►
From YouTube: 2021-08-02 Multi Large Working Group Weekly
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
Good
morning,
good
afternoon
and
good
evening,
everyone
so
today
is
august
2nd,
it's
already
august
wow.
A
A
Welcome
everyone.
Let's
get
started
what's
being
done,
amy.
B
Yes,
thank
you.
So
the
web
fleet
migration
is
still
progressing,
so
we've
completed
the
configuration
comparison
and
staging,
which
is
good.
We
didn't
find
any
major
differences
which
is
really
good.
So
hopefully,
when
we
do
the
same
on
production
like
it
will
be
pretty
clean
as
well,
and
we
have
done
some
setup
in
production
already
and
canary
and
then
the
next
thing
will
be
to
well.
I've
got
my
next
thing,
so
yeah
we're
making
good
progress
on
getting
things
set
up
and
moving
towards
canary
as
the
next
step.
A
B
So
the
next
things
well,
so
the
next
things
it's
actually
engine
x
is
what
we're
focusing
on
for
now.
So
over
the
last
couple
of
weeks
we've
been
investigating,
we've
got
an
issue
that
comes
around
on
nginx,
it's
affecting
the
api,
but
not
in
huge
numbers.
So
it's
it's
not
a
issue
right
now,
but
we're
expecting
it
will
be
a
bigger
problem
for
web,
so
we've
been
investigating
that,
but
actually
good
news
is,
as
of
yesterday.
I
know
hang
on
wait
today.
B
Apac
today
we
don't
think
we
actually
need
nginx.
So
we
are
currently
doing
an
investigation
to
see
if
we
can
just
remove
nginx
entirely
and
if
we
can,
then
that's
great
news
simplifies
the
tech
stack
as
well
as
avoid
this
possible
incident
that
we've
been
we've
been
trying
to
deal
with.
So
we're
looking
into
that
and
then,
as
sort
of
alongside
that,
we'll
be
continuing
to
push
towards
canary
in
the
next
few
days.
C
Yeah,
it's
maybe
a
bit
of
a
a
selfish
slug
on
my
side,
but
we
are
running
fast.
Is
there
any
incremental
updates
that
the
team
here
can
vet
that
it's
good
to
go,
so
we
can
run
on
top
of
get
we're
gonna
use
to
set
up
staging
too
any
mature
building
blocks
that
we
can
just
pass
through,
and
we
know
this
is
great.
Let's
just
set
it
up
in
a
new
environment
we're
setting
up.
So
we
we
continue
to
catch
up
at
a
an
appropriate
pace
to
what
all
of
you
are
doing
here.
B
I'm
not
sure
so
I
guess
this
is
probably
one
like
with
distribution.
I
guess
like
what
do
you
do
like?
I
suppose,
we're
still
adding
in
possible
changes,
I
suppose
the
config
and
into
the
charts,
but
like
how
solid
do
you
do
you
think
like
do?
B
B
So
I
suppose,
updates
from
here
are
probably
the
best
place
as
we
as
we
move
step
through
the
environment,
but
we
will.
We
do
have
them
on
the
on
the
epic
as
well,
but
certainly
we
can.
We
can
update
on
here,
but
I
think
once
we
get
to
canary,
we
will
gradually
be
increasing
the
traffic,
and
that
gives
us
the
confidence
and
we'll
reach
a
point
where
we
are
we're
fairly
sure
that,
like
what
we
have
set
up
is
it
is
solid.
Okay,.
C
B
C
A
Yeah
yeah
mac.
I
agree:
let's
decouple
this
to
at
first,
because
there
is
urgency
we
need
a
new
staging,
but
this
is
a
pretty
great
field
for
everyone.
C
A
B
Yes,
so
yes,
so
yeah,
so
absolutely
so
we
are
working
on.
Nginx
has
been
the
sort
of
big
question
mark
over
the
last
few
weeks
and
we
have
an
issue
which
does
affect
the
api.
B
We
have
been
trying
to
sort
of
diagnose,
we're,
still
continuing
to
try
and
diagnose
it,
but
current
states
we
think
cloudflare
gives
us
the
current
coverage
that
nginx
is
meant
to
be
giving
us
so
this
week,
that's
just
totally
new.
So
actually
now
we're
looking
at
whether
we'll
be
able
to
continue
the
web
migration
without
having
nginx
in
place.
E
So
I
actually
have
distinct
concerns
about
that,
because
we're
intentionally
moving
away
from
what
we
deem
as
good
and
official
architectures
we're,
relying
now
on
a
cdn
that
many
customers
of
reference
architectures
do
not
have
we're
effectively
taking
a
chunk
and
being
like
you
handle
it,
which
I'm
not
saying
understand
is
invalid.
E
D
To
level
set
on
what
were
like,
I
I
looked
through
that
investigating
an
issue
with
nginx
and
the
original
agenda
item,
and
so
I'm
trying
to
understand,
like
I
saw
this
thing
about
the
pods
not
being
ready
and
then
this
thundering
herd
when
they
finally
do
come
online.
Is
that
what
we're
are
we
aligned
on
what
you're
thinking?
This
is
jason?
B
D
B
To
the
bottom
of
this
investigation
issue-
and
let
me
see-
I
believe
that
so
he
eliminated
it
being
a
race
condition
and
yeah.
So
the
next
change
that
is
being
made
around
this
is
to
increase
the
number
of
ports
available
to
nginx.
So
it
sounds
like
there's,
maybe
some
additional
stuff
there-
that
we
could
also
do.
