►
From YouTube: Metrics Models Working Group Meeting May 23, 2023
Description
Meeting Summary can be found on Discourse: https://chaoss.discourse.group/t/metrics-models-working-group-meeting-summary-may-23-2023/148
A
A
No
so
hi
everybody
Welcome
to
the
chaos
metrics
model
meeting
it's
good
to
have
everybody
here,
I
think
we
have
quite
a
few
things
on
the
agenda
for
today.
So
there
was
the
this
is
the
seemed
like
there
were
two
kind
of
open
questions
and
Don
I'm
kind
of
looking
at
you
one.
It
looked
like
around
the
project,
starter
Health
metric
model-
you
had
a
conversation
with,
was
it
was
it
Luis?
Was
it
that
kind
of
about
the
metrics
that
are
in
there
as
well?
C
Yes,
I,
but
the
the
conversation
with
Luis
is
specifically
about
how
we're
measuring
the
change
closure
ratio.
Okay,
he
had
a
very
different
idea
of
what
he
thought
we
meant
by
that
and
he
was
trying
to
clarify
it,
but
what
he
meant
by
it
wasn't
at
all
what
we
actually
meant
by
the
metrics
so
I'm,
just
trying
to
clarify
with.
C
A
Problem
thanks
for
that,
though,
and
then
how
about
the
other
part
the
one
with
yahui
that
was
I!
Think
more
specific
to
this.
C
Yeah
and
I
think
I
think
this
brings
up
a
broader
question
for
for
the
chaos
project
as
a
whole,
because
as
far
as
I'm
concerned,
when
when
we
that
we,
the
Google
Docs
that
we
use
to
draft
the
metrics
are
a
way
of
collaborating
on
a
metric
until
we
get
it
in
GitHub
and
then
from
that
point.
C
B
C
Yeah
but
I
know
people
like
to
go
back
to
the
word
doc.
The
next
time
we
revise
the
metric
and
I
think
that's
also
a
mistake,
because
the
the
metric
could
have
a
different
direction,
even
if
it's
even
if
it's
just
you
know
a
matter
of
like
a
few,
you
know
typos
or
errors.
You
know
we.
Those
would
end
up
creeping
back
in
if
we
continue
to
use
those
those
documents.
C
So
the
question
that
yahui
had
was
hey.
I
was
looking
at
the
word
doc,
and
it
still
has
time
to
close
in
it,
and
I
was
like
yeah
that
shouldn't
be
there,
that
we
should
be
using
the
one
in
in
the
on
the
website
for
the
metric
yeah.
D
Yeah,
it's
like.
Actually,
the
question
from
me
actually
comes
up
from
the
compass
live
I
I'm,
looking
at
Matrix
exits
like
in
that
model,
I
I,
compared
with
the
with
the
models
existing
chaos.
They
are
slightly
different.
I
I
find
one
more
Matrix
exactly
in
canvas
live
which
is
not
existing
in
the
in
the
chaos
matrix
model.
So
I'm
trying
to
look
in
bike,
the
Google
Docs
we
had
so
I
saw
that
made
a
mistake.
D
We
made
Implement
such
matrix
model,
so
I
found
that
when
we
implement
this
matrix
model,
I
mean
the
starter
project
house.
We
do
have
that
metric,
but
the
ones
we
upload
to
the
a
I
mean
the
deployed
in
the
chaos
website.
This
magic
is,
is
removed,
so
I'm
thinking
it's
a
collaboration
between
the
chaos
and
Compass
live.
So
once
there's
one
any
update
something
like
this.
In
this
case,
we
removed
one
metric
from
the
matrix
model.
C
C
Other
changes
might
not
be
so
obvious,
like
we
might
change
the
definition
of
a
metric,
that's
used
within
a
metric
model
and
then
how
does
compass
know
that
they
need
to
change
that
that
metrics
model
with
an
updated
definition
of
a
metric
I
mean
it's
a
bigger
problem,
even
than
just
like
the
you
know,
the
pull
requests
to
you
know
to
this
working
group
to
the
metrics
models
working
group.
It
could
also
Compass.
