►
A
All
right
so
we're
talking
about
the
initial
retro
feedback
on
the
first
responder
pilot
program,
of
which
many
on
here
are
users
of
or
helped
create
and
are
working
on,
etc,
so
good
stuff.
So
I've
got
the
issue
in
the
meeting
notes.
I'll
also
put
it
in
the
chat
and
zoom
in
case
folks.
A
Don't
have
it
handy,
so
my
impression
is
overall,
it's
actually
gone
pretty
well
with
some
notable
exceptions,
so
the
parts
that
have
gone
well
in
kicking
this
up
is
that
it's,
it's
primarily
working
as
designed
where
people
in
the
slack
room
on
a
slack
channel,
are
getting
the
escalations
before
people
have
to
go
to
the
spreadsheet
and
that's
providing
a
better
work.
Life
balance
for
the
at
least
during
the
weekdays,
for
all
the
engineers
doing
support
during
the
week.
A
We
still,
of
course,
have
the
spreadsheet
for
the
weekends,
because
generally
nobody's
online
in
that
slack
channel
on
the
weekend,
so
we
still
have
that
for
the
weekend.
So
overall,
that's
going
well.
However,
a
couple
escalations
one
in
particular,
but
but
more
than
one,
took
a
really
long
time
to
find
somebody
to
respond
during
the
week
and
it
looks
like
the
cause
primarily.
A
Is
that
not
all
engineers
doing
doing
support
have
been
asked
to
nor
done
proper
slack
configuration,
so
the
bot
knows
if,
if
they're
available
and
online,
and
also
we
seem
to
be
the
bot,
let's
say
person
number
one
misses
the
escalation
right
and
then
a
day
later,
another
escalation
comes.
A
If
we
have
a
list
of
people
who
don't
have
slack
configured
correctly
or
don't
know
how
the
program
works
and
they're
not
responding
escalations,
every
subsequent
escalation
is
going
to
take
a
really
long
time
to
escalate
because
it's
going
to
the
people
that
are
not
fully
in
the
program
properly.
Yet
so,
there's
all
sorts
of
ways
to
solve
that
that
we've
thrown
around
so
anyway,
that's
my
impressions.
What
are
what
are
other
folks
in
question
so
far.
B
A
B
Your
observation
is,
one
conclusion
is
that
the
folks
are
didn't
set
correctly.
Is
that
the
one
thing
you
observe.
A
I
I
think
that
that's
the
first
part
of
it
that
slack
is
not
configured
because
we
actually
haven't
asked
them
to
right.
So
so
we
haven't,
or
if
we
have,
we
haven't
given
great
recommendations
to
them
and
we're
working
on
updating
the
handbook
for
that
and
reaching
out
to
them.
So
one
is
people
have
slack
configured
correctly
and
when
they're
online,
when
they're
working
etc.
A
The
other
is,
I
still
think
if,
let's
say
everybody
has
it
configured
properly,
which
is
not
the
case.
Currently,
let's
say
the
person
who
has
it
configured
properly
is
choosing
to
not
participate
in
the
program
right
where
they're
not
responding
the
bots
always
going
to
choose
them
first,
once
it
chooses
them
one
time,
the
first
time
it
chooses
them.
So,
let's
say
I'll
say
I'm
the
person
not
participating.
So
the
bot
chooses
me
I'm
the
least
recently
escalated
to.
I
ignore
it,
I'm
planning
to
ignore
it.
A
I
ignore
it
every
time
the
bot's
always
going
to
choose
me
first
for
every
escalation,
even
if
I've
slack
configured
correctly.
So
I
think
that's
a
challenge
as
well
many
different
ways
to
handle
that
to
address
that,
but
you.
D
C
Yeah
I
mean
to
me
like
that's,
definitely
something
that
we
can.
We
can
program
in
right
that
that-
and
I
guess
like
is-
is
this
a
voluntary
thing
to
me?
This
is
supposed
to
be
replacing
the
current
on-call
schedule
during
the
week.
So
it's
good.
C
Yeah
so,
and
and
like
with,
with
all
the
things
that
you
brought
up
to
me,
I
think
that
what
we've
realized
is
that
there
were
a
few
prerequisites
that
maybe
weren't
communicated
and
just
learnings,
for
maybe
this
next
iteration
that
would
have
helped
get
towards
a
better
success
rate.
