►
From YouTube: 2022-02-02 Workspace meeting
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
B
A
Okay,
so
again
that
I
will
find
it
in
a
second,
but
my
first
question
is
regarding
the
locking
problem:
how
are
you
standing
with
the
locking
problem.
B
Well,
we're
still
investigating
and
let
me
link,
I
don't
know
if
I
think
you
should
have
the
issue,
but
let
me
link
the
issue
where
we
discuss
most
of
this
stuff.
Give
me
one
second.
B
So
like
we
now
basically
well,
I
don't
know
that
this
is
a
problem
on
production
to
be
totally
like
blunt,
I
think
it's
more
of
a
qa
issue.
B
B
So
the
locking
is
happening
when
a
new
name
space
is
created
because
we
need
to
create
this
traversal
ids
collection,
so
we're
basically
locking
the
hierarchy
so
that
we
are
making
sure
nothing
is
modified
while
we
compute
these
traversal
ideas
and
update
the
the
record
itself
so
because
we
were
not
doing
that
during
project
creation
before
this
was
not
an
issue
either
in
qa
or
for
anyone.
B
Now
we
started
doing
that
for
project
creation,
so
obviously
they
started
popping
up
and-
and
I
think
the
main
problem
like
it's-
not
the
the
only
situation,
the
only
scenario,
but
I
think
the
main
scenario
is
that
we
run
these
tests
in
parallel
on
qa.
So
what
happens
and-
and
I
think
most
of
them
most
for
most
of
the
cases
we
create
the
project
for
a
for
qa
user
on
its
personal
name
space.
So
what
happens
is
the
personal
name?
B
Space
gets
blocked,
we
create,
like
I
don't
know,
maybe
like
thousand
maybe
hundreds
of
projects
and
that
one
gets
locked.
So
we
get
a
long
queue
of
projects
that
are
waiting
for
the
for
the
vlog
to
be
released
and,
like
maybe
first
I
don't
know
10
20
100
projects
do
get
created,
but
at
some
point
still
some
project
hits
the
the
the
lock.
B
So
it
feels
more
like
a
this,
this
kind
of
issue
where
we
run
these
things
in
parallel,
so
more
of
a
question
issue,
but
it
it
still
can
happen
also
if,
if
some
customer
will
run
a
spirit
to
through
api
or
whatnot
trying
to
create
many
many
of
these
projects.
B
So
my
I
was
discussing
this
with
jan
as
well
this
week
so
like
maybe
one
alternative
is
to
let
qa
know
about
this
issue
enable
the
feature
flag,
back,
watch
it
for
a
bit
more
on
production
and
if
we
don't
like,
maybe
do
some
sort
of
a
percentage
comparison
see
how
many
times
we
we're
actually
hitting
this
this
issue,
if
it's
like
a
a
very,
very,
very
small
fraction
of
a
percent
or
something
like
that,
we
can
move
on
and
unblock
everyone
from
just
moving
on
with
the
project
itself
because
like
currently,
this
is
blocking
basically
everything.
B
So
that's
that's
one
option,
while
in
the
meantime
we
try
and
reduce
this
problem,
we'll
I
don't
think
we'll
be
capable
of
like
removing
the
problem
completely,
because
we
need
to
create
this
namespaces
when
the
project
is
created
and
we
cannot
remove
the
log
because
then
we
kind
of
end
up
in
this
invalid
state
with
the
traversal
ids
or
we
have
the
potential
of
ending
up
there,
we're
discussing
some
options
of
moving
trying
to
move
this
update
and
the
locking
as
low
as
possible
in
the
transaction.
B
So
like
right
now.
What
happens?
Is
like
first,
the
namespace
is
created
and
then
a
bunch
of
other
inserts
updates
and
things
are
happening,
so
the
lock
is
very
up
in
the
transaction.
So
obviously,
instead
of
like
10,
milliseconds
or
100
milliseconds,
it's
it
takes
more
time
until
the
transaction
is
committed,
so
moving
it
as
low
as
possible
in
the
in
the
like
queries
that
happen
within
the
transaction
would
kind
of
help
make
the
problem
smaller.
