►
From YouTube: Monitor:APM 12.4 Issue Overview
Description
Discuss the planned issues for the Monitor:APM team for the 12.4 milestone.
A
Okay
yeah,
so
this
meeting
is
to
talk
about
the
12.4
issues
for
the
APM
team.
We
are
just
gonna.
We
have
about
20
issues,
we're
just
gonna
kind
of
walk
through
them
quickly
because
we
don't
have
a
lot
of
time.
So
just
can
briefly
walk
through
each
one.
Just
give
a
little
high-level
introduction
to
each
issue,
there's
questions
or
details.
We
can
kind
of
follow
up
afterwards.
There
are
detailed
questions,
but
if
there's
questions
we'll
try
to
answer
them
right
here,
I'm
gonna
be
uploading
this
to
YouTube.
A
So
there's
a
couple
security
related
issues,
I'm
gonna
just
skip
over
for
now
and
then
I
think
that's
about
it.
The
only
other
thing
is
I
I'll
share.
My
screen
here
in
a
second
I,
also
labeled
everything
we
added
new
labels
for
their
like
category
metrics,
category
logging
or
tracing.
So
there's
one
of
those
three
labels
I
think
on
all
the
issues
412
for
now.
So
just
another
way
for
product
and
engineering
to
coordinate
is
and
see
what
we're
working
on.
B
A
A
C
A
D
Seconds
yeah
sure
I
think
that
all
is
required
here
is
going
to
be
to
come
up
with
a
chart
with
a
set
of
values.
Why
male
for
the
the
standard
get?
Let
are
elasticsearch
chart
2
for
this
thing
to
be
very
useful.
It
needs
to
have
a
persistent
value
claim,
so
there's
some
way
to
just
store
persisted
values
and
for
the
logging
in
the
rip,
because
normally
we
don't
have
any
persistent
body
claims
in
any
of
the
other
charts
we
have
so
they're
just
kind
of
ephemeral
or
the
logging
doesn't
need
to
be.
D
Maybe
is
like
for
me
to
use
and
stuff
so
this
will
so
I
did
a
little
bit
of
work
on
some
of
the
helm,
charts
I
made
of
mr2
the
helm,
charts
that
needed
to
be
changed,
but
long
I
think
it
was
gave
me
some
really
good
feedback
on
that.
So
I
think
that
this
can
certainly
be
a
ok
here's,
here's!
What
we
think
needs
to
be
the
the
values
for
this
and
how
it
goes
into
the
to
the
to
the
repo
and
then
give
you
configure
involved
to
make
sure
that
everything's
copasetic
sounds.
A
Sounds
good,
alright!
Moving
on
the
next
one
is
another
one
around
logging.
This
is
to
kind
of
dig
into
how
we
can
just
kind
of
build
on
that
last
one,
but
we
can
feel
kind
of
start
aggregating
logs,
so
I
think
there's
a
lot
of
comments
and
history
in
here
that
we
can
we're
not
gonna
go
through
right
now,
but
I
think
there's
enough
to
start
working
on
this
we
might
have
to
once
somebody
picks
it
up.
Adrienne
said
he
was
gonna,
take
a
look
at
it
so
be
able
to
him
dig
in
and
see.
A
The
next
one
is
around
tracing
so
for
a
first
pass
where
we've
been
talking
about
just
exposing
kinda
jäger
tracing
in
an
iframe
within
the
gait
lab
product,
but
there's
some
security
concerns
there.
So
we
want
to
try
to
address
that.
So
that's
what
this
one
is:
there's
a
number
of
kind
of
some
details
about
some
of
the
security
concerns
about
pulling
an
iframe
in
and
what
it
exposes.
So
that's
what
this
one
is
and
to
solve.
A
That's
a
tracing
one,
pretty
much.
The
rest
I
think
are
around
metrics.
So
the
next
one
is
auto
detection,
her
namespace.
So
we
want
to
there's
a
I
think,
there's
already
a
setting
that
we
support
metrics
with
namespace
deployments,
but
we
don't
auto,
detect
the
map.
You
know
we
can
create
metric
that
we
don't
kind
of
detect
them
automatically.
I
think
is
the
problem
for
Auto
DevOps.
There's.
D
A
Perfect,
okay,
that
makes
sense.
Okay.
Next
one
is
another
Prometheus,
but
a
storage
class,
so
some
deployments
use
or
there's
different
types
of
persistent
volumes.
So
there's
you
know
we
kind
of
make.