I
think
the
success
rate's
been
pretty
good,
but
given
that
we
didn't
really
have
total
compliance
with
people,
updating
their
slack
notification
that
definitely
impacted
how
successful
this
could
be
initially.
C
But
I
mean
with
with
with
some
of
this
it's
hard,
because
we
need
to
try
it
in
order
to
get
shake
the
bugs
out
now
we're
shaking
the
bugs
out,
and
so
I
wonder
if
we
don't
need
to
set
back
the
pilot
like
another
month,
because
now
we've
we
kind
of
understand
the
prerequisites
and
we're
really
understanding
the
bugs.
A
I'm
not
I'm
not
in
favor
of
definitely
I'm
not
in
favor
of
actually
taking
a
step
backwards,
although
we
surely
can
and
like
ending
the
pilot
and
going
back
to
the
spreadsheet.
While
we
fix
things,
I
think
we
could
go
through
and
ask
the
managers
of
all
back-end
engineers
to
first
update
the
the
handbook
on
what
we
recommend
it,
how
they're
recommending
the
updates
their
slack
config
and
then
asked
each
engineering
manager
to
confirm
their
backend
engineers
have
done
so.
A
I
think
the
escalations
have
only
been
later
than
what
we
wanted,
maybe
two
out
of
10
or
something
like
that
which
is
going
to
be
better
if
it
was
zero,
but
the
other
the
majority.
It's
worked.
Just
fine.
B
All
right,
so
I
think
we
are
jumping
into
the
solutions
here,
but
I
think
we
need
to
be
clear
about
the
several
things.
First
is,
of
course,
the
tooling
thing
I
mean
everybody
should
should
set
up
their
configuration
right.
B
Thing
is
uncle
is
not
optional,
it's
it's
mandatory
for
all
engineers,
let's
be
clear,
then
I
think
the
solution
will
be
here.
The
two
problems,
one
is,
you
know,
select
settings
and
the
other
one
is
if
they
say
I'm
on
call.
I
never
responded
that
I
will
be
always
on
top
of
the
list.
I
don't
think
we
we
want
to
optimize
this
algorithm,
but
yeah.
A
And
exchange
an
example
for
you
is
like
just
use
the
example.
So
let's
say
let's
say
you,
you
have
a
tendency
to
miss
your
slack
notifications
right.
You've
set
up
slack
correctly,
but
you
miss
it.
Yeah
people
look
like
it
happens
right,
it
happens
to
all
of
us
and
let's
say
you
have
a
tendency
to
miss
it,
and
let's
say
you
get
escalated.
Three
different
support
issues
in
three
days,
you're
going
to
be
first
on
the
list,
even
though
you,
though,
you've
configured
slack
correctly.
That's
what
I
think.
Perhaps
we
should
not
pick.
A
The
least
least
recently
escalated
two
person,
but
maybe
take
the
let's
say
a
number.
Let's
say
five
of
the
least
recently
escalated
two
people
and
randomly
choose
someone
to
add
a
little
jitter
into
the
system
so
that
it's
not
always
choosing
that
one
person
now
one
person
will
still
get
escalated
to
again,
but
it
won't
be
that
person
every
time,
potentially
delaying
the
support
teams
getting
a
quick,
a
quick
escalation,
they're,
better
solutions
than
a
random
thing,
but
perhaps
we
that's
just
one
way
to
potentially
do
it.
D
Yeah,
I
think
I
think
we
need
to
think
about
about
this
a
bit
more.
They
should
be
the
random
things
that
you
will
end
up
with
the
top
five
being
those
who
do
not
respond.
I
know
so
we
don't
know
if
this
is.
If
this
is
really
a
problem
like
right
now,
most
likely,
they
don't
have
that
configured
and
that's
the
feedback.
We've
got
from
most
of
the
people
where
that
oh
yeah,
my
electrifications
weren't
configured
or
they
forgot
to
have
the
dnd
setting.
D
I
think
when
we,
when
we
do
this,
and
this
this
was
one
of
the
reasons
that
I
made
a
bot.
We
need
to
be
fair,
so
we
we
still
want
the
the
people
who
were
lesbian
to
be
there
to
be
first.
Basically.
D
So,
if
there's
a
particular
problem
with
someone
who
continually
ignores
that,
then
I
think
we
may
want
to
address
that,
and
it
could
be
automated.
As
I
mentioned
it,
maybe
pages
like
could
send
a
private
message
or
something
and
another
thing
now
that
I'm
talking
about.