It
will
not
solve
the
problem.
B
Obviously,
because
you
can
you
can
still,
depending
on
on
the
number
of
the
projects
that
you
create.
You
can
still
hit
this
locking
thing.
Another
thing
is
like
moving
it
totally
out
of
the
transaction
into
after
commit,
but
then
you
kind
of
can
end
up
in
this
situation
where
traversal,
ladies,
are
not
valid,
or
you
are
with
the
list
of
like
not
complete
or
invalid
traversal
ids,
which
then
needs
to
be
solved
with
some
cron
job
that
goes
and
runs,
and
checks
and
and
updates
those
which
is
not
ideal
as
well.
B
So
I
think
there
are
some
a
couple
other
options
there,
where
these
are
mainly
two
aspects
that
we
were
like
discussing,
moving
it
as
low
as
possible
in
the
transaction
or
moving
it
out
of
the
transaction
completely
yeah.
There
is
some
other
options
of
not
locking
the
root
namespace,
but
looking
like
closer
to
what
we
like
to
to
the
namespace,
where
we
create
the
projects
where
the
new
namespace
is
created,
that
does
have
its
own
complications
as
well.
So
that's
where
the
discussion
is
mainly
right.
B
A
B
In
the
issue,
it's
me
jan
and
alex
okay,
so
yeah.
He
said
that
he
went
through
a
lot
of
these
things
that
we
discussed
when
he
was
developing
that
so
some
of
the
things
are
not
exactly
clear
to
me,
but
hopefully
he'll
he'll
shade
some
light
on
yeah.
A
I
think
that's
apart
from
this
being
a
qa
issue
that
you
mentioned
that
yeah.
A
This
is
like
a
particular
situation
when
you
create
a
lot
of
projects
in
the
parallel
which
I
understand,
but
I
think
we
can
also
hit
that
when
importing
groups
with
like
when
we
import
groups
a
lot
of
projects,
so
I
think
it
would
be
good
to
like
ask
george
from
import
group
to
just
like
like
validate
if
we
are,
if
we're
not
liking
them,
so
making
things
harder
for
them,
like
I
don't
know
exactly
what's
a
kind
of
if
they
are
using
the
api,
endpoint
or
they're
using
something
in
the
code.
A
B
We
kind
of
surfaced
when
we
started
breaking
down
the
phases,
so
we
don't
like.
We
don't
create
projects
on
grouping
like
it's
very
weird
to
me,
but
like
groups,
imports
and
project
imports
are
separated
so
to
say,
but
the
this
problem
of
locking
is
still
there.
It's
like
it's
not,
and
it
was
there
and
if
you
can,
I
probably
can
share
my
screen.
B
Yeah,
let
me
do
this
now
sure.
B
Hopefully,
you
can
see
the
browser
now
yeah,
okay,
so
what
you
can
see
here
is
I
filter
it
for
this
locking
problem.
Only
for
the
group's
end
point-
and
you
can
you
can
see
even
after
we
disabled
the
flat
the
feature
flag,
we're
still
hitting
this
issue
like
it's
very
it's
not
as
often
but
like
my
explanation
for
it,
for
it
not
being
as
often
is
because,
basically
we
don't
create
as
many
groups
in
space.
B
So
the
problem
is
still
there
and
it
was
there,
but
it
was
not
as
obvious
as
when
we
introduced
the
namespace
creation
in
the
projects
where
we
can
see
these
spikes
there.
B
It
would
be
nice
to
create
a
report
here
and
see
like
how
big
of
a
percentage.
This
was
compared
to
overall
project
creation
on
group
creation
and
so
on.
So
because,
like
we
have
like
19
errors,
but
if
we
created
200
000
projects
in
the
meantime,
it's
like.
Is
it
really
that
do
we
really
care
to
block
everything
because
of
these
19
euros?
On
the
other
hand,
if
you
create
20
projects
and
19
of
them
fail,
then
yeah,
you
kind
of
have
to
think
how
to
make
it
less
of
a
problem.