D
Yeah
I
mean
I
I,
noticed
like
to
Sharon
will
be
demo
a
new
sus
service
today
I,
and
we
made
my
the
similar
question
problem
after
after
we,
we
are
trying
to
deploy
new
matrix
model
on
the
exist
matrix
model
whenever
they
would
be
revised
in
time.
So
so
there's
one
solution
that
we
at
our.
D
B
B
However,
we
could
create
a
reference
point
for
any
tool
to
say
this
is
the
version
of
metrics
that
we're
using
simply
by
doing
GitHub
releases
on
the
repo
or
the
metric
is
contained
periodically,
and
then
just
say
this
is
part
of
you
know,
repository
release,
1.1
or
whatever,
and
that
covers
everything
at
the
repo
level.
That's
how
software
is
released,
at
least
then.
B
If
there's
a
difference
between
what
people
see
in
the
metrics
model
on
our
website
and
the
metrics
model
in
the
tool
they
can,
they
can
line
up
the
versions
of
the
repositories
in
some
way
and
that
that
keeps
us
from
having
to
do
all
that
granular
maintenance.
If
the
metric
or
metric
model
level
and
does
give
us
a
reference
point
for
people
to
have
some
trust
in
what
they're
seeing.
A
D
F
C
What's
changed
and
I
think
the
problem
is
not
knowing
when
something
when
something
changes
but
I
think
that
there
are.
There
are
better
automated
ways
to
do
this
right.
I
mean
I'm,
not
sure,
I'm,
not
sure
exactly
what
happens
within
WordPress
when
it
pulls
in
a
GitHub
markdown
file.
That's
changed
whether
there's
something
we
could
trigger
on
that.
We
could
also
I
mean
we
could
use
something
like
like
GitHub
actions
on
the
metrics
reposed
to
to
take
some
action.
When
you
know
when
a
pull
request
is
merged,
when
a
metric
is
changed,
yeah.
A
C
D
Yeah
actually
Shane
just
remember
me,
remember
me
that
actually
it's
a
metrics
and
Matrix
Matrix
model
we
had
in
in
chaos.
Actually
we
if
we
traded
that
trading
as
the
software
they
they
should
have
a
Version
Control.
Actually,
so,
when
we
release
a
new
metric
matrix
model,
which
is
like
so
you
know,
version
1.0
and
we
should
we
could
have,
we
could
have
update
the
end
time
after
that,
so
any
service
depend
on
this
Matrix
or
matrix
model.
D
As
a
software,
we
should
know
that
the
version
will
depend
on
it's
version,
one
or
version
two.
D
So
I
mean
yeah,
I'm
thinking
the
question.
We
had
it's
a
problem,
we
have.
Should
we
introduce
the
motion,
control
on
actual
single
matrix
model
or
Matrix
and
The
Matrix.
A
I
definitely
cannot
type
in
talk
at
the
same
time.
So
maybe
the
first
issue
Don
to
your
point
of
this
whole
thing
like
having
the
Google
the
link
to
the
Google
Docs
in
each
of
the
things,
along
with
the
link
to
the
GitHub
metric.
You
know
like
what
we
have
here
and
agree
that
these
will
diverge.
Would
it
make
sense
that,
even
just
in
this
tracking
spreadsheet,
to
start
once
a
metric
is
released,
you
remove
this
foreign.
C
A
A
A
Okay,
I
think
that's
completely
Fair
issue
two,
which
was
how
to
inform
Downstream
software
Solutions
that
something
has
been
updated.
Elizabeth
do
you
know
if
there's
anything
in
WordPress
that
could
serve
as
a
trigger
that
would
send
anything
to
anybody
like.
F
I
think
it
might
be
better
to
do
it
in
GitHub,
through,
like
notifications
or
like
Don,
said
some
some
kind
of
action
that
would
that
would
notify
folks,
just
because,
like
WordPress
I,
don't
know
how
like
auger
or
Compass
would
connect
to
Wordpress.
I
guess
is
my
question.
I
feel
like
they
would
connect
better
to
GitHub.
Somehow.
C
B
F
Yeah
and
we've
been
having
some
issues
just
to
be
perfectly
honest:
we've
been
having
some
confusion
about
caching
in
WordPress.