A
Yes,
sir,
look
at
somebody's
slack
config,
and
is
it
able
to
pull
the
configuration
to
see
if
it's
configured
according
to
the
best
practices
that
we
want?
At
least
some
of
them
are,
if
not.
D
Yeah,
this
is
there's
a
way
to
say
whether
they
have
disabled
or
not,
not
enable
any
dnd
setting,
and
then
we
could
even
check
the
actual
settings
and
we
could
say
hey
if
I
think
the
default
is
12
hours
working
time
or
something
like
that.
We
can
check,
maybe
if
they
have
the
defaults
as
well.
So
there's
a
way
for
the
bot
to
check
to
check.
D
So
here's.
B
My
opinion
we
shouldn't
do
that
we
should
we
should.
We
shall
ask
all
the
managers
to
work
with
team
members
to
ensure
they
set
their
slack
correctly
and
for
those
people
you
know,
tend
to
miss
all
the
calls.
Then
we
shall
we
can
do.
We
can
detect
that
data
and
ping
the
person
and
their
manager
to
work
together
to
get
a
result,
get
it
resolved.
So
it's
a.
We
need
to
be
fair
for
all
the
people
and
be
honest.
B
The
goal
one
of
the
goals
to
set
up
the
uncle
is
to
spread
out
the
workload
to
more
people
in
engineering
department
award
in
development
department,
sub
department
department
yeah
because
it
used
to
be
just
a
few
people
responding
to
all
the
escalations
after
this
uncle
more
people
got
involved.
So
that's
one
problem
we
want
to
solve.
If
we
just,
if
we
shrink
back
to
a
volunteering
base,
it
will
shrink
back
to
a
small
subset
of
the
developers
again.
So
then
uncle
is
well.
We
don't
solve
the
problem.
We
wanted
to
solve.
A
So
perhaps
is
we
give
the
engineers
good
recommendations
on
how
to
configure
slack?
We
haven't
done
that
yet
right,
so
we
need
the
at
first,
so
some
already
had
it
set,
which
is
great,
but
we
haven't.
Actually
we
haven't
done
that
ask
them
to
do.
You
know
an
engineering
week
in
review
and
you
know,
and
maybe
a
couple
slack
chat,
a
slack
channel
or
two,
and
then
we
also
ask
engineering
managers
to
confirm
it.
So
the
engineers
don't
feel
like
it's.
You
know,
you
know,
oh,
you
didn't
do
something
right.
A
Well,
we
haven't,
we
haven't
even
asked
them
to
do
it
in
the
first
place,
so
we
asked
them.
Give
it
a
week
then
ask
the
engine
engineering
managers
with
back-end
engineers
to
go,
confirm
that
they
verified
it
and
just
those
two
things.
Those
things
may
so
not
changing
the
bot
at
all.
Just
doing
those
kind
of
procedural
things
might
help
a
lot
but
amy
and
michelle.
C
C
Yeah
I
mean
to
be
clear:
I'm
pretty
sure
that
we
we
did
ask,
but
you
know
it's
a
type
of
thing
where
it
takes
a
while
for
some
of
the
stuff
to
sink
in
as
well.
I
think
the
getting
injury
managers
to
confirm
that
people
have
done
it,
that
that's
the
step
that
we
missed
right
and
then
that
makes
sense
to.
D
We
also
fixed
a
couple
of
things
that
were
related
to
this,
like
being
the
the
original
ones
going
on
people
again
like
for
the
first
and
second
thing.
So
the
second
thing,
the
first
thing
person
will
will
appear
there
again
while
before
they
may
come
just
a
bit
later
like
two
minutes
later,
and
then
they
won't
see
that
and
they
wouldn't
know
what
was
happening
so
so
now
they
should
see
their
names
all
of
all
of
the
time.
Basically
so
that
may
help
as
well
with
this.
A
Agreed
amy
and
michelle:
what
do
you
think.
E
What
I
mean,
I
think,
yeah
it
all
sounds
really
sensible,
like
I
think,
definitely
feel
like
it
should
be
fair
and
people
should
get
the
chance
to
be
sort
of
told
to
set
things
up
correctly.
One
thing
that
we
that
has
been
kind
of
useful
from
my
side
is
knowing
we
have
the
spreadsheet
to
fall
back
on.
D
Yeah,
so
I
think
for
that
I
mean
this.
