe
to
a
calendar.
There's
some
issue
with
it
and
so
kubernetes
had
to
go,
create
a
g
suite
organization.
E
Basically,
so
I
have
an
action
to
go
reach
out
to
amy
to
see.
If
we
can
get
the
same
thing,
it
would
explain
why
the
events
on
the
calendar
are
sometimes
blocked
from
being
forwarded
to
the
mailing
list.
That's
supposed
to
just.
E
Subscribe
to
it
and
there's
various
ones
that
there's
an
issue,
some
getting
deleted
from
the
calendar
there's
another
issue
where
some
are
still
on
the
calendar,
but
the
mailing
list
will
refuse
to
acknowledge
that
they
exist.
Even
if
you
edit
them,
you
have
to
delete
them
and
recreate
them,
and
it's
just
wacky
and
then
there's.
E
E
One
one
I
mean
what
kubernetes
ended
up
doing
was
one
zoom
account
per
sig,
which
is
very
expensive
for
the
cncf,
but
I
mean
I'm
sure
the
bigger
expenses,
but
but
they
couldn't
find
any
way
around
it.
I
try
experimenting
with
what
we're
doing
now,
but
the
odd
time
it
gets
in
a
conflict,
and
it
says:
oh,
a
single
meeting
owner
can't
have
two
meetings
so
really.
E
D
E
F
D
D
It
has
account
limits,
though
I
think
you
have
to
switch
to
some
other
kind
of
enterprise
account.
D
G
E
F
D
C
E
I
I
think
the
cncf
has
the
tools
I
haven't
had
a
chance
to
reach
out
to
amy
in
the
last
week,
because
I've
been
absolutely
slammed
because
we're
in
like
month
or
quarterly
planning
starting
next
week.
I
should
have
a
bit
more
time.
I'll
talk
to
amy,
hopefully
we'll
get
a
g
suite
subscription
that
will
solve
the
calendaring
issues.
E
E
It
just
vanishes
from
into
nothing
and
secondly,
I'll
get
those
zoom
accounts
back
well
more
than
getting
back.
We
still
have
the
three
or
four
that
we
had,
but
we
need
more
and
I
will
get
us
more.
So
it's.
B
D
E
Same
one,
okay,
all
right,
you
know
what
I'm
gonna
go
recreate
some
of
the
meetings
tomorrow
then,
with
the
four
zoom
accounts,
we
have
and
just
roll
that
back
quickly,
because
this
is
the
problem
of
zoom.
This
is
the
zoom
specific
one
where
we
have
one,
we
tried
to
get
clever
and
have
one
zoom
account
making
multiple
meetings.
Zoom
gives
very
cryptic
errors.
You
actually
get
a
different
error
on
the
website
versus
the
app
one
of.
E
I
will
go
fix
this
friday,
daniel
apologies.
I
didn't
realize
you're
still
running
into
that.
B
Yeah
no
problem.
Well
I
mean
it
is
a
problem,
but
no
problem.
D
Okay,
second,
moving
on
so
morgan,
it's
your
action
item,
yeah.
A
Yeah,
so
the
first
one
is
just
it's
just
a
follow-up,
so
we
talked
about
this
email
from
tin
last
week
and
I
I
emailed
in-
and
they
said
they
would
send
an
email,
an
update
this
week
and
they
just
sent
so
it
should
be
on
your
inboxes
if
you
haven't
seen
yet
look
for
an
email,
titled,
open,
telemetry,
maintainers
reminder
and
so
on
so
forth.
A
Basically,
there
is
a
link
here
on
the
second
item,
so
both
chaos
and
virtual
presence
there's
a
link
there
with
all
the
information
that
we
need.
Basically,
I
think
the
the
most
important
piece
of
information
here
is
that
priority
is
given
based
on
importance,
so
graduated
projects
have
a
precedence
over
incubating,
which
has
which
have
residence
over
sandbox
and
so
on
so
forth.