Also,
yesterday,
I
was
I
was
putting
some
PRS
through
and
it
did
not
update
the
WordPress
page
automatically
and
then
I
did
another
one,
and
some
I
did
four
files
and
some
of
the
files
updated
automatically
and
some
of
them
didn't
so
I
had
to
manually
go
in
and
Trigger
the
update
on
some.
So
that
piece
is
a
little
wonky
right
now,
I,
don't
I,
don't
know.
E
F
E
C
A
B
E
B
A
Like
that
might
be
the
most
sensible
thing
that
given
a
change
to
a
metric
or
a
metric
model,
if
we
can
localize
it
down
to
that
level,
like
just
say,
if
there's
a
change
in
this
folder
I,
don't
want
to
check
on
actions
that
post
an
issue
in
the
OSS
Compass,
some
repository
there,
as
well
as
in
auger,
as
well
as
wherever
else
we'd
want
to
place
an
issue,
and
the
issue
could
just
simply
be.
This
has
changed
with
a
link
to
the
to
the
metric
or
the
metric
model.
We.
C
Maybe
do
we
have
I,
don't
know
if
it's
anybody
on
this
call,
but
does
somebody
want
to
I
feel
like
this
is
something
we
should
take
offline,
so
we
can
get
to
the
rest
of
the
agenda.
C
B
Is
is
there
a
web
I
think
the
website,
like
I'm
I'm,
actually
sitting
here,
looking
through
the
different
plugins
and
it's
unclear
to
me,
which
plugin
is
being
used
to
bring
things
from
GitHub
I
think
it
might
be
the
just
GitHub
shortcode,
no.
F
It's
a
special
one
that
Linux
foundation
made
just
for
us,
so
there's
very
little
documentation
on
how
it
works
and
it's
a
little
bit
wonky.
But
it's
cool
like
when
it
works.
It's
awesome
because
it
does
pull
that
GitHub
straight
through,
but
yeah
there's
like
zero
documentation
on
it.
Do.
F
F
F
E
B
And
my
issue
would
be
nice,
it
would
be
time
like
I
I,
know,
Kevin
and
I
spent
probably
a
day
dealing
with
this
Linux
Foundation
plug-in
Elizabeth
mentioned
when
we
moved
the
website
or
we
did
the
website
so
I
I,
don't.
B
This
month,
over
the
summer,
at
some.
D
Point
yeah,
maybe
I
I,
also
don't
have
time
this
month,
but
I
I
can't
do
next
month.
Actually
yeah
like.
C
This
one,
you
know
what
this
isn't
urgent.
Why
don't
we
just
file
this
as
an
issue
and
either
the
common
or
the
metrics
model
working
group
and
see
if
somebody
has
time
to
pick
it
up
and
then,
if
not
one
of
us
can
maybe
get
to
it
in
a
month
or
two
over
the
summer,
when
we've
got
a
little
more
time,
because
it's
not
it's.
E
C
E
B
A
B
It's
it's
nominally.
It's
not
yet
fully
related
to
metric
models,
because
we
haven't
labeled
or
implemented
metric
models,
but
we
have
implemented
a
number
of
things
that
are
two
two,
a
couple
of
tech
companies,
the
things
they
want
to
see
right
away.
B
So
if
we
look
at
the
chaos
org,
you
can
see
commits
issues,
pull
requests
at
a
real
high
level
and
then
they
have
three
dash
plotly
Pages.
This
is
all
based
on
auger
data,
where
they
generate
Graphics
contributor
growth
by
engagement,
new
contributors
by
month.
Of
course,
always
during
a
demo,
it
goes
slow
for
me
issue.
Staleness
issues
over
time
commits
over
time.
I
think
yahui,
where
this
is
not
as
well
architected
as
compasses.
B
C
B
So
it's
in
it's
in
much
it's
an
earlier
Alpha
type
of
release,
and
eventually
all
of
this
will
come
up,
but
it
is
not
coming
up
for
whatever
reason
right
now,
Beauty
women.
