►
From YouTube: Monitor:APM Weekly Meeting - 2020-06-03
Description
Weekly meeting for the Monitor:APM team
A
B
All
right,
so
this
is
the
APM
weekly
meeting
for
June
3rd.
Let's
see
the
first
topic
was
a
carryover
from
last
time
it
was
a
documentation
card
sort
from
Nadia,
I,
don't
know,
there's
any
updates.
We
want
to
talk
about
this
I,
don't
know.
If
there's
more
that
we
wanted
to
talk
about,
we
only
have
briefly
touched
on
it.
Last
time,
I.
C
C
D
Ok
in
the
last
couple
of
comments
on
the
issue
I've
outlined
I've
tried
to
outline
where
we
stand.
This
issue
blocks
an
immense
amount
of
work
for
me,
because
the
Prometheus
page
I'm
just
going
to
say
that
and
leave
it
once
once
we
have
it
settled
on
what
the
URL
path
should
be.
I
can
start
moving
things,
but
until
we
settle
on
a
path
I,
don't
really
want
to
shift
things
around
only
to
move
them
again
shortly
thereafter.
D
B
D
D
Sorry
I
was
completely
blanking
right.
I
totally
did
need
to
mention
that
I'm
not
intending
on
taking
vacation
anytime
soon,
but
we
have
changed
things
around
in
the
documentation
team
so
that
we
have
a
defined
backup
person
instead
of
catch-as-catch-can
Evan
Reed
will
be
serving
as
my
backup
anytime
that
I'm
out
for
a
significant
period
of
time.
So
if,
for
some
reason,
I
disappear
I'll,
let
you
all
know
in
advance
and
he'll
handle
handle
things
while
I'm
out.
C
They
want
to
understand
where
we
at
where
we're
heading,
and
sometimes
you
know
there
are
some
of
the
team
members
do
assigning
to
a
specific
issue
and
don't
really
understand
well
fits
in
the
overall
picture.
So
I
wanted
to
take
a
couple
of
minutes
to
go
over
in
a
really
high-level
burst
of
an
eye
yeah.
C
C
C
Cool
okay,
so
this
is
really
really
high
level.
Okay,
obviously
the
devil
is
in
the
details
and
if
I
would
have
gone
to
each
and
every
section
and
go
over
all
the
issues,
it
will
probably
take
us
hours
and
also
since
we
iterate
some
of
the
ideas
that
we
have
are
not
well-defined.
We
maybe
have
like
a
couple
of
issues
that
we
want
to
take
on
in
the
future
and
then
obviously
we
will
get
feedback
in
and
either.
C
Now,
in
order
to
get
to
this
goal
and
I've
look
at
it
from
the
eyes
of
the
user,
the
user
needs
to
do
a
couple
of
key
things
and
I'm
going
going
to
describe
the
actions
the
users
are
doing,
but
it
doesn't
necessarily
work
that
we
need
to
do
and
I
can
talk
about
it
afterward,
but
to
start
with
user
need
to
get
the
data
in
and
what
we
need
when
I
listed
here
is
a
user
to
use
it.
I
would
like
to
collect
the
important
metrics
for
my
application
and
services.
C
There
are
some
vendors
that
already
have
like
predefined
metrics
that
you
can
collect
like
you're
new
you
install
an
agent
and
you
can
automatically
select.
I
won't,
do
I,
want
memory,
usage
and
so
on,
and
they
have
like
a
list
of
like
hundreds
of
metrics
that
you
can
pick
from
there's
like
one
example.
So
if
we
look
at
what
our
users
are
doing,
obviously
we
ask
them
to
install
Prometheus,
external
or
thermal
quality.
So
then
I
want
a.
We
asked
them
to
implement
the
application
if
they
would
like
to
collect
like
specific
metrics.
C
They
can
deploy
exporters
to
the
hand
chart
and
they
can
enable
they
should
enable
metrics
came
in
once
they
are
done
with
those
four
steps
they
get
the
data
in
okay.
Now
we
have
the
data
and
yeah
by
the
way,
I
forgot
to
tell
you
that
what
you
see
here
on
the
black,
the
black
circle,
those
are
different
researchers,
research
issues
that
me
and
Nadia
have
done.
I
think
all
of
them
are
complete
and
we
have
more
research.
C
Into
so
what
I'm
talking
about
is
based
on
interviews
that
we
did
with
users
and
obviously
you
can
click
on
the
research,
and
you
can
see
interviews
and
finding
and
to
some
extent
we
have
also
epics
hello.
Here
we
have
some
epics
that
we
we've
defined
and
obviously
ability
to
iterate
on
those.
So
this
is
this
is
the
first
first
phase
that
the
customer
is
doing
once
we
have
the
data
in
the
next
obvious
thing
is
we
want
to
see
the
data
and
the
user
I
like
to
see
them?
