►
From YouTube: Service APIs Office Hours 20200701
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
We're
recording
this
is
a
relatively
sparse
office
hours
for
July
1,
happy
Canada.
Today,
everyone
and
I
know
a
lot
of
people
are
out
of
office
today
and
for
the
rest
of
the
week,
but
for
the
few
that
are
here,
I
think
it'd
be
good
just
to
go
through
any
outstanding
pr's
and
see
if
we
can
close
them
out
I'm
open
to
any
other
ideas
for
things
we
should
discuss
today.
But
I
know
there
are
few
PRS
that
could
use
some
attention
so
yeah,
let's,
let's
get
into
it,
I
know
this.
A
B
C
C
C
C
B
A
B
A
B
A
B
Okay
thanks,
so
that's
interesting,
I,
wonder
whether
we
should
be
using
that
and
the
reference
types
we
define,
probably
not
I,
don't
think
based
on
upstream
guidance
that
we
want
to
use
it
for
object
references.
So
it's
it's
surprising
to
me
to
see
that
here,
I
wonder
whether
that
should
be
three
fields
or
whether
we
should
be
using
the
meta,
v1
type.
A
B
D
D
B
B
A
B
I
can
do
that
basically
I
guess.
The
two
concerns
are:
is
this?
The
right
way
is
using
group
resorts
group
version
resource
the
right
way
to
do
this
sort
of
thing
and
the
what
you
end
up
with
you
have
the
nested
resources
are
resources
that
something
we
want.
I
can
yes,
add
a
comment
asking
about
this
point.
B
A
D
A
All
right:
well,
let's
move
on
then
to
the
next
PR
and
I
think
this
one
is
I.
Think
I
can
lgt
him
this
one
today,
I
just
wanted
to
run
it
by
everyone,
one
more
one
more
time
or
actually
already
has
lgt
images
as
a
hold,
but
James.
This
is
your
PR,
I
I
know:
we've
already
discussed
this
I.
Don't
think
too
much
has
changed
that
since
last
week,
right
no.
C
I
fixed
it
I
think
I
fixed
the
typo
there,
but
we've
found
yeah
and
every
based
yeah
I
once
his
Lance
I'll
do
the
following
work
to
you:
propose
a
DNS
type
address,
okay,
think
about
what
that
means.
The
belief
that's
used.
That
seems
like
useful
and
I
believe
yeah
Andrew
put
a
comment
describing
how
AWS
uses
that.
E
A
Thank
you
for
the
PR
this
one
I
know
it
was
either
nearly
the
proof
that
it
felt
like
it
was
so
close
to
getting
in
but
yeah.
Now
it
has
all
the
conflicts
in
the
world.
I
think
there
was
enough
support
for
this.
I
know.
James
you've
mentioned
that
p.m.
what
I
think
is
James.
This
someone
mentioned
that
you're
moving
to
dynamic,
client,
yeah
contour.
D
C
A
A
So
what
I
mean
what
I'm
hearing
here
is?
There's
no
one
absolutely
needs
this,
but
no
one
hates
it
heater
Oh
which-
and
it
sounds
like
it-
would
help
his
use
case
out
at
least
and
potentially
maybe
a
bit
more
scalable
if
you're
dealing
with
lots
of
resources
here
and
a
bit
easier
to
work
with,
but
it
does
mean
more
cogent
for
actually
developing
service
api's.
A
A
A
Okay,
I
will
I
will
leave
that
to
them
for
a
while.
That's
cute
and
traffic
mirroring
I,
guess
yeah
I
guess.
A
lot
of
these
I
should
just
wait
for
Dan
Ian
to
be
on
a
call
to
discuss,
because
at
least
traffic
nearing
I
think
we
got
pause.
We
got
hung
up
a
little
bit
on.
Does
this
even
make
sense
in
other
use
cases
like
Multi
cluster
and
the
answer
was
well,
it
doesn't
really
break
anything.
It
should
work.
Fine
and
I've
got
wrote
a
talk
about
how
it
could
work
and
I.
C
C
A
Yeah,
that's
agreed,
I
wish
I
know
they
have
a
doc
where
they
we're
proposing
updates
to
this
and
splitting
and
more
but
I,
don't
see
a
LinkedIn.
This
key
are
all
right.
I
will
follow
up
on
that
one
and
that's
similar
like
oh
all,
of
these
PRS
are
all
kind
of
linked
together
and
part
of
their
newer
single
doc
as
I
recall,
we're
both
Harry
and
day
and
head
had
been
working.
A
C
A
Yeah
I
agreed
agreed,
I
I'll,
make
a
note.