So
to
say.
C
I
think
I'm
thinking
of
too,
if
this
is
an
issue
that
won't
be
as
common
in
production
and
we
do
determine
it
is
mainly
seen
like
through
qa.
Maybe
it
warrants
going
through
our
specs
and
seeing
where
we
might
be
able
to
share
some
projects
instead
of
creating
like
a
new
one,
every
single
time.
I'm
not
sure
if
that
can
help
at
all
with
this,
but
that
might
be
something
to
look
into.
A
When
you're
not
creating
we're
not
trying
the
problem
so
yeah,
this
will
definitely
help
but
yeah.
Well,
we
don't
like,
but
we
also
need
to
like
understand
that
if
it's
only
a
qa
issue,
I
understand
that
we
it's
rarely
that
that
some,
like
groups
or
noises,
are
creating
many
projects
at
once.
Like
it's
probably
it's
probably
super
rare.
B
Yeah,
it's
not
only
a
qa
problem
just
to
make
it
clearer.
It's
like
it's
not
only
qa
and
and
only
because
it's
running
parallel
but
like
I
think
it's
surface
is
more
just
because
we
do
that
with
the
qa,
because,
like
I've
been
running
some
query
yeah
in
the
center,
so
I
I've
been
running
somehow
I
was
even
on
kibana
and
on
center.
I
was
really
hitting
the
qa
user
appearing
a
lot
in
these
years,
both
on
production
and
so
like.
Let
me
re-share
my
screen.
B
So
like,
for
instance,
here
on
on
sentry,
we
have
this
here
on
groups
and
we
have
17
events
but
like
the
single
user
that
is
is
related
to.
That
is
good
luck,
qa
on
production.
So
this
is
what
led
to
me
like
to
think
that
this
is
probably
somehow
related
to
how
we
run
tests
and
specs,
and
things
like
that
so
yeah.
B
B
And
probably
like
next,
this
next
days
I'll
try
to
create
this
report.
So
I'm
trying
to
see
how
many
post
requests
we
have
to
the
projects
and
groups
and
then
see
how
many,
what's
the
count
of
the
errors
and
see
if
we
can,
we
can
get
a
percentage
there
and
at
least
like
how
see
how
much
it
really
affects
the
project
creation.
B
Because
again,
if
it's
a
fraction
of
a
percentage,
maybe
we
were
able
to
make
the
decision
that
yeah.
Let's,
let's
move
on
with
what
we
have
and
work
on
the
issue
in
parallel.
B
If
it's
needed,
no,
I
don't.
I
don't
think
we
need
database
with
that
yeah
I
can.
I
can
pull
the
projects
being
created
either
from
kibana
or
from
from
pgai
like
by
david
or
something
like
that
and
have
some
sort.
A
A
No,
I
I'll
just
switch
on
the
notification
for
the
issue.
Thanks
yeah.
A
That's
a
good
call,
okay,
so
there's
one
more
thing
from
me:
it's
regarding
the
phase
two.
I
don't.
I
know
that
alexandria,
not
you're,
involved
in
that
still
slightly
one,
but
I
hopefully
but
the
phase
two
we
I
would.
A
I
will
post
the
same
thing
on
on
slack
because
there's
not
not
many
people
here,
but
we
break
down
phase
two
like
three
aspects
of
phase
two,
so
memberships
policies
and
routing
into
like
smaller
issues,
but
there
are
still
like
things
from
phase
two
that
are
not
broken
down
into
smaller
issues
and
aren't
researched.
A
I
think
this
was
something
with
api,
and
so-
and
I
call
this
like
face
to
a
and
phase
two
b
so
face
to
a
is
already
broken
down
and
face
to
b
is
the
things
we
still
need
to
broke
down,
and
I
would
say,
like
probably
the
end
of
february,
we
need
to
like
start
breaking
down,
also
phase
two
to
understand
the
scope
of
what
we
need
to
do
in
this
like
face
to
be
as
well.
