►
From YouTube: Runner 12 4 Kickoff
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
Hi
I'm
Josh
Lambert,
a
product
manager
here
at
get
led
and
I'm
responsible
for
the
verify
rudder
group
and
I'll
be
walking
through
the
issues
that
a
plan
to
deliver
and
our
12.4
release.
We
are
primarily
focused
on
addressing
high-priority
bugs
in
this
release
in
four
in
particular.
The
first
and
most
important
is
an
issue
where,
in
our
company's
executor,
there
are
some
instances
where
a
job
which
is
terminated
or
potentially
fails
in
the
middle
is,
is
still
reported
as
a
successful
job
back
up
to
get
live
itself.
A
This
can
lead
to
a
number
of
undesirable
behaviors,
as
you
can
imagine
that
the
pipeline
center
may
still
continue
where
they
partially
built
or
bad
code,
that
failed
unit
testing,
and
so
we
want
to
make
sure
we
solve
this.
For
our
users,
we
don't
know
exactly
what's
causing
it
or
exactly
all
the
instances
when
it
can
happen,
and
so
we've
actually
allocated
quite
a
bit
of
time
in
this
release
to
get
to
root
cause
and
then
fixing
it.
So
this
is
an
important
one
for
us
to
address.
The
next
two
are
related
to
job
cancellation.
A
Similarly,
we
also
want
to
solve
the
issue
where,
in
some
cases,
when
you
click
cancel
and
the
gitlab
UI,
the
runner
continues
to
run
the
job.
It'll
be
canceled
and
so
future
later
stages
and
jobs
will
not
continue
to
run,
but
that
job
itself
would
continue
execute,
and
this
again
could
cause
some
usage
and
compute
costs,
as
well
as
potentially
unexpected
behavior,
where
you
cancel
the
job,
but
it
still
might
run
to
completion
which
again
might
not
be
expected.
A
We've
traced
this
back
actually
to
issues
with
the
dock
machine
in
de
based
plug-in
itself
and
how
it
interacts
with
the
AOS
API,
but
essentially,
once
you
get
to
this
state,
it
just
continues
to
utilize
the
API
and
continues
to
trigger
the
essentially
the
light
of
it
and
doesn't
have
proper
back-office
in
place,
and
it's
also
making
some
inefficient
usage
of
of
some
API
calls.
And
so
what
we're
looking
to
do
here
is
to
actually
looking
to
solve
this
issue
rather
than
trying
to
band-aid
it
at
the
good
WebRunner
side.
A
We're
actually
looking
to
solve
it
at
the
duct
machine
side.
Now,
since
dr.
machine
is
in
maintenance
mode-
and
they
haven't
been
quick
to
merge
our
previous
contributions,
we're
actually
going
to
be
forking,
dr.
machine
and
addressing
some
of
the
important
bugs
that
were
hitting
in
this
fork
and
offering
out
to
customers
that
they
then
can
use,
and
this
will
essentially
continue
this
for
again
hi,
pretty
bugs
dr.
machine,
while
they've
worked
to
have
a
alternative
in
place
for
dr.
A
Machine,
going
forward
we're
making
progress
on
the
windows,
auto
scaling
runner
right
now
and
that
is
being
utilized
to
help
explore
other
articling
runners
that
we
can
use
or
executives
in
menus
for
other
systems
like
Linux
and
Mac
as
well.
So
we're
exploring
that
now,
but
the
interim
again
for
our
users
we're
using
that
machine,
which
is
a
lot
of
folks.
We
wanted
to
improve
this
particular
pain
point
here
on
AWS,
so
those
are
the
four
issues
that
we're
playing
to
deliver
here
again
to
the
floor
primarily
focused
on
closing
some
high-priority
bugs.