A
So
I
assume
we
would
be
quite
good
in
in
the
priority
queue
because
you
know
we
are
not
graduated,
but
there
are
not
that
many
graduated
projects
yet
so
I
think
we
just
need
to
define
what
do
we
actually
need,
or
what
do
we
actually
want?
So
is
it
only
chaos
plus
bug,
bash
or
just
project
office
hours,
and
so
on?
A
From
my
side,
I
cannot
commit
to
any
to
anything
for
this
one
here,
I'm
not
attending
cubicon,
but
I'll
be
more
than
happy
to
help
you
all
up
before
the
event
until
yeah
mark
and.
D
E
E
D
Yeah,
I
think
it
was
great,
so
I'm
just
saying
morgan
will
put
a
proposal
together
and
join
it.
D
Great
yeah
submit
a
submit,
and,
and
this
is
a
town
hall.
E
C
This
is
only
if
people
want
to
hear
about
it.
Last
week
I
thought
we
had
an
interesting
discussion
about
projects
that
are
more
really
adjacent
to
hotel.
I
really
in
fact
strongly
believe
it
should
not
be
part
of
hotel
around
the
standardization
of
I
would
call
like
totally
commoditized
basic
monitoring.
You
know,
like
dashboard
type
stuff,
alerting
thresholds
slo
that
category
of
things,
so
I
did
have
a
conversation
with
some
sort
of
end-usery
people
about
this,
and
I
just
wanted
to
report
back
about
what
I
heard.
C
C
Not
anymore,
oh
good
ali,
can
you
do
your
thing?
First,
that's
actually
kind
of
more
important
than
a
generalization
of
what
I'm
talking
about
this
customer
user
group
stuff.
D
Yeah
sure
so
just
wanted
to
let
you
know
that
you
know
again:
I've
been
working
with
char
from
new
relic
to
get
organized
about.
You
know,
setting
up
a
user
group.
We
had
to
discuss
this
with
ted
in
the
communications
group
there
and
also
at
the
gc
meetings,
and
there
is
a
meeting
scheduled
that
we're
planning
to
have
starting
next
next
week
and
we
want
to
do
it
at
the
monthly
cadence.
D
Just
you
know
an
hour
where
some
of
us
can
join
in
for
customer
questions
or
user
questions,
and
it's
a
user
group
kind
of
meeting
again.
We
don't
want
to
impact
any
of
the
hotel.
You
know
maintainers,
but
it
really
is.
You
know
where
a
lot
of
users
or
customers
are.
You
know
using
hotel
at
this
point
and
they
have
questions.
So
it's
a
good
way
of
actually
listening
and
having
you
know
those
discussions
and
being
able
to
identify.
D
You
know
what
are
the
ways
we
could
actually
provide:
more
knowledge,
more
training
or
configurations
or
whatever
conveniences,
that
we
can
provide
for
adoption
of
hotels
so
that
that's
just
that's.
That
was
the
heads
up
that
I
just
wanted
to
make
sure
everybody
was
aware
of.
C
Cool,
that's
great,
I'm
really
glad
you're
organizing
that
alita.
Thank
you.
I
think
that's
maybe
two
or
three
years
overdue,
but
I
think
it's
really
important.
So
I
guess
the
interesting
feedback
about
the
idea
of
like
creating
some
some
sort
of
initiative
hotel
or
not
around
standardization
and
monitoring.
This
came
up
again
to
just
review
from
last
week,
because
not
one
but
like
two
or
maybe
depending
how
you
count
like
three
customer
conversations.
C
I've
had
in
the
last
couple
of
months
have
involved
people
doing
homegrown
things
to
create
kind
of,
like,
I
would
say,
broadly
speaking,
like
declarative
definitions
of
monitoring
that
can
be
kind
of
stamped
out.
Jurassic
mentioned
the
mix-ins
stuff,
which
I
frankly
wasn't
aware
of
before
from
grafana
and
talk
through
these
things
anyway.
It's
interesting
that
the
conclusion
from
the
conversation
was
that
this
is
going
to
be
very
difficult
to
do
in
a
general
purpose
way.
Right
now,
without
the
quote.