B
Well,
we
don't
need
to
waste
time
on
this,
but
you
can
create
a
user
and
if
you
click
manage
groups,
it
opens
up
a
tab
to
Auger
and
you
can
create
a
new
group
and
you
can
put
in
any
org
any
org
repo
combo,
full
URL,
and
if
the
data
is
already
in
the
database,
it'll
just
give
you
that
data
right
away.
If
it's
not
it'll
cue,
the
data
for
collection,
so
probably
the
most
useful
part
of
it
is
how
easy
it
is
to
just
add
repos
to
it.
B
I
just
wanted
to
point
it
out,
because
I
did
open
for
all
of
our
released
metrics
models.
I
I
have
issues
open
on.
This
is
a
tool
called
oss-as
aspen8not
that
reads
auger
data
and
for
each
tool
for
each
metric
model
that
we've
released.
There's
an
open
issue
to
create
the
metric
model
here
as
well,
and
we
did
Matt
so
separate
from
this-
not
sure
why
I'm
going
slow
this
morning
but
I'm
sure
it's
because
I'm
loading,
a
ton
of
data,
so
I'll
stop
sharing
and
Status
time
when
the
data
comes
back.
B
They've
done
yeah
yeah
see.
My
intention
is
to
deploy
metrics
models,
I
think
from
the
perspective
of
red
hat
and
a
few
other
companies.
B
The
purpose
is
to
deploy
the
metrics
that
they
think
are
important
to
look
at
together
and
those
don't
necessarily
those
don't
precisely
correspond
with
what
we
have
defined
as
metrics
models.
But
the
origins
of
our
metrics
models
are
very
similar.
B
There
are
groups
of
metrics
deployed
together
for
a
purpose
and,
like
I,
said,
I've
opened
eight
issues
to
create
the
metrics
models
that
we've
released
in
this
tool
and
to
try
to
make
just
like
with
compass
to
try
to
make
some
things
available
for
people
that
they
can
experiment
with,
but
obviously
it's
it's
not
as
far
along
as
compass
in
terms
of
the
performance
engineering
right
now
so.
B
I
think
so
what
I
need
to
do
is
a
sort
of
a
comparison
between
which
things
are
on
which
pages
and
what
things
are
in
what
metrics
models
and
see?
Okay,
if
there's
so
perhaps
just
a
few
things
missing
to
make
it
an
implementation
of
a
metrics
modeled,
but
I
haven't
done
that
analysis.
Our
goal
was
to
have
this
up
for
chaoscon,
which
we
did
and
then
Red
Hat
has
their
Summit
starting
today
and
so
they're
demoing
their
version
of
it
there.
A
Okay
questions
comments
for
Sean
on
this.
A
This
does
raise
just
kind
of
one
ongoing
thing
that
I
think
about
a
lot
which
is
like
we
have
tools
sometimes
or
is
gay
Eric's
still
on,
like
you
have
your,
you
know
how
you
have
your
like
metric
stuff
of
the
month.
You
know
what
I
mean
or
like
models
like
I'm,
not
quite
sure
how
we,
how
we
connect
or
like
the
dashboards
from
grimora
Lab,
how
we
connect,
what's
been
developed
in
software,
to
to
kind
of
a
written
specification
that
we
track
here.
A
B
We've
I
think
we've
laid
the
ground
so
at
the
very
beginning
of
chaos,
project
auger
actually
linked
back
to
the
metrics
that
were
affiliated
with
our
API
endpoints.
Then
we
just
stopped
doing
that,
because
the
location
of
the
metric
kept
changing
in
the
GitHub
repository
around
the
website.
But
now
we
have
stable
URLs
for
each
of
the
metrics.
B
So
now
I
think
the
tools
are
able
to
to
provide
that
chaos
reference
again
and
that's
a
pretty
recent
development
that
we
have,
that
consistent
metric
and
so
I
think
in
a
perfect
world
than
every
project
would
use
the
the
consistent
permanent
URL
to
indicate
that
this
is
an
implementation
of
this
chaos
metric
model
or
this
chaos
metric
that
that
would
be
the
road
map.
And
you
know
we
only
had
those
stable
URLs
for
maybe
six
months.