C
C
This
is
like
many
windows,
take
pride
on
the
amount
of
out-of-the-box
dashboard
that
you
can
get,
and
customers
really
like
that
because
they
don't
like
to
mess
around
with
configure
dashboard,
and
you
know
moving
things
around
if
they
can
get
like
whatever
they
need
in
a
default
dashboard,
and
it
doesn't
matter
if,
like
half
of
the
dashboard
is
not
working
and
half
is
walking,
it's
good
enough.
It's
good
enough
because
it
does
provide
value
and
I
think
also
give
them
confident
that
the
monitoring
system
is
working.
We
set
everything
up.
C
We
did
all
this
hard
work
and
now
I
can
see
something
come
in.
So,
like
feels
good,
when
you
set
up
your
monitoring
solution-
and
suddenly
you
see
like
colors
in
the
dashboard
gives
you
like
I,
guess
it's
a
mental
thing,
so
the
more
out-of-the-box
dashboard
we
can
provide
to
the
user
the
better.
Secondly,
is
all
the
data
visualization
and
manipulation.
C
It
was
like
the
main
focus,
and
the
next
thing
is
adding
metric
and
I've
added
like
adding,
add
metric
here
as
a
workflow,
because
in
one
of
our
research
research
that
we've
done
me
Nadia
realized
that
a
lot
of
our
users
actually
starts
with
the
metric
dashboard,
and
then
they
quickly
move
on
to
Havana
why
they
move
on
to
kavana,
because
it's
easy,
if
you
look
at
our
dashboard
and
you
want
to
do
something
simple
change-
it
rearrange
it
or
add
a
single
metric.
That
is
not
it's.
C
Not
it's
not
one
of
those
out-of-the-box
that
we
provide.
They
need
to.
First
of
all,
they
need
to
go
to
the
commentation,
because
there
is
no
easy
way,
even
in
the
UI.
We
have
like
the
eighth
item.
Hotel-Motel,
then
you
you
have
this
huge
form.
Okay,
what
do
I
do
and
it's
easier
for
the
user
to
simply
right-click
and
go
to
go
fauna,
but
we
bundleco
fauna
with
with
good
luck,
and
we
have
a
way
to
do
this.
C
We
have
a
way
to
set
up
this
integration,
so
user
already
did
this
part
and
when
the
user
go
to
the
dashboard,
you
realize
hey
I
can
go
to
the
fauna
and
what
we
believe
is
that
if
we
allow
our
users
to
easily
set
up
different
metrics
that
they've
already
collected,
then
we
will
probably
get
more
adoption.
This
is
again
based
on
the
research
that
we
did
so
add.
C
C
The
tone
like
create
an
empty,
no
decay,
the
dashboard
and
they
add
metric,
also
on
the
way
and
configure
a
little
thing,
which
is
that
the
next
thing,
but
basically
I,
see
those
two
workflow
go-go's
handing
that,
although
I
separate
those
two
because
I
think
the
add
metric
deserve
like
a
place
of
its
own
and
obviously
we
have
the
research
and
we
have
several
epics
and
it's
amazing
design
issue
that
Nadia
created
about
the
add
metrics
and
encourage
all
of
you
to
go.
Look
at
it
and
add
your
comments
and
pretty
sure
you
saw
it.
C
But
if
not
then
do
that
and
the
next
part
on
monitoring
adventure
is
logging.
Logging
is
important
because
matrix
is
not
enough.
My
twigs
provide
you
with
an
indicator
that
something
goes
on.
Sometimes
metal
can
help.
You
find
the
problem,
but
in
many
cases
we've
realized
the
users
need
to
go
to
some
sort
of
a
logging
system
and
today,
with
the
hope,
this
whole
move
to
micro
services,
architecture
and
kubernetes,
and
everything
is
moving.
It's
very
hard
to
do
that
manually,
practically
impossible.
C
The
old
days
when
you
have
like
a
little
application
and
a
big
model,
if
you
just
dive
into
like
SSH
into
a
single
server,
and
then
you
just
tell
the
log,
go
to
like
a
huge
log
that
went
on
and
on
and
on
today.
Logs
are
split,
of
course,
hundreds
of
ports
and
like
a
single
service,
can
call
like
multiple
boards
and
activate
multiple
multiple
services,
and
it's
particular
impossible
to
do
that.
So
this
is
why
the
log
aggregation
use
case
emerged
form.
C
So
you
have
some
sort
of
logging
capability
that
aggregates
on
your
log
and
present
them
in
the
filter.
By
you
know,
the
easiest
thing
is
to
filter
them
by
by
timestamp,
but
obviously
allow
the
user
to
conduct
sophisticated
search
and
filtering
capability.
This
is
what
we
are
doing
with
our
logging
solution
with
our
log
in
solution
in
the
long
Explorer.
