►
From YouTube: GMT20200110 k8s migration sidekiq: project export queue
A
B
C
B
So
last
week
we
set
the
goals
to
work
on
monitoring
our
metrics
I
had
a
bunch
of
merger
quest
ready
to
go
prior
to
the
holiday
season,
Andrew,
jarv
and
I
think
Benton
reviewed
at
least
one
of
them.
Those
are
all
put
into
place.
So
if
there's
anything
missing
out
of
metrics
and
our
dashboards
today,
that
just
means
we
missed
something.
B
The
auto
deploy
process-
this
is
an
epoch
where
we
are
trying
to
do
a
few
things.
One
is
creating
the
auto
deploy
branches
for
helm
and
CNG,
followed
by
upgrading,
just
like
we
do
with
omnibus
upgrading
the
version
of
our
components
inside
of
the
scene,
minimally,
the
cng
repo
and
then
tagging
that
repo
that
way,
the
build,
starts.
I
got
the
auto,
deploy
branches,
merged,
I,
need
to
work
on
updating
the
components
and
then
performing
a
tag
for
minimally.
The
CNG
repo.
This
epic
does
go
into
the
helm.
B
B
We
discussed
that,
but
I
don't
think
we
actually
really
wanted
to
do
it
yet
or
maybe
we
just
didn't
take
up
that
work.
We
discussed
it
because
it
was
a
easy
queue
that
we
could
potentially
do.
Instead
a
psychic
exporter,
but
we've
invested
so
much
time
in
the
export
of
queue.
I,
don't
think
we'll
get
to
this
anytime
shortly.
So
I
do
not
work
on
this
at
all.
B
C
Cleanup
work,
I,
think,
what's
happening
from
what
I
can
tell
is
that
there
are
cases
where
we
leave
temp
files
on
disk.
So
in
this
case
the
cleanup
worker
is
necessary.
I
haven't
dug
into
the
code
yet,
but
I
have
a
feeling
that
what
happens
is
is
that
you
start
the
export
something
blows
up
and
then
files
are
left
around,
which.
C
B
And
if
the
job
is
successful,
it
cleans
itself
up
cause
if
anything.
The
only
reason
why
we
see
more
disk
space
use
is
because
they
created
a
directory
structure
in
the
directory
structure,
did
not
clean
get
cleaned
up,
but
the
files
inside
of
those
did.
This
is
relatively
unfortunate
to
learn
about,
but.
A
C
Yeah
I
think
that's
what
we
should
do,
so
maybe
we
should
the
repairs
and
cleanup
I
need
to
see
in
the
code
how
exactly
we're
creating
the
dump
file
and
what
the
what
it
looks
like
before,
but
after
I
do
that
I'll
open
up
an
issue,
okay
and
crate,
staging
configuration
and
kubernetes
I
didn't
assign
this
to
myself
until,
like
I
only
assigned
this
to
myself
this
morning.
So
this
is
moving
forward.
I
have
an
mr2
prep,
soon
and
and
right
now,
I'm
just
I
was
just
talking
to
skarbek.
C
How
do
we
create
auto,
deploy
CNG
images
for
testing
and
the
way
I'm
gonna?
Do
it
right
now
is
I'm
just
going
to
create
an
auto
deploy
branch
manually,
I'm
gonna
bump
the
versions
I'm
gonna
tag
it
I'm
gonna
tag
it
with
the
same
tag
we
use
for
homnibus
and
get
that
or
just
omnibus,
and
then
then,
hopefully
we'll
just
use
that,
and
until
we
have
that
automated
as
part
of
Auto
deploy.
C
C
C
So
I'm
not
sure,
what's
going
on
here,
like
maybe
we're
just
being
overly
aggressive
and
CNG
about
converting
characters
like
globbing
lots
of
lots
of
them,
including
dots,
I,
think
it
would
be
nice
if
our
registry
tag
matched
exactly
or
at
least
very
close
to,
because
we
have
like
different
versions
of
you
have
we
already
have.
This
would
make
the
third
version
of
this
tag,
which
kind
of
sucks
cuz
we
lose
the
plus
as
well.
Maybe.
A
B
C
B
C
C
B
This
is
a
major
blocker
to
keeping
us
from
going
to
production,
I
feel
so
yeah.
Oh
I,
guess
something
I
didn't
mention
in
here.
I
started
the
readiness
review
and
I
created
a
few
issues,
but
like
that,
that's
that's
halfway
done
at
this
point.
So
okay
I'll
work
on
that
throughout
the
week
as
I
get
tired
of
writing.
Ruby
sure.
C
For
the
investigating
and
the
fest
mounts,
I
think
it
would
be
nice
I
think
we
have
a
grab
bag
of
issues
in
the
Gilad
tracker,
for
things
related
to
NFS,
like
upload
file,
locking
CI,
traces
and
and
then
of
course,
this
issue,
where
we
have
the
cron
cleanup,
I'm
gonna
try
to
get
those
under
like
or
at
least
make
sure
they
had
their
more
visible
and
easy
to
query.
So
we
can
track
of
them.
C
Note
that
currently
temp
is
unmounted
across
the
fleet,
but
that's
all
we
can't
unmount
cache.
It
sounds
like
we
can't
unmount
uploads.
You
definitely
can't
unmount
the
pages
artifacts
and
LFS
objects.
I
have
no
known
issues
that
but
I'm
also
kind
of
scared
to
do
anything
right
now.
So
I
don't
know
it's
a
good
idea
to
move
forward
and
unmount
artifacts
and
LFS
objects.