E
The
conversation
for
taking
the
the
work
that
is
done
in
the
implementation
and
the
software
back
into
the
specification
I
know
we
have
the
conversation
on
slack
for
the
risk
assessment
framework
blog
post,
that
I
authored
and
my
hope
right
now
is
that
I'll
take
some
time
off
next
week
and
then
I
would
be
happy
to
start
the
draft
to
get
that
in
okay.
A
E
A
That
connection
right,
that
that
I
was
talking
about,
we
don't
have
that
reference
point
at
all,
but
yeah
that'd
be
great.
B
A
If
we
don't
have
like
documented
here
as
a
reference
point
so
like
how
because
like
even
those
graph
that
you
were
showing
in
eight,
not
Aug
or
eight
knot,
yeah,
both
those
are
a
little
bit
different
than
what
we
have
here
documents,
and
so
how
do
we
get
like
what
you
have
deployed
in
in
your
software
environment
to
be
documented?
Here?
As
a
reference
point,
that's
the
question.
There
seems
to
be
some
problem
and
again
it's
not
like
I
think
about
this,
like
how
do
we
improve
that
process?
B
I
think
I
think
most
of
the
most
of
the
graphs
do
have
a
chaos
metric
behind
them.
I,
don't
think
the
pages
have
a
metric
model
behind
them.
B
B
D
B
So
I
mean
I,
think
I
think
this
is
almost
on
the
scale
of
an
initiative
to
systematically
support.
Try
like
for
each
project
to
prove
like
I,
think
actually
Compass
does
the
best
job
of
linking
the
chaos
information
directly
to
what's
on
the
page
and
I
think
for
more
lab
and
auger
and
ain't.
B
Not
it's
an
initiative
to
you
know
for
somebody
that
works
on
each
of
those
projects
to
say
here's
how
you
identify
whether
or
not
there's
a
metric
and
here's
how
you
include
the
link
so
that
metric,
if
it
exists
and
here's
how
you
propose
a
new
metric
if
it's
on
the
page,
but
it
doesn't
exist
right
and
you
know
we
provide
a
process
for
each
tool
and
engage
the
community
in
in
making
those
additions.
B
A
C
When
I,
if.
A
C
Or,
if
nothing
else,
you
know
put
exactly
how
it's
been
calculated
within
the
tool.
That's
that's
using
it.
So.
B
Yeah
I'm
giving
her
the
list
of
of
metrics
in
auger,
but,
like
we
don't
have
I
can't
we
don't
have
any
reference
points
to
where
in
auger
it
exists
like
on
the
page.
B
F
B
Yeah
and
we've
been
we've
been
corresponding
non-slack,
but
an
email
and
I'm
gonna
get
her
that
list
this
morning.
A
A
I,
like
this
I,
like
the
idea
of
an
mentorship
opportunity
too,
to
help
kind
of
structure
that
work,
so,
okay,
Compass
lab
metrics
components.
Did
you
put
this
here.
D
Yeah,
it's
me:
okay,
go
ahead!
Actually,
I
I
have
one
one
more
question
to
share
that.
Have
you
ever
talked
to
students
in
in
European
countries
or
in
USA
that
they
are
willing
to
help
us
to
to
doing
the
auger
implementation,
algorithm
level,
implementation.
B
D
I
mean
I
mean
I,
actually
have
a
conversation
with
the
students
in
in
China
from
different
universities.
D
Actually,
they
do
not
have
so
much
show
so
much
interest
on
on
helping
us
to
to
Implement
those
metrics
directly.
They
are
more
care
about
the
data
analytics
after
you
know,
week
out
the
week
out
the
refund
the
data
instead
of
to
to
implement
how
to
fetch
the
data
and
clean
the
data.
That's
the
big
issue
for
me
to
implement
Compass
to
Trend
when
I'm
trying
to
get
support
from
those
students.
What.
B
Yeah
I
mean
I,
I,
I'll,
I'll,
say
agars,
never
had
a
beautiful
front
end
and
at
chaos
kind
of
characterized
it
more
as
a
data
engineering
tool
which
is
really
what
it
is,
because
the
work
that
we've
accomplished
is
every
time.
There's
a
new
data
anomaly
that
comes
from
a
platform
or
a
git
log
we
handle
it.