This
ball
is
really
a
prototype
and
it
can
be
configured
right
now.
We
have
six
people
because
we
have
we
give
enough
time
for
escalation,
like
maybe
that
takes
like
12
minutes,
and
then
we
still
got
a
few
minutes
for
finding
the
right
people
after
that.
So
we
may
want
to
increase
that
after
this
pilot
to
maybe
more
people-
and
the
other
thing
is
so
there's
more
things
that
we
can
do
there.
D
For
instance,
after
dinging,
say
10
people,
then
you
can
add
here,
for
instance,
nobody
replied
and
then
after
10
people
said,
let's
just
add
here
and
that
will
bring
all
of
the
online
engineers
that
are
in
in
the
in
the
channel.
It
will
be
making
other
people
as
well,
but
that
may
not
do
any
any
harm
if
we
reach
that
stage.
B
Can
we
actually
implement
that
at
here
to
replace
the
sixth
ping
now
and
see
how
that
works?
I'm
what
I
I
control,
whereas
amy
is
a
concern,
I'm
worried.
If,
if
we
turn
to
production
and
we
drop
the
spreadsheet
sign
up,
then
we
don't
have
a
way
to
ensure
there.
A
A
Yeah
and
we
said
already
we're
going
to
give
a
spreadsheet
through
september
and
october,
you
know
november
is
a
big
holiday
month
as
well
as
we
as
we
work
out
the
kinks.
With
this
you
know,
I
think
we
keep
the
spreadsheet
during
the
weekdays,
in
addition
to
the
weekends,
probably
through
november,
and
probably
through
december
too,
since
obviously
it's
a
big
holiday
month
as
well,
and
only
if
we
really
see
the
you
know
the,
and
then
we
reassess
getting
rid
of
the
spreadsheet
during
the
week.
Perhaps
in
january.
A
Well,
while
we
continue
to
iteratively
improve
things,
yep
yeah.
C
Why
can't
we
just
get
a
list
of
people's
working
hours
their
standard
working
hours?
You
know,
for
example,
you're
going
to
have
a
set
of
people
in
europe
and
america
and
aipac,
and
then
because
to
me
like,
like
you,
should
be
able
to
fall
back
to
a
sheet
of
the
people
that
you
would
kind
of
know
would
be
online
or
like
that
that
typically
work
during
that
period.
B
C
C
It's
new
it's
it's
supposed
to,
but
we
didn't,
we
haven't,
had
the
prerequisites.
I
feel
like
we're
kind
of
judging
this
from
the
standpoint
of
like,
but
it,
but
we
didn't
have
the
prerequisites
in
place
right
so
so
so
now
we
we've
had
some
incidents
where
we
haven't
had
responses,
but
we
actually
didn't
have
the
prerequisites
in
place.
So
I
don't
feel
like
we
can
make
a
real
clear,
like
judgment
of
it.
At
this
point.
B
You
said
we
have,
we
don't
have
prerequisite.
I
actually
in
the
process
documentation
of
this
pilot.
We
do
describe
how
to
how
to
set
the
select
settings
right.
C
B
Guys
exactly
this
imr
is
doing
ask
people
to
set
their.
I
mean
more
clear
instruction
how
to
set
slack
and
then
we
I
think
we
are
going
to
extend
our
pilot
program
and
basically
it's
a
reset.
We
discover
the
issues
we
make
enhancements.
Let's
continuously
pilot
till
the
end
of
the
year
and.
D
B
I
I
suggest
the
previous
suggestion:
all
the
managers
work
with
their
with
their
team
members
ensure
that
their
slide
is
set.
F
Correctly
so
one
thing
that
I
do
want
to
ask
in
terms
of
setting
up
your
slack
right:
do
we
need
a
solution
for
being
away
because
one
of
the
things
that
has
been
vocalized
to
me-
which
I
feel
like,
makes
a
lot
of
sense-
the
spreadsheet
today,
if
you're
on
call
and
you're
on
that
spreadsheet
you're,
I
don't
want
to
say
on
edge.
That's
not
the
right
term,
but
you're,
aware
you're,
hyper
aware
that
at
any
time
you
could
get
an
incident
with
the
first
responder
program.