C
Unquote,
declarative
configuration
is
taking
a
very
hard
dependency
on
the
particular
observability
tooling,
because
semantic
conventions
are
not
mature,
and
that
was
really
the.
If
I
could
summarize
what
was
a
very
interesting
like
long
conversation,
the
most
important
thing
was
that
the
semantic
conventions
work
in
hotel
until
that's
really
stabilized.
C
It
makes
it
almost
impossible
to
pursue
this
kind
of
effort,
as
you
end
up
having
to
at
some
level,
define
the
actual
semantic
queries
you're
making
and
if
you
can't
find
a
surefire
portable
way
of
describing
what
whether
you're
monitoring,
infrastructure
or
slos
or
whatever
until
that
stuff
is
really
predictably
named
and
predictably
shaped,
it's
almost
impossible
to
create,
like
a
portable
declarative
configuration
for
something.
C
So
I
thought
it
was
kind
of
interesting
and
kind
of
worth
reporting
on,
because
that
I
mean
I
know
we're
already
engaging
that
work
and
I
think
people
realize
it's
important,
but
it
was
maybe
just
more
of
a
realization
that,
once
that
work
is
truly
complete
and
like
out
in
the
world,
I
think
a
lot
of
pretty
interesting
adjacent
projects
become
unblocked.
C
That
would
be
quite
powerful
and
I
think
disruptive
in
similar
ways
to
hotel
in
that
they
can
commoditize
something
that
shouldn't
be
very
vendor
specific
and,
I
think
won't
be.
Does
that
make
sense
anyway?
That's
like
the
main
thing.
I
learned
I'm
happy
to
discuss
it,
but
I
just
wanted
to
share
that.
A
C
I
should
probably
invite
this
person
to
the
the
user
thing
that
we're
talking
about,
but
this
person
is
like
super
bought
in
on
hotels
being
a
good
idea
and
they've
been
pushing
for
it
for
a
long
time,
like
you
know,
certainly
for
multiple
years
at
this
point,
and
I
think
that
they've
found
that
with
this
person
feels
like
they're
about
to
lose
a
war
of
attrition
and
that
they
cannot
force
engineers
to
use
this
thing
and
that
they
feel
like
they
kind
of,
did
it
wrong
and
they
should
have
gone
in
with
a
value
prop.
C
That
was
more
about
the
developer.
Experience
the
good
side
of
it
like
fixing
a
monitoring
problem,
or
something
like
that,
which
is
how
this
whole
discussion
came
about
and
that
pushing
on
hotel,
even
though
from
a
you
know,
if
you're
thinking
green
field
totally
makes
sense
like
get
the
telemetry
layer
sorted
out
first
and
then
move
to
storage
and
then
move
to
you
know
value,
propositions
and
workflows.
C
You
know
trying
to
actually
push
this
technology
on
a
complicated
organization
with
many
views
and
stuff
like
that,
was
pretty
painful
to
start
with
hotel,
even
if
everyone
kind
of
agrees
as
an
engineer,
it
makes
sense,
and
so
so
I
guess
the
other.
You
know
feedback
was
just
that.
You
know.
Hotel
is
really
hard
to
deploy
without
a
carrot
for
the
developers
that
need
still
need
to
do
work,
at
least
until
auto
instrumentation
is
like
truly
seamless,
which
it
really
isn't
right
now
so
yeah.
D
I
mean
again
and
practically
I
totally
agree
with
you
ben,
because
you
know
one
of
the
things
that
I'm
sure
all
of
us
are
seeing
in
you
know
respective
and
especially
given
we
have
a
distribution,
you
know
which
is
being
pushed
and
being
used
in
fraud.
Even
is
that
again
the
two
big
barriers
you
know
or
just
hardships
that
customers
face
constantly
is
configuration
management
and
you
know
having
a
seamless
way
of
interoperating
with
existing.
D
D
D
So
I
mean
ben:
do
you
see
any
practical
ways
that
you
know
this
could
be
actually
taken
up
as
a
you
know,
as
a
project
under
the
hotel
umbrella,
because
we
do
need
to
address
it
for
our
own
needs
right
I
mean
for
for
open
telemetry
the
collector,
as
well
as
the
to
be
easily
configurable
and
at
the
same
time
also
manageable
and
and
the
sdk
is
instrumental
right.