B
You
know
we
put,
we
do
a
pull
request
and
we
fix
it
and
you'd
be
surprised
at
how
often
there
are
new
data
anomalies
or
subtle
changes
to
the
types
of
data
returned
by
apis
that
need
to
be
addressed
so
and
I
think
grimoir
lab.
Similarly,
for
the
way
they
collect,
data
with
Percival
has
to
deal
with
that
routinely
as
well.
So
I
do
think.
We
have
some
really
good
data
engineering
tools
that
they
could
use
and
perhaps
help
to
improve
compass
in
that
in
that
way,
yeah
is.
D
A
Is
it
students
that
have
an
interest
in
asking
questions
against
the
data
that
has
been
collected
so
less
about
like
actually
collecting
the
data
but
asking
say
protecting
questions?
Okay,
so
I
mean
as
part
of
it.
You
know
we
are
going
to
be
moving
towards.
You
know
kind
of
a
data
science
approach,
probably
a
little
bit
later
this
summer,
and
that's
probably
something
we
could
think
about
as
to
how
to
include
people
that
may
have
questions
against
the
data
that
is
collected
and
how
to
do
that
effectively.
A
So
I
think
probably
more
to
come
on
that
in
the
next
couple
months
and
and
these
questions
would
be
I
mean
I.
Think
they'd
potentially
be
you
know,
platform
agnostic.
You
know
like
whether
it's
on
Compass
or
auger
or
more
lab.
You
know
how
we
think
about
the
data
that
is
generated
from
them.
I
think
the
question
still
might
be
the
same
and
I
think
that'd
be
very
interesting
and
I
think
that's
something
we're
very
interested
in
doing
in
the
chaos
project,
starting
a
little
bit
later.
This
summer,
I.
A
Okay,
great,
so
that's
not
this!
So
do
you
want
to
talk
about
this
a
little
bit.
D
D
A
little
bit
more
to
remind
you,
I
I'm,
not
sure
if
you
remember
it,
but
but
I
I
would
quickly
go
through
the
whole
instruction
of
the
compass
lab.
No
okay,
here's
the
thing
the
this
the
compass
lab.
If
we
we
already
done
for
the
starter
project,
Health
matrix
model
and
but
the
way
we
collaborate
with
the
between
the
chaos
and
and
the
compact
is
that
after
this
com,
matrix
model,
D1
definition
is
completed
in
in
chaos
and
the
members
from
Compass
would
start.
D
The
validation
and
the
implementation
in
the
compass
I
mean
under
the
compass
lab,
but
I'm
thinking
that
if
we
could
implement
or
discuss
a
new
matrix
model
and
mean
mean
times
like
we
can
create
bioset
of
student
discussion.
That
would
be
great,
so
I
mentioned
here.
A
little
bit
create
your
own
models
in
five
minutes.
Yep
select
the
data
site.
We
provide
and
select
Matrix.
We
we
had
that
support
metrics
and
under
under
matrix
model.
Sorry
actually
I
think
we
already
have
this
this
one.
D
Yep
select,
Matrix
five
100
Matrix,
originated
from
chaos
and
select
the
algorithm
and
publish
models,
and
that's
that's
our
thinkings
create
this
matrix
model.
So
the
precisely
diagram-
and
you
can
name
your
model
here
and
the
select
the
data
site-
will
provide
some
data
sites
from
different
categories
and
select
the
Matrix
we
already
had
and
Implement
in
compass
and
originated
from
chaos,
and
actually
that's
the
topic
I
want
to
discuss
today
and
how
to
implement
how
to
how
to
add
some
priority
out
there
as
different
priorities
and
then
select
the
algorithm.
D
So
here
is
the
The
Next
Step,
but
you
know
how
to
add
a
data
site.
You
want
to
choose
for
a
new
Matchbox
model
and
and
in
in
this
matrix
model,
which
Matrix
you
would
like
to
use.
You
can
select
one
of
the
or
three
or
five
or
10
yeah.
That's
the
data
sites
and
unselect
the
metrics
and
select
up
to
10,
metrics
or
more
and
then
select
the
algorithm
you
can.
You
can
have
some.