We
currently
think
that
it's
gonna
be
a
certain
storage
class
signa
cloud
provider,
but
there's
always
there's
settings
in
Prometheus
that
you
can
change
to
your
storage
class
and
especially
as
you're
growing
I.
Think
that
often
will
change.
So
we
need
to
support
that,
so
we
can
also
pull
those
metrics
in
it's.
A
A
B
A
All
right
breezing
through
these,
so
this
is
another
bug,
so
this
is
actually
so
a
customer
reported
this.
It's
a
basically
there's
a
list
of
kind
of
concerns
bugs
around
metrics
and
how
they
are
working.
I
know,
I,
think
it's
about
a
month
ago,
age
really
went
through
and
found
a
number
of
related
issues
and
address
some
of
these
I
think
there
was
still
I
think
we'll
need
to
update
the
description
a
little
bit
to
talk
about
which
ones
have
related
issues
and
which
ones
will
be
addressed
in
this
issue,
but
I
think
so.
A
Each
of
these
bullet
points
is
basically
different,
bug
first
and
I.
Think
three
or
four
of
them
already
have
issues
created
so
we'll
link
to
those
and
I
think
there's,
maybe
one
or
two
that
maybe
we
can
address
with
this
issue,
but
they're
all
on
that
kind
of
same
general,
whole
theme
that
they're
not
metrics,
aren't
quite
working.
A
The
way
we'd
expect
those
to
be
a
great
one
if
we
can
kind
of
address
some
of
those
in
this
release
and
in
the
comments
here
there's
listing
of
there's
some
of
the
issues
that
are
linked
here
and
there's
some
more
details,
some
of
them,
okay.
Next
one.
This
is
a
big
one,
around
source,
controlled,
alerting
for
auto
dev,
ops,
so,
basically
being
able
to
add
a
custom
alert
manager,
config
file.
A
There's
I
think
there's
one
more
issue
that
is
related
to
this,
which
we'll
get
to
I'll
jump
ahead.
It's
this
one
happened
that
was
lucky
click,
so
there's
these
two,
it
I
know:
Reuben
I,
think
you
looked
into
it
a
little
bit,
I
think
they're,
related
I,
don't
know
if
they're
exactly
the
same
or
not.
E
A
One
where
we
want
to
load
I'm
not
fully
up
speed
on
this
one,
but
get
it.
Can
you
open
this
one
be
able
to
load
kind
of
test
data
into
environments,
so
he
had
an
idea
that
we
could
have
like
a
Prometheus
data
file
that
gets
loaded,
I,
don't
know
if
anyone
else
has
looked
into
this
at
all.
I
have
a
I.
C
A
D
Like
to
I'd
like
to
make
a
little
rails
app,
that
is
a
is
a
fake
that
does
some
fake
arm,
Prometheus
metrics,
so
that
you
just
hand
it
the
metrics
of
your
endpoint.
Have
it
read
in
all
of
the
different
metrics
and
give
you
an
ability
to
say
over
time,
make
this
increase
or
make
this
decrease
or
make
this
thing
go
whatever
I
think
it
would
be
fairly
very
useful
for
the
for
the
world,
but
certainly
be
useful.
I
know.
D
A
Yeah,
it's
discussion
about
that
yeah
we
might
need
to
for
this.
One
I
know
yeah.
You
had
some
good
comments
in
here
and
back
and
forth
with
Kenny
and
stuff.
So
we
might
need
to
think
about
for
this
one,
what's
kind
of
an
MVC
that
we
can
deliver
in
this
release
and
kind
of
break
it
down
a
little
bit.
So
this
one
might
need
a
little
bit
more
definition.
A
A
B
Yeah
I,
don't
I
obviously
commented
on
this.
Oh.
B
Yeah
this
looks
like
there
were
two
things
kind
of
getting
conflated
here.
If
I
read
correctly,
it's
about
you
know,
people
were
running
into
errors
on
the
dashboard
as
a
result
of
probably
as
a
result
of
invalid
queries,
but
it's
not
clear
to
the
user.
When
that
happens,
why
there's
no
data
showing
up
so
really
this?
B
We
just
need
to
clarify
here,
what's
happening
on
a
dashboard
when
we
see
no
data,
oh
yeah,
which
is
a
two
part
thing
I
mean
that's
gonna,
involve
front-end
and
back-end
here
back
in
to
determine
why
we
get
errors
and
passed
that
information
to
the
front-end,
and
then
we
can
show
more
specific
data
and
you
know
basically
alert
the
user
to
hey.