The
user
first
needs
need
to
install
elasticsearch
and
we
do
it
as
a
manager
app.
So
it's
a
with
the
click
of
a
button.
C
The
user
can
deploy
elasticsearch
in
the
cluster
and
it
will
automatically
start
collecting
all
the
logs
that
are
coming
from
this
cluster
and
then
obviously,
those
logs
need
to
be
surfaces
in
the
UI.
This
is
why
we
have
the
roam
Explorer
and
we
need
to
provide
the
user
with
advanced,
search
and
filtering
capabilities,
because
the
amount
of
logs
that
are
generated
by
those
application,
sometimes
it
can
be
like
petabytes
of
logs
per
per
day,
and
it's
again
it's
impossible
to
scan
or
to
scare
them.
C
So
it's
very
easy
to
see
what,
like
things,
that
all
five
windows
have,
and
naturally
this
is
something
that
we
should
have
as
well.
So
it
was
very
helpful
doing
that
to
understand
what
are
our
gaps
along
with
customer
interviews
that
we've
done
is
we
mature
our
login
solution,
and
the
next
thing
that
users
would
like
to
do
obviously
is
to
set
up
alerts.
They
want
to
be
alerted
when
there
is
some
sort
of
a
violation
of
a
matter.
C
If
you
ask
someone
in
the
NOC
or
the
VP
of
hopes,
what
is
the
thing
that
he
hate
the
most
he
cut
off
God
and
he
surprised
getting
any
calls
from
customers
saying:
hey
your
applications,
not
working
they're,
going
to
be
alerted
when
things
go
wrong
before
someone
is
calling
them
and
getting
mad
at
them,
because
basically
they're
unable
to
do
things
in
the
application
and
to
that.
Obviously,
we
need
an
easy
way
to
set
up
an
alert
to
our
to
our
solution
and
define
some
sort
of
notification.
C
Sometimes
alert
does
not
involve
notification,
it
doesn't
go
like
it
doesn't
one.
Sometimes,
when,
when
you
trigger
an
alert
it
is,
it
is
doing
that
something
like
running
a
run,
some
sort
of
a
playbook
or
set
message
to
slack.
So
there
are
multiple
ways:
multiple
actions
that
is.
Do
you
know
that
this
is
more
like
the
healthy
responsibilities,
but
the
setting
up
of
the
alert.
This
is
something
that
is
critical
for
our
solutions
well
and
yeah.
And
lastly,
we
are
cloud
native.
C
We
want
to
make
sure
that
users
that
adopt
that
are
going
cloud
native
or
using
kubernetes
are
able
to
use
our
monitoring
solution,
which
means
that
they
need
specific
things,
for
instance
cluster
inside.
So
we
have
a
couple
of
things
where
we
can
provide
our
users
first
of
all
with
the
cluster
health,
and
we
have
an
issue
where
we
discuss
what
we
need
to
move
the
cluster
health
dashboard
that
we
have
on
the
cluster
setting,
which
is
really
buried
somewhere
to
our
dashboard,
because
it's
a
metric
job.
C
Basically,
so
this
is
one
one
thing
that
we
can
quickly
do
and
I
think
the
big
thing
that
is
needed
for
my
users
is,
if
you
are
familiar
with
the
infrastructure,
dashboard,
epic
and
think
about
like
a
birds
of
an
eye,
yeah
of
course,
you're
across
your
clusters,
you'll
be
easily
be
able
to
see
all
your
ports
and
their
health.
We
have
something
very
similar
with
the
environment
tab
where
you
can
actually
see
your
pods
but
think
about
something
like
that.
We
will.
C
C
That's
like
a
as
I
mentioned,
like
a
birth
of
an
eye
view
and
the
details
here,
metals,
and
today,
as
I
mentioned,
we
are
focusing
on
the
data
visualization
in
the
receive
symmetric
in
the
dashboard,
but
as
we
proceed
I
would
it
I
would
like
to
move
to
the
next
big
thing
with
which
would
probably
be
like
the
add
metric
and
then
obviously
we
can
iterate
and
see
where
we
need
to
go
next
either.
It
would
be
like
too
old,
like
getting
data
in
an
instrumentation
or
maybe
log
application.
C
It
really
depends
on
the
feedback
that
will
get
from
the
customers
and
the
trends
that
we
see
the
market
and
the
opportunities
that
we
have.
Obviously
we
are
I
am
I'm
measuring
all
of
the
work
that
we
are
doing
to
our
north-south
metric
and
with
that
I
can
move,
and
still
this
focus
and
I
think
this
is
more
or
less
what
I
wanted
to
cover.
And
if
you
have
questions
you
know
how
to
reach
me
and
yeah
I
mean
this
mule
is
is
open,
for
you
can
click
through
the
research
we
can
in
every
meetings.