So.
C
I
mean
sort
of
stating
the
obvious,
but
I
think
that's
that's
the
I
mean
that's
been
the
primary
blocker.
I
think
for
a
lot
of
things.
Hotel-Wise
in
the
last
year,
I'd
say
and
that's
not
meant
to
be
a
dig
on
anyone.
I
mean
I'm
certainly
not
helping
alida
you're
doing
more
than
probably
anyone
on
this
call.
I
think,
as
far
as
that
sort
of
work
is
concerned,
but
I
do
think
that
that's,
like
probably
more
of
the
rate
limiting
factor
than
feasibility,
but
I'm
curious
other
people
think
it's
a
good
question.
G
We
are
trying
to
define
the
structure
of
the
metrics
without
defining
these
rules
and
is
more
or
less
hypothetical.
We
everyone
imagines
them
their
their
own
rules
or
whatever
the
instead
of
having
concrete
monitoring
rules
for
for
this.
C
Yeah
bogdan,
I
I
didn't
want
to
bring
up
other
thing
that
came
up
in
my
hour-long
conversation
with
this
person,
which
is
that
this
is
sort
of
like
a
you
know,
a
true
believer
on
the
observability
front
and
then
there's
this
whole
other
conversation
about
like
what
are
metrics
and
logs
really
for
anyway,
and
so
there's
like
this
other
point
of
view,
that
metrics
should
primarily
be
derived
from
transactional
data
and
so
on
so
forth,
which
is
valid.
I
mean
I
I
mean
like
from
a
academic
standpoint.
C
I
totally
agree,
but
you're
well
anyway,
you're
totally
right
about,
I
think
like
without,
without
having
a
clear
sense
of
what
the
queries
whether
they're
you
know,
ad
hoc
queries,
are
sort
of
the
sort
of
standing
query
of
an
slo
actually
look
like
and
having
that
workload
kind
of
sitting
in
front
of
you
can
be
difficult
to
reason
about
these
things
like.
I
personally
believe
that
metrics
have
a
long
and
glorious
future
for
measuring
all
sorts
of
infrastructure
and
things
like
that.
C
But
for
measuring
transactional
things
like
you
know,
queries
per
second
or
latency
or
whatever
it
probably
does
make
more
sense
to
use
the
tracing,
instrumentation
and
tease
something
off
in
the
data
pipeline
or
in
the
collector
or
whatever.
But
it's
it
doesn't.
It
doesn't
make
sense
that,
starting
with
the
query
workload.
G
Wait
wait,
wait
but
but
no
matter
what
is
the
source
could
be,
the
direct
instrumentation
could
be
tracing.
We
still.
You
are
still
producing
the
method
out
of
this
and
that
metric
that
metric
that
how
we
name
it,
what
are
the
the
attributes
or
dimensions
that
you
have
on
that?
What
are
the
the
whatever
format
of
that
metric
and
all
of
these
are
coming
from
the
the
rules
like?
Yes,
I
think
I
think
I
think
I
think
it's
it's
exactly
like
how
people
are
describing
how
you
build
a
product.
F
Yeah
right,
like
I
think,
the
challenge
here
is:
we
need
to
not
start
from
the
data
model.
We
need
to
start
from
the
problems
that
the
customer
is
having
as
well
as
boxing
is
saying,
but
I
think
the
thing
that
I
would
challenge
on
is
that
in
general
we
shouldn't
need
to
be
constrained
to
kind
of
pre-aggregating
the
slo
right
like
in
this
future
glorious
worlds,
which
you
know
I
fortunately
live
in
right.
F
Like
you
can
say,
I
want
to
include
in
scope
for
my
slo
only
queries
from
the
specific
back
end
to
the
specific
path
with
this
specific,
but
not
these
particular
client
ids
right.
