►
A
B
A
A
A
C
Hold
up
so
that's
fine.
I
was
a
bit
distracted.
This
morning,
I
tried
to
get
a
website
live
shelter
in
place
because
I
thought
like
if
it's
going
exponential
there's
some
of
the
stage
should
be
shutting
down
just
like
we
are
in
San
Francisco
yeah
just
looked
at
the
latest
data
from
Washington,
and
it
seems
it's
linear,
so
not
exponential.
So
it
seems
that
the
distancing
is
is
working
good
enough.
That
I'm
not
sure
I
want
to
stand
on
a
barricade
and
shout
about
something
yeah.
B
I
will
say:
Kansas
actually
proactively,
canceled
all
schools
through
the
end
of
the
year,
which
is
not
something
a
bunch
of
other
states
have
done,
but
in
like
an
abundance
of
caution
and
I.
Imagine
will
get
to
shut
down
shortly,
even
though
we
haven't
seen
that
kind
of
rate
of
transmission.
Here,
no.
B
Cool
so
I
think
we're
already
recording
I
I
guess
I'll
give
the
overview
I
did
this
post
hog,
walkthrough
and
I
think
it
highlighted
to
me
a
really
cool
gap
in
our
single
application
between,
like
we
say,
monitor
and
monitor,
is
typically
very
focused
on
application,
performance
and
availability,
but
post
august
kind
of
like
a
feedback
loop
for
your
product,
instrumentation,
something
we
struggle
with
at
gitlab
in
terms
of
having
a
separate
tool
that
instruments
our
application
from
where
we
create
and
develop
our
applications.
So
I
thought
I
was
really
interesting.
B
C
B
And
so,
when
you
pointed
that
out
totally
makes
sense
to
me,
it
would
be
a
place
where
we
could
immediately
dog
food
I
had
made
this
point
about
I.
Think
there's
a
little
bit
of
conflict
and
you'd
responded
here.
I,
don't
know
where
there
is
conflict,
and
let
me
voiceover
where
I
seen
this
conflict
and
I
was
actually
the
person
who
wrote
the
like
product
principle
in
here,
because
I
spotted
this
before,
but
one
of
the
things
that
seems
to
be
happening
in
as
we're
building
more
and
more
completeness
to
our
DevOps
tool.
B
There's
lots
of
times
where
we're
confronted
with
the
question
of.
Should
we
build
this
for
the
user.
Who's
already
found
a
solution
for
this
integrated
problem
and
add
it
in
or
for
the
user,
who
has
who
has
no
exposure
to
this
and
is
kind
of
starting,
fresh
and
I
have
proposed
that
we
start
with
the
person
who's
starting
fresh,
because
in
lots
of
these
cases,
especially
in
the
opposite.
C
B
And
the
example
that
I
give
is
kubernetes
management
right,
like
our
communities,
management
integration
is
relatively
simple
compared
to
someone
who's
been
doing
kubernetes
for
three
years,
but
for
the
fact
that
we
know
the
majority
of
the
market
is
still
yet
to
adopt
kubernetes.
It's
it's
very
useful
in
that
context,
and
so
that
was
what
I
think
in
I'm.
Applying
that
same
logic
in
the
post
hog
example,
Louis
tag
would
be
like
easy
to
install
one-click.
Install
get
you
started
with
product.
Instrumentation
seems
like
more
valuable
than
well.
B
C
Mean,
first
of
all,
when
you
apply
when
you
appeal
to
a
principal,
it's
really
important
to
call
the
principal,
for
example,
if
we
hire
someone
and
make
a
hiring
decision,
see
that
isn't
isn't
aligned
with
our
values,
that's
useless.
That
is
like
what
do
you
mean?
We
have
six
values
like
which
one
of
the
values.
So
when
I
read
your
comment,
I
was
like.
What
are
you
talking
about?
One
of
our
products
principles
is
that
we
have
a
single
application
in
your
you're,
proposing
to
have
a
separate
application.
B
B
C
I
think
make
sure
what
part
we're
talking
about
snow,
snowplow
JavaScript,
let's
call
it.
The
JavaScript
part
super
easy,
hopefully
snowplow
back-end
by
kafka
receiving
everything
else
super
heavy-duty
and
you
have
to
do
it
all
yourself.
So,
basically,
that's
not
part
of
snowplow.
You
just
have
to
make
your
own
so,
but
what
we're
talking
about
is
which
client
library
to
use
I'm
missing
a
client
and
JavaScript
I'm
using
them
interchangeably,
but
I,
think
the
client
library,
a
snowplow,
is
not
harder
than
front
post
hog.
B
C
You
get
the
back
end.
What
is
the
back
end?
It's
receiving,
storing
and
displaying
I
think
our
problem
is.
We
want
a
single
application,
so
the
displaying
we
can't
do
in
post
hoc
anyway,
needs
to
be
in
gift,
might
be,
I
want
a
single
application
and
the
storing
we
could
just
do
any
post
grass
snow
plowers
from
anything
you
can
start
in
Postgres
and
they
all
supported.
So
it's
just
that.
C
Do
we
use
post
hog
and
are
stuck
with
another
interface,
which
is
super
bad
and
now
we're
gonna
I
framed
that
and
everything
else,
and
it's
not
it's
not
that
fancy
of
a
displacement
and
display
anyway
vs.
do
we
receive
the
events
ourselves.
It
seems
pretty
trivial
to
me
to
to
received
events
at
a
low
volume
and
a
high
volume.
It's
you
get
super
complex
with
Kafka
and
everything
else.
Post
hog
is
not
super.
Complex
post
dog
doesn't
solve
that
problem
writer,
so
we
can
quickly
get
to
kind
of
parody
there.
Yeah.
B
And
I
think
you
know
in
my
mind,
I
can
see
it
going
either
way
in
terms
of
prioritizing
one
or
the
other,
and
the
dog
fooding
use
cases
like
is
eminently
more
valuable
to
me.
Then
the
other
principles
and
we
we
specifically
highlighted
in
our
product
principles
that
we
will
choose
dog
fruiting
initiatives,
even
if
that
means
not
doing
you
cloud
native
first
or
something
like
that,
because
we
get
much
better
feedback
and
can
drive
improvements
quicker,
so
I
can
get
behind
that.
That
makes
sense
to
me.
Yeah
I
think
it's.
C
B
Makes
sense
I'm
still
going
to
engage
with
the
post
hog
team
and,
if
they're
willing
to
community
contribute
the
manage
tab,
that's
great
I,
don't
think
we
would
be
honestly
we're
not
playing
I,
don't
know
if
not
here
but
I
doubt
dev
would
be
prioritizing
this
type
of
work
in
the
next
coming
releases.
But
that
makes
sense
to
me
that
if
we're
gonna
put
our
resources
on
it,
it
would
be
for
the
dogfooding
principles
and
effort.
But
we
can
leave
that
accepting
merge
request
on
post
Hogg
issue,
cool.