D
B
With
it
yeah,
that's
totally
fair
point
jason
thanks
for
for
making
that
one
would
you
mind,
we've
got
this
issue
1924,
where
we're
kind
of
going
to
be
working
through
the
changes
and
kind
of
looking
at
what
we
can
do
to
test
it.
Would
you
mind
just
putting
a
summary
of
of
your
concerns
that
from
above
onto
that,
so
we
can
wrap
those
in.
A
Let's
collaborate
in
that
issue
1924,
because
if
we
remove
the
injects,
that
could
be
a
big
one.
So
I
don't.
A
Else
we
want
to
involve
in
this
discussion,
jason,
or
maybe
do
you
have
ideas
of
who
else
in
the
engineering
group
should
be
involved
in
this
discussion
because
injects
is
a
quite
a
critical
piece
in
our
stack.
B
Here
I
mean
we
can
certainly
discuss
some
issues,
an
interesting
one,
the
thing
that
led,
so
we
only
at
the
moment,
use
it
for
api
and
web
and
at
the
moment,
the
the
thing
that
led
graham
to
this
current
latest
understanding
was
actually
api
hasn't
been
set
up
correctly,
which
means
actually
for
a
lot
of
the
calls
we're
not
using
it
for
api,
we're
only
using
cloudflare,
so
not
to
say
we
shouldn't
use
it,
but
just
right
now
it's
not
it's
not
actually
a
it's,
not
as
central
as
say
h,
a
proxy
in
the
in
the.
B
A
Let's
continue
the
research
and
then
whatever
it
is
to
be
aware
of
this.
So
let's
continue
the
discussion
in
the
middle
issue
yeah.
Thank
you.
E
The
only
thing
I
would
add
to
this
is,
as
we
are
looking
at
aj
proxy,
and
this
is
something
that
actually
production
has
been
using
in
in
mass
for
quite
some
time
to
handle
at
our
load,
and
I
believe
we
have
this
integrated
into
some
features
of
get.
E
Maybe
we
should
start
to
consider
not
necessarily
dropping
nginx
but
replacing
nginx
with
aj
proxy
white.
We
have
the
ability
to
do
that
technically
within
omnibus,
once
we
package
aj
proxy.
Now
we
would
have
two
ingress
providers
in
the
package
and
inside
of
the
charts
we
technically
have
the
ability
to
swap
out
nginx
for
aj
proxy
as
well.
B
A
Thank
you,
jason,
okay.
Moving
on
to
next
one,
the
discussion
item
will
show
us
from
me:
it's
basically
a
food
for
that
everyone,
pretty
much
got
the
contacts
right
now
right
in
q3
and
probably
moving
forward
too.
Our
focus
will
be
everyday
issues,
everybody
related
issues
and
also
the
corrective
corrective
actions
from
the
production
incidents.
A
So
that
leads
me
to
seeing,
if
does
it,
make
sense
to
label
all
the
multi-large
issues,
development
related
issues
as
infra
dive
issues
and
then
close
out
this
working
group?
That's
what
I'm
thinking
because
of
this
of
the
focus
shift.
The
reasons
behind
my
thinking
are
the
bullet
a
b
and
c.
So,
basically,
of
course,
this
working
group
every
divided
issue
from
this
group.
They
are
inferred
dive
related,
so
they
are
related
to
infer
dev
and
also
you
know,
without
the
infradab
label.
For
those
issues
there
could
be.
A
These
issues
could
be
de-prioritized,
based
on
our
current
focus
of
different
error,
budget
and
corrective
actions,
and
also
the
third
reason
is
the
existing
depth
process.
Possibly
it
possibly
makes
this
working
group
redundant
because
the
priority
is
given
to
interdep
and
if
we
label
it
and
also
our
issues
are
infradive
related.
So
if
we
label
the
our
issues
as
different
issues,
then
we
have
two
processes
here.
A
D
It
seems
like
the
working
groups,
kind
of
change
and
then
the
players
change
over
time
and
I'm
not
sure
we
always
do
a
good
job
of
kind
of
re-level
setting
as
people
change
out
and
focus
change
to
make
sure
that
we
actually
are
like
okay,
here's
here's
what
we
want
to
accomplish
in
this
working
group.
So
that's
really
my
only.
E
B
Yeah
I
sort
of
tend
to
agree.
Like
I
don't
mind:
if
labels
like
issues
have
additional
labels,
I
think
that's
if
that
keeps
them
in
sync
totally
fine.
They
do
think
this
is
interesting
group
in
terms
of
like
discussions
today,
we've
had,
and
particularly
as
we
go
through
italy
and
and
things
like
that,
like
it,
it
does
seem
that
being
able
to
keep
in
sync
on
those
things
will
be
will
be
useful
additions.
A
Yeah,
okay,
good
good
points.
I
let's
continue
this
working
group
for
now
yeah
the
discussions
today.
Definitely
if
we
don't
have
this
working
group,
probably
we
don't
have
a
forum
to
discuss
these
topics.
A
Good
points.
Okay,
thank
you
and
moving
on
to
the
follow-up
items
here.
So,
basically,
that's.
We
need
a
issue
to
record
the
pre-effective
testing
results.
Cj
is
not
here
I'll
I'll
reach
out
in
offline
in
the
working
group
channel.
A
A
Recharge
yeah,
that's
the
end
of
the
agenda,
any
other
stuff
we
want
to
go
through
today.