So
I
think
there
is
value
in
saying
you
know
for
the
for
the
purpose
of
the
slo
standardizing
things
like
you
know:
okay,
sure,
you
know
latency
and
error
right
like
those
are.
Those
are
two
of
the
simplest
things
that
people
need
to
need
to
have
a
definition
of
that.
F
C
Yeah
I
mean
I
think
in
that
world
liz
the
the
quote-unquote
metric,
which
I
think
I
would
call
time
series
in
this
case,
because
it's
not
instrumentation
is
an
optimization
and
for
like
a
certain
type
of
storage
and
so
on
so
forth.
But
I
don't
know
I
mean
I
said
I
I
intentionally
omitted
that
piece.
C
That's
such
a
rat
hole
when
I
was
initially
describing
the
conversation,
but
but
that's
sort
of
the
challenge
we
have
is,
I
think,
if
we
start
talking
about
customer
problems,
I
think
we
have
people
who
are
you
know
living
in
the
future
like
liz
or
whatever,
and
people
who
are
very
much
living
in
an
ultra
brownfield
kind
of
itops
world
trying
to
just
sort
of
drag
people
into
the
cloud
or
something
like
that
where
it's
like
a
complete
non-starter
right
so
which
customers,
I
think,
might
be
the
question
liz.
C
It's
like
you
know
what
I
mean
and
that's
going
to
probably
drive
the
direction
of
something
like
this
right.
Like
are
you?
Do
we
have
the
luxury
of
of
of
thinking
of
this
as
a
screen
field?.
F
Yeah
and
I
think
that
that's
kind
of
where
the
folks
from
noble
nine
and
open
slo
are
doing
good
work
to
make
sure
that
there
is
kind
of
that,
as
I
was
describing
the
lowest
common
denominator
right.
D
So
I
mean
as
an
as
an
evaluation.
You
know
that
one
would
do.
Is
it
worth
actually
looking
at
the
open,
slo
work,
that's
being
done
and
intersecting
the
lowest
common
denominator
with
our
conventions,
you
know
again,
how
do
we
make
sure
that
that's
something
we
are
aware
of
in
in
what
we
build
in
what
we
spec.
C
D
C
Solve
that
without
saying,
slos
and
slow
definition
are
now
in
scope
for
open
telemetry,
I
mean
we
could
say
a
job
to
be
done
for
open.
Telemetry
is,
to
you
know,
be
totally
friendly
and
compatible
with
the
approach
taken
by
oslo
or
open
that
slow,
whatever,
for
instance,
and
then
we
could.
We
could
be
evaluating
that
as
we're
kind
of
figuring
out
what
we
build,
but
I
am
a
little
rough.
I
agree
right.
F
Stakeholder
as
a
consumer
of
the
data
that
ultimately
winds
up,
passing
through
a
telemetry
system
of
some
kind
right,
like
you
know,
we
are
deliberately
agnostic
about
the
form
of
the
telemetry
system
on
the
on
the
end.
But
if
those
semantic
conventions
are
being
repeated
verbatim
right
like
then,
we
have
some
duty
to
make
sure
that
it
is
as
easily
consumable
for
downstream
api
use
like
open,
open.
C
I
want
to
say
one
thing
because
I
have
to
drop
actually,
ironically
for
a
customer
call
in
a
minute
but
elite
about
the
meeting
with
end
users.
I
mean-
maybe
not
in
the
first
meeting,
but
one
thing
that
could
be
incredibly
valuable
is
for
the
end
users
to
help
curate
some
list
of
jobs
to
be
done
for
otel,
basically
like
if
hotel
doesn't
do
these
things,
it's
not
useful
for
me,
yeah.
D
Yeah,
I
agree,
I
agree
ben
and
I
mean
that's
the
dialogue
again,
which
I
think
would
give
us
some
valuable
feedback
in
also
prioritizing
or
being
aware
of.
You
know
some
of
the
limitations
that
customers
face
as
pain
points
right
which
which
the
project
does
need
to
address
in
order
to
be
widely
adoptable.
D
Good,
I
think,
we're
at
time
anything
else
we're
good
thanks.
Everyone
take
care.