I'm
gonna,
follow
up
on
this
issue
or
the
slack
thread
and
invite
him
to
one
of
our
meetings
or
office
hours.
Just
it
might
be
good
to
have
a
broader
discussion
about
other
protocols,
even
though
I
don't
think
at
least
a
LPN
or
DTLS
are
part
of
an
initial
alpha.
It
would
be
interesting
to
at
least
understand
the
use
cases
and
the
goals
so.
C
One
interesting
thing,
I
think
with
some
of
these
is
that
they're,
really
more
I
think
there
might
be
easier
to
implement
with
the
network
load
balancer
infrastructure,
then
for
people
doing
you
know
in
cluster
proxies
so
yeah
you
might
be
some
MIB
some
interesting
choices
to
make
around
how
we
can
express
what
a
particular
gateway
class
is
actually
able
to
do
with
these
things.
Yeah.
A
And
Jeremy
already
has
a
PR
to
address
this
issue
conflict
resolution.
We
may
want
to
wait
till
we
have
a
few
more
people,
but
I
think
it
would
be
good
to
go
back
through.
You
know
we
identified
a
number
of
potential
conflicts
and
at
some
point
we
need
to
follow
up
on
these
and
identify
how
each
potential
conflict
should
be
resolved
and
I
and
I've
identified
so
much
some
of
the
examples
right
here
and
just
like
a
guiding
principle,
and
maybe
this
turns
into
a
very
rough.
A
You
know
PR
so
I'm
like
just
a
guiding
principles
for
the
API
for
conflict
resolution,
but
this
doesn't
answer
every
individual
potential
conflict
and
I'm
not
sure
the
best.
The
best
way
to
do
that
in
some
sense
you
could
have
you
for
every
specific
field
that
could
have
a
conflict.
You
could
choose
to
specify
it
in
a
documentation
for
that
field.
How
conflict
should
be
handled
or
maybe
we
need
some
kind
of
overarching
documentation
of
all
the
potential
conflicts
and
how
each
one
should
be
resolved.
B
A
C
I
think
what
you
have
they're
saying
that,
if
something
invalid
comes
in,
then
you
keep,
the
valid
ones,
is
a
good
principle,
so
that
that
gives
implementers
guidance
that
they
shouldn't
nuke.
The
whole
gateway
goes
one
element
right,
because
you
can
make
it.
You
can
construct
an
argument
that
dropping
an
element
could
create
a
security,
critical
vulnerability
in.
A
A
You
know
with
the
with
the
proposed
updates
to
routes,
one
of
the
new
concepts
in
that
in
that
doc
was
the
idea
that
you
might
be
able
to
add
some
additional
custom
layers,
like
maybe
you
might
add,
some
custom
authentication
at
some
point
in
the
processing
of
the
route,
and
obviously
there
are
some
layers
where,
if
that
specific,
let's
say
layer
is
invalid,
you
probably
do
want
to
drop
everything
after
it
and
so,
like
all
guidelines
or
principles,
they're
going
to
be
examples
where
it
makes
sense
to
disregard
it
and
just
drop
so
yeah
I'm
interested
in
the
exceptions
to
the
rule
here
because,
like
you
say,
James
there
are,
there
are
certainly
potential
security
considerations
of
leaving
partially
valid
things
in
place.
A
Yeah,
okay,
I
think
I
think
you
know
we're.
Probably
we've
already
covered
just
about
everything
else.
We've
got
a
fairly
small
group
today,
so
I
think
this
is
enough
to
cover
today,
unless
anyone
wants
to
bring
up
other
issues
or
pull
requests
to
review.
I
think
I'll
leave
the
rest
for
next
week
when
we
have
a
few
more
people
back
in
the
office,
well,
not
in
the
actual
office,
but
whatever
their
office
happens
to
be
yeah
any
anything
else.
We
should
discuss
or
this
this
good
and
give
everyone
back
half
an
hour.
I.
A
Imagine
others
are
in
my
place,
but
with
code
freeze
coming
up
this
coming
week,
I've
been
more
distracted
than
I'd
like,
but
I
hope
that
coming
into
Wednesday
Thursday
of
next
week,
I
have
more
time
to
to
contribute
and
we'll
be
able
to
take
it
moving
again
on
alpha
I'm
narrowing
down
the
alpha
scope
and
actually
making
sure
that
the
various
items
we
have
in
there
get
in
you
know,
we
have
I
think
we
have
a
lot
of
things
in
flight
and
you
know,
instead
of
adding
more
things
to
do,
we
just
need
to
be
helping
helping
things
move
forward
that
are
already
already
close
to
done
or
proposed
and
need
an
implementation
to
go
along
with
it.