F
D
I
mean
I
think
that
would
be
ideal
because
the
this
so
it's
like
has
two
modes,
one
or
three
with
the
online
offline,
but
if
you're
not
offline,
you
can
also
be
away
if
you're
away,
then
the
bot
won't
ping
you
and
if
you
set
yourself
dnd,
which
is
like,
has
some
default
settings
like
away
for
half
an
hour
or
something
further,
then
they
won't
be
in
you
either.
So
we
ideally
if
the
engineer
goes
just
grab
a
couple
of
water
or
they
go
for
lunch
for
an
hour.
D
F
D
F
F
D
Ideal,
if
we
were
if
we
can
recommend
that
as
well,
that
would
be
perfect
because
it
happened
the
other
day
as
well.
I
think
someone
went
for
a
glass
of
water
and
then
missed
the
notification.
They
did
reply
afterwards
anyway,
because
it
was
just
a
couple
of
minutes.
So
it's
fine.
B
A
D
Yeah,
the
other
thing
is,
I
think
it
was
mentioned.
I
don't
know
if
there's
an
issue
for
this
to
being
the
top
five
or
top
ten
engineers
at
the
beginning
of
the
week,
so
so
they're
aware
that
they're
aware
that
they
may
get
things
very
soon,
because
to
be
honest,
I
think
this
business
command
top.
That
shows
you,
the
top
25
people
may
be
pinged
next,
so
we
can
pick
those.
D
So
we
can
send
a
reminder
of
something
and
so
they're
aware
that
the
next
week
they
may
get
things
as
well,
there's
definitely
more
probabilities,
while
the
others
can
probably
forget,
because
they
won't
get
being.
I
mean
out
anyone
after
20
position,
probably
won't
get
being
at
all.
B
I
think,
while
we
are
enhancing
the
tool,
I
think
this
is
a
fundamental
question
here
initially,
when
we
did
the
poll
survey.
Indeed,
some
people
prefer
to
predict
predictability
or
deterministic
of
their
their
schedule.
So
some
people
like
like
this
idea
of
this
pilot
program.
Some
people
prefer
the
predictability
so
probably
after
running
this
four
months,
we
should
do
that
survey
again
because
more
people
are
involved
now,
so
they
can.
They
have
a
better
idea
what
this
means
today.
B
So
let's
do
I
suggest
we
do
a
survey
again
to
see
where
the
preference
is.
If
majority
of
the
people
prefers
the
predictability,
we
may
think
about
it
differently.
I
mean
think
about
the
pilot,
whether
the
truly
works
well
or
not.
It's
still.
It's
still
something.
This
is
a
fundamental
question.
It's
not
about
the
process
itself.
D
Well,
one
thing
I
noticed
is
that
not
many
people
are
using
the
private
commands,
such
as
top
or
position
to
check
where
they
are
at
the
list.
I
saw
very
few
people
using
it,
so
maybe
we
should
also
make
sure
that
they
know
about
these.
D
D
B
Any
other
topics
to
go
about
maybe.
E
I
was
just
gonna
say
like
I
just
want
to
say.
Thank
you
all
of
you
for
this.
The
interaction
with
the
process
is
so
much
easier
having
a
bot
and
people
being
pinged
rather
than
having
to
like
you
know,
go
off
and
find
a
spreadsheet
and
hope
that
one
person's
around
so
yeah,
like,
even
even
in
kind
of
the
early
stages
like
yeah,
definitely
thank
you
so
much.
E
It's
been
hugely
easy
and
enjoyable
to
use
it,
and
on
this
on
the
whole
like
when
people
do
are
available
and
respond
to
the
notifications,
like
it's
brilliant,
like
people,
pile
in,
and
collaborate
well
and
kind
of
get
things
resolved
really
quickly,
so
yeah
really
enjoying
this
new
process.
A
Think
in
the
issue-
so
I
I
tagged
all
of
us
here
so
please
review
edit
my
notes
and
add
things
to
it
and
correct
it,
etc.
I'll
also
post
the
recording
as
well
so
cool.
B
So
what
are
the
next
steps?
We're
taking
continue
to
improve
the
the
bot.
A
So,
actually,
let
me
hit
the
save
here,
so
you
can
see
it
in
the
comments
so
but
a
quick
summary
after
the
handbook
updates
are
merged
for
recommended.
Slack
configuration
remind
all
back
engineers
to
do
this.
Configuration
about
a
week
later
create
an
issue
to
ask
all
ems
with
back-end
engineers
to
confirm
that
the
team
has
set
the
slack
configuration
extend.