We
can
provide
some
default
algorithm
or
you
can.
D
You
can
Define
it
by
yourself
and
after
that
you
can
let
people
to
invite
people
to
working
together
with
you
on
this
new
matrix
model
creation
and
and
actually
I'm
start
doing
so
from
sniper
by
step
by
spice
type
and
the
data
size
really
high.
Actually,
so
that's
not
a
problem.
The
thing
is
that
to
be
you
know
in
case
we
have
many
many
metrics
that
already
defined
in
chaos
and
I've.
Only
I
I
want
to
pre-ordalized
those
metrics
which
should
be
deployed
or
implemented
at
first.
So.
D
Yeah,
this
was
I,
think
it's
created
by
by
much
right,
so
we
have
matrix
model
and
each
metrics
model
have
different
Matrix
and
I
copy
this
sheet,
and
here
we
have
the
matrix
model
in
chaos
and
also
deployed
in
in
compass.
At
the
same
time
as
first
the
column
I
default,
I
listed
The
Matrix
we
already
supported
in
the
compass
and
here
is
the
link
I'm,
showing
sorry.
D
This
the
metric
is,
is
mapping
to
this
the
matrix
definition
in
chaos.
This
is
the
deployment
I
am
showing
here
it's
using
the
kubernetes.
D
Yeah
this
one
so.
D
B
I
mentioned
here
so
that,
from
a
software
engineering
perspective,
that's
kind
of
a
brilliant
way
to
do
it,
because
then
you
have
a
component
for
each
metric
and
if
we're
using
the
metric
in
multiple
Matrix
models,
you
just
drop
that
component
on
the
metric
model,
page
right.
D
Yeah
all
right
so
I
mentioned,
because
we
have
so
many
metrics
already
I
wanted
chaos.
People
help
help
me
to
define
the
prayer
priority,
which
one
should
be
implemented.
First,
as
a
compass
lab
component.
Then
they
can.
They
can
start
implementation,
those
Matrix
components
in
compass.
After
that
people
could
could
select
the
matrix
model.
Sorry,
Matrix
components.
D
B
So
this
I'll
just
say
my
first.
My
first
answer
would
be
identify
any
metrics
that
are
part
of
more
than
one
metric
model
and
the
metric
with
the
highest
end
of
participating
in
metrics
models.
I
would
put
those
the
highest
yeah
sure
yeah,
but
it's
yeah
it's
it's.
Obviously
you
know
I
mean.
Obviously
this
is
one
of
the
tracking
it
spreadsheets
are
great
answering
that
question
spreadsheets
make
it
happen,
it's
not
as
easy
to
figure
it
out.
B
A
From
my
perspective,
I
agree
with
Sean
that
any
metric
that
is
appearing
in
more
than
one
metric
model
seems
to
be
a
good
candidate.
I
would
also
just
say
kind
of
from
experience.
It
does
seem
like
a
lot
of
the
metrics
that
were
released
in
the
evolution.
Working
group
are
kind
of
very
Atomic
metrics
generally
and
might
be
good
candidates
for
for
this,
and
there
are
things
we
end
up
talking
about
a
lot
pretty
often
as
they
seem
to
to
be
pretty
Broad,
and
then
the
second
working
group
is
common.
A
B
So,
for
example,
once
you
build
your
first
issue
or
pull
requester
commit
component
there's
there
may
be
like
sub
components
or
core
Library
pieces
that
can
go
into
any
component
related
to
a
PR,
an
issue
or
a
commit,
because
these
are
core
objects
and
be
reused
and
provided
different
parameters
in
the
component
for
the
specific
metric
about
those
things.
D
A
D
A
A
A
One
was
around
project
influence,
so
we'll
pick
that
up
next
time
or
I'll
put
it
in
the
in
the
slack
Channel
and
then
likewise
I
do
think
we
do
I
think
we
have
a
couple
metrics
models
looking
at
them
today
that
are
done,
for
all
intents
and
purposes,
so
it'd
be
great
to
get
some
closure
on
some
of
those
as
well
as
we
continue
to
not
only
develop
software
around
these
models,
but
the
models
themselves.
So
anyway,.