A
B
Yeah,
I
think
the
problem
with
face
to
b
is
that
we
don't
really
know
what
is
going
to
go
in
there.
Some
of
the
things
on
top
of
my
head.
I
think
there
were
some
discussions
on
query
performance
specifically,
for
instance,
for
epics
or
some
other
things
yeah
like
yeah.
A
Okay,
but
we
just
you,
know,
I
think,
to
understand
how
big
the
like
this
unbroken
issues
can
get
and
and
without
like
any
research,
it's
hard
to
say
how
how
much
of
the
work
it
still
needs
to
be
seems
to
be
done
in
phase
two.
So
I
understand
that,
like
I
understand
that,
asking
developers
to
estimate
issues
like
the
worst
thing,
because
everyone
loves
doing
that,
but.
A
But
but
I
just
wonder
if,
but
I
just
like,
I
know
that
at
some
point
we
need
to
also
like,
let's
like
understand,
the
scope
of
the
work.
B
Okay,
like
a
good
part
of
the
phase,
2,
is
also
replacing
usages
of
project
with
project
namespace.
So,
basically,
instead
of
instead
of
listing
projects,
for
instance
on
the
project,
page
you'd
list
actually
project
namespaces
with
a
reference
to
the
project
and-
and
things
like
that-
and
it's
hard
for
me
to
quantify
like
how
much
work
that
is
but
yeah,
maybe
maybe
you
can
run
some
grip
or
something
like
that,
and
just
like
at
least
have
a
list
of
files
that
I
don't
like.
I
don't.
B
B
Now,
let's,
let's,
instead
of
moving
using
the
project,
let's
start
using
the
project,
namespace
everywhere
that
we
can
right.
So
that's
that's
more
of
a
entire
company
effort,
rather
than
then
workspaces
group
greatly
effort
where
you
need
to
go
and
change
these
things
because
then
like
there
is
nothing
blocking
other
teams
to
actually
start
doing
that.
If
we
have
everything
backfilled
and
yeah,
and
things
like
that,
so
yeah.
A
B
Yeah
yeah,
so
so,
like
maybe
phase
to
b,
is
actually
try
and
investigate
what
else
needs
to
be
like
migrated
before
we
can
hand
it
off
to
the
to
the
entire
company
like,
like
memberships,
needs
to
be
backfield
yeah.
What
we
have
there
is
to
be
back
here.
Maybe
there
is
something
else
that
also
needs
to
be
backfilled
before
people
can
start
working
on
that.
So
so
maybe
that's
something
that
we
should
be
doing
for
them
to
be
basically
identify
these.
A
Okay,
I'll
start
this
discussion
like
how
can
we
do
it,
the
in
the
most
efficient
way,
not
interrupting
your
work?
That
much
and
but
also
like
having
like
some
answers
rather
than
more
questions.
I
know,
maybe
it's
better
to
wait
a
little
bit
for
like
phase
two
to
unravel
more,
but
we
cannot
like
wait
with
this
like
too
long,
because
that
means
also
to
understand
the
scope
of
work.
B
There
is
a
in
my
mind.
At
least.
There
is
a
scenario
where
we
can
start
handling
it
off
even
after
we
do
just
like
membership
and
routing
and
things
like
that,
and
then
we
have
these
examples,
how
it
is
done
so
then,
if
if
we
need
to
like,
if
a
team
needs
to
backfill,
I
don't
know
like
vulnerabilities-
or
I
I
don't
know-
is
the
backfilling-
they
can
take
these
examples
and
and
do
it
with
the
resources
of
their
own
team
in
a
much
smaller
school.
I.
A
Think
there's
this
is
the
idea,
and
this
is
what
like,
we
would
love
to
do
and
like
what
teams
would
love
to
do
because,
like
there's
always
like
a
couple
of
teams,
that's
like
looking
forward
to
backfilling
the
project
namespaces
because
they
want
to
use
this
like
the
foundations
that
we
are
creating,
but
yeah
the
phase
two
is
still
like
critical
and
we
need
to
understand
right
right.
I
agree.
Yeah,
okay!