This
is
an
invalid
query
or
we
couldn't
connect
to
your
Prometheus
instance,
or
you
know
whatever
it
is.
That
gives
us
more
just
a
better
understanding
of
how
to
troubleshoot
our
dashboard
yeah.
A
Okay,
great
thanks,
okay,
so
this
one
is
basically
just
moving
where
the
Prometheus
integration
is
in
the
settings,
so
we're
I
think
it
is
just
as
simple
of
this
kind
of
moving
where
you
can
find
this,
but
so
which
I
think
is
mostly
just
a
front-end
thing.
But
if
I'm
wrong
on
that,
we
can
change
that
so
right
now
it's
you
go
to
settings
and
integrations
and
Prometheus,
and
they
we
want
to
move
it's
just
like
settings.
Operations
I
think
is
yeah
is
the
goal.
A
A
A
Basically,
we
lazy,
initialize
metrics
right
now
too,
so
the
first
time
metrics
start
coming
in
and
we'll
provide
data,
but
that
was
causing
some
issues
where,
if
they
can't
find
a
metric
they're
expecting
to
find
we
get
some
errors.
Basically,
we
want
to
initialize
all
these
to
for
this
specific
class
to
just
initialize
them
to
0.
A
So
then
there's
a
whole
list
in
this
comment
in
this
ticket
or
this
issue
about
which,
which
wants
to
actually
do,
but
then
that
way
when
the
server
starts,
we
won't
get
these
kind
of
errors
but
kind
of
missing,
missing,
metrics
or
missing
here
or
alerts
or
missing
data.
That
would
power
the
alert
okay.
A
A
The
next
one
is
another
documentation,
one,
but
just
a
little
bit
about
how
to
set
up
Prometheus
this
one's
kind
of
old.
So
we
should
probably
look
through
it
to
see
how
things
have
changed
in
the
last
two
years,
since
this
is
open,
I
know
we've
added
documentation
since
then,
but
it's
probably
a
good
time
to
look
at
it
again.
An
axel
I
think
is
planning
to
kind
of
dig
into
it
a
little
bit
this
milestone
for
us,
which
is
great
and
then
the
last
one
I
think
we
just
added
the
last
day
or
something.
A
But
basically
this
is
goes
along
with
another
bug
that
we
were
seeing
in
12.3,
but
basically
just
validating.
You
can
see
the
little
animation
here
validating
that
you've
configured
this
correctly
before,
because
right
now,
if
you
watch
it
again,
if
you
don't
select
an
operator
like
greater
than
or
less
than
or
equal,
you
know
it's
like
one
of
these,
then
you
just
get
a
error
messages
and
it
says
air
creating
alert.
You
don't
know
why.
So
if
we
can
do
some
front
end
validation
on
that
we'll
make
that
error
go
away.
A
D
Are
a
couple
of
them?
They
need
go
ahead,
so
there
are
a
couple
of
them
that
require
that
we're
saying
we
want
to
have
source,
controlled,
alerting
or
a
source
control
this
that
or
the
other
thing-
and
maybe
this
question
for
configuration-
or
maybe
it's
our
question,
but
we
need
to
come
up
with
a
kind
of
okay.
This
is
where
we're
gonna
store.
This
is
where
we're
gonna,
you
know,
store
all
these
configuration
files
and
how
we're
going
to
store
them
and
in
the
in
the
repository
who
were
controlling
this
stuff.
A
B
A
C
A
E
A
All
right,
okay,
okay,
so
maybe
we
should
flag
those
to
doe.
He
is
there's
a
holiday
for
him
today.
That's
why
he's
not
here?
Otherwise,
normally
he
would
be
at
this
meeting,
but
we
can,
let's
find
those
that
we
can
kind
of
flag
them
to
see
if
we
can
either
pull
those
into
a
closer
release,
and
we
before
this
milestone
before
he
went
through,
there
were
I
think
over
50
issues
and
now
we're
down
to
about
20.
A
So
he
had
to
do
a
lot
of
shuffling
things
around,
which
was
great
that
so
there's
things
that
we
want
to
reprioritize
I
think
we
just
need
to
kind
of
let
him
know
and
why
they're
needed
a
little
bit
more.
He
basically
just
had
to
rely
on
a
lot
of
the
description
and
his
best
understanding
at
a
high
level
of
to
prioritize
things
so,
but
yeah
so
I
think,
let's
just
grab
find
those
two
issues.
There's
I
think
right.