C
The
weekly
meeting
that
we'll
have
we
can
go
over
a
specific
research
that
we
did.
You
think
me
Nadia
spoke
about
and
we
did
a
research
of
understanding
how
users
our
users
are
using
fire,
because
it's
very
related
to
the
work
that
we
do
here
today.
So
this
is
likely.
This
is
probably
the
next
research
that
we
will
talk
to
thing,
but
you
can
obviously
can
go.
You
can
click
you
can
read
through
their
insights,
yeah.
E
That's
a
little
buddies
with
sorry
go
ahead.
Yeah
yeah
I
am
every
question.
Firstly,
thank
you
for
sharing
the
setting.
This
gives
a
good
overall
picture
and
a
good
direction.
I
know
this
is
still
work
in
progress,
but
when
you
say
as
a
user,
so
is
there
a
target.
Persona
is
desiccated
lab
developer
or
is
an
expert
and
I'm
confused
about
this
a
little
bit
because,
right
now,
with
the
dogfighting
efforts,
yeah
target
audience
are
SRT
teams
who
are
expert
users.
E
C
You
all,
probably
next
week,
by
the
way,
two
more
arm
of
quantification
the
next
week,
I'll
probably
share
with
you,
which
exactly
speak
about
who,
who
is
our
target
user?
Today
we
focus
on
our
internal
s-sarah
team,
but
this
is
like
some
sort
of
I
said
like
there's
a
temple
Oracle
and
because
we
want
to
be
successful
with
with
dogfooding,
but
the
based
on
some
of
the
work
that
we've
done.
I
think
that
we
should
focus
first
on
on
developers,
I
mean,
of
course,
gitlab.
C
Whenever
we
start
with
a
new
project
or
weary
solution,
we
focus
on
developer
first
because
doesn't
mean
that
we
won't
beat
features
that
will
appeal
for
the
DevOps,
but
we
need
to
think
about
it
from
a
developer
point
of
view
and
in
this
MRI
I'm
actually
going
to
explain
why
why
those
are
the
target
persona,
so
it's
probably
be
like
based
on
our
personas.
You
know
handbook
the
developers
and
DevOps
engineers
which
are
similar
to
DevOps
2
to
12
to
us
it
will
be
our
own
histories,
but
there
is
there
is.
C
E
A
A
So
the
way
correct
me
if
I'm
wrong,
but
my
understanding
is
more
we're
trying
to
do
for
right
now
we're
trying
to
help
dogfooding
so
we're
trying
to
do
the
minimal
things
that
that
team
needs
to
get
in
order
for
them
to
get
running,
and
then
we're
gonna
focus
more
on
rounding
up.
The
C,
metrics
and
dashboard
section
that
gives
you
seeing
in
the
Bureau
is
that
incorrect.
C
I
mentioned,
focusing
he
only
hill
and
by
the
way,
I
said
dogfooding
and
more
because
the
footing
is,
is
a
partial
list
of
things
that
a
mojito
institutions
should
have
like,
for
instance,
drag
and
drop,
and
draggable
and
resize.
Your
your
dashboard
is
I
believe,
is
a
basic
capability,
any
monotone
solution
solution
you
should
have
so.
C
We
have
like
tools
like
solution,
validation
that
I'm
doing
together
with
with
Nadia
and
more,
and
we
can
set
up
our
own
goals.
You
can
set
up
our
own
goals.
Okay,
once
we
are
done
with
doing
X,
Y
&
Z,
let's
move
on
it,
doesn't
mean
that
we'll
stop
development.
I
mean
we
still
have,
as
I
mentioned,
customer
enhancement
that
they
will
ask
us
to
do
things
that
are
outside
of
this
area,
which
is
fine,
but
the
idea
is
that,
where
are
we
going
to
concentrate
most
of
our
effort?
C
A
C
C
Good
idea
and
I
was
thinking
about
adding
these
ones
are
done
with
two
today
metrics
mutually
fits
in
the
metric
Direction
page,
because
it's
a
good
question
where
and
maybe
in
the
APM
group,
maybe
because
I
want
to
add
to
the
metric
the
action
page.
This
thing,
and
probably
also
here
but
like
log
aggregation,
does
not
belong
in
the
metrics
section.
C
C
B
B
A
Yeah
I
realized
I,
don't
think
we
announced
this
on
the
call
yet,
but
we're
having
second
family
friends
and
family
day
on
June
12.
So
please
take
that
day
off
if
similar
to
the
first
one,
if
you
have
a
holiday
on
that
day
pick
a
day.
That's
around
that
time.
That
goal
is
to
encourage
us
to
have
better
work-life
balance
and
focus
on
friends
and
family.
It's
a
little
bit
more
effective
when
everyone
takes
off
because
then
you
don't
sign
on
and
everyone
else
is
working.