►
From YouTube: 2022-02-22 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).
B
Absent
for
a
few
meetings
because
of
ski
trips
and
then
unfortunate,
coming
down
with
covid
so
past.
That
now,
which
is
good
yeah.
B
Hi,
hey
dave
tiger
before
we
get
to
introduce
dave
he's
on
my
team,
also
working
on
some
op-up
stuff,
and
he
he
made
the
pr
for
the
spec
hi
dave.
C
Seemed
like
a
good
good
first
issue,
so
I
was
happy
to
do
it
commit
with
just
markdown
code.
I
like
it.
B
A
I
don't
know
if
shamek
is
joining,
because
the
issue
that
I
wanted
to
talk
about
is
somewhat
related
to
what
he
proposed
earlier,
or
maybe
we
can.
Okay,
maybe
wait
one
more
minute
and
if
he
doesn't
join,
we
can
move
forward
or
maybe
we
start
with
the
second
one
first
anyway,
let's
give
him
a
minute.
A
Okay,
he's
still
not
here,
I
don't
know.
If
he's
going
to
join
andy
david,
we
want
to
maybe
start
with
with
a
restart
shutdown
issue.
First,.
B
Sure
so,
just
a
little
history
for
people
that
might
not
have
been
paying
attention
to
the
repo
I
submitted
a
pr
about
a
month
ago
to
add
the
capability
of
restart
shutdown.
After
discussion
with
tigran,
we
decided
to
take
the
approach
of
a
server
to
agent
command
and
then
that's
in
pr
in
the
op
amp
go
repo,
and
then
we
sort
of
back
ported
that
implementation
to
the
op
amp
spec,
there's
a
pr.
B
The
ibm,
spec
repo
as
well
basic
ideas
of
server
and
agent
command,
has
two
command
options
right
now,
either
restart
or
shutdown
and
when
receiving
that
command.
The
agent
should
either
restart
or
shut
down
as
specified.
A
A
Okay,
anyway,
please
review
it
if
you're
interested.
I
will
review
myself
as
well.
A
A
However,
since
then,
I
think
so
a
couple
of
things
right,
one
is
that
the
the
metric
was
never
added
in
to
the
open
country
spec.
So
there
is
nothing
to
which
we
can
refer
to
or
delegate
to
this,
this
notion
of
healthiness,
but
also
I've.
A
I've
been
thinking
about
some
of
the
the
control
logic
that
you
would
want
to
have
on
the
on
the
on
the
server
side,
on
the
management
server
side,
and
so
when
you're
doing,
for
example,
things
like
rolling
updates
of
of
the
agent
right
you're
updating
the
fleet
and
the
server
probably
wants
to
monitor
how
the
updates
are
are
going,
and
maybe,
if
something
goes
wrong,
the
server
probably
wants
to
make
a
decision
about
stopping
the
updates,
maybe
rolling
back
what
was
done
if
something
goes
wrong,
and
and
for
this
the
server
needs
information
like
this
right.
A
So
I
I
posted
an
up
an
example
of
the
kind
of
a
logic
that
the
server
may
want
to
have.
So
something
like
this
when,
when
you're
doing
rolling
updates
of
the
fleet,
you
want
to
see
if,
if
90
of
the
agents
don't
become
healthy
within,
let's
say
five
minutes
after
the
package
is
installed.
The
new
version
is
installed,
then,
probably
something
is
wrong
right.
Maybe
you
want
to
pause
the
the
updates
or
you
even
want
to
roll
back
to
the
previous
version.
A
So
logic
like
this
requires
some
very
basic
health
information.
At
least
to
be
available
to
the
to
the
server
right,
which
technically
may
be
available
somewhere,
where
the
metrics
are
being
sent,
but
that
that
metric
backhand
is
not
necessarily
available
to
the
server
right,
with
the
way
that
open
spec
is
written.
It
says
that
the
the
management
server
just
tells
the
agent
where
to
send
the
metrics
and
and
doesn't
care
about
that
anymore
right.
A
A
A
I
don't
know,
but
very,
very
limited
set
of
things,
which
are
also
very,
very,
very
precisely
defined,
mostly
boolean
things
right,
not
not
things
like
cpu
utilization,
which
which
is
very
variable,
which
we,
which
you
probably
don't
know
how
to
act
on
but
stuff,
which
is
either
good
or
bad
right
and
yeah
and
which
is
also
which,
which
you
can
maybe
report
very
easily
from
the
agent
yeah.
B
I
I
personally
think
it's
valuable.
We
we
hope
to
report
that
kind
of
information
right
now.
It's
based
solely
on
whether
the
agent's
connected
and
if
we
remember
agents
that
have
connected
and
but
we'd
like
to
be
able
to
have
agents
connected
but
unhealthy,
which
is
not
represented
right
now.
A
B
I
don't
want
to
go
off
topic
too
much
or
get
in
the
weeds,
but
I
did
notice
in
the
op
implement
implementation
and
when
the
configuration
is
applied
and
there's
an
error,
there's
no
response
to
the
server
it
seems
possibly
related.
But.
A
B
I'd
have
to
double
check.
What's
in
the
spec,
I
do
know
what's
in
the
receiver
implementation
and
when,
if
an
error
is
received,
it
doesn't
require
a
status.
A
A
But
it
is
in
the
stack,
yeah,
okay
and
you're
right.
This
is
one
of
the
things
that
you
probably
would
also
want
to
to
base
the
decisions
on
right
when,
when
you're
updating
the
configurations
yeah,
so
this
does
exist,
we
do
have
some
some
sort
of
information
already
reported
there.
It's
just
that.
I
think
we
need
to
maybe
report
a
bit
more
than
we
than
we
do,
particularly
about
the
healthiness
of
of
the
process
of
the
agent.