►
From YouTube: July 22, 2021 AUS TZ - Ortelius Architecture 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
B
Okay,
okay,
perfect
yeah,
I
was
just
gonna
say:
it'll
be
nice
to
collaborate,
but
it
is
so
that's
fine.
C
Yeah,
I
just
added
reason
I
just
got
signed
out.
Let
me
get
signed
back
into
that
doc.
C
All
right
so
today
is
the
july
22nd
artelias
architecture
for
the
australian
time
zone,
while
we're
waiting
for
ahmed
to
come
on
to
the
zoom
I'll,
just
kind
of
review
some
of
the
stuff
that
happened
at
the
mornings
meeting,
so
the
service
catalog
is
moving
along,
really
well.
Ukarsh
has
found
stuff
that
we
can
use
for
rendering
swagger
into
the
ui.
C
He
basically
wrote
a
go
program
to
kind
of
interface
with
docker
and
run
commands
inside
of
the
docker
image
for
us
to
grab
some
files
out
of
there.
Basically,
the
results
of
the
scanning
stuff
so
and
that
will
then
what
the
next
step
will
be
to
push
that
into
the
data
the
artillious
database,
one
of
the
things
that
will
that
that
program
will
eventually
just
keep
on
being
added
on.
C
So
if
we
need
to
add
on
a
new
scanner
that
we
want
to
interact
like
with
trivia
or
start
scanning,
node
js,
that
one
will
evolve
to
handle
those
things
and
we'll
be
shortly,
we'll
be
getting
into
testing
and
documentation
updates
for
the
service
catalog.
C
So
that's
kind
of
what's
happened
from
this
morning's
meeting
in
the
states.
So
did
I.
A
B
B
B
Yeah
sure
do
you
want
to.
You,
did
most
of
the
work
well
all
of
the
work,
because
I
was
busy
moving.
Do
you
want
to
give
us
an
update.
D
Okay,
perfect,
so
we
have
converted
from
apple
apps.
To
application.
Set
application
set
is
still
not
lga,
it's
going
to
be
released
in
alco
cd
0.2.
So
currently
I
have
pulled
from
the
dev
repo,
but
it
works
like
a
charm.
The
advantage
of
application
set.
Is
you
don't
need
to
really
create
an
app
of
apps
and
you
directly
just
create
an
application
set
you
create
a
matrix
and
what
the
matrix
does
is.
D
You
can
basically
say
what
applications
you
want
to
install
and
what's
the
matrix
of
all
the
systems
you
want
to
install
it
on
so
similar
to
production.
I
would
have
another.
This
is
basically
a
list
right,
so
you
can
have
production,
dev,
staging
so
on
and
so
forth.
C
D
So
the
matrix
is
up
to
you,
where
you
want
to
put
it.
You
can
create
that
matrix,
so
I've
currently
created
this
matrix
within
a
kubernetes
reaper.
So
whatever
is
kubernetes
infrastructure
I
put
into
the
matrix
and
that
matrix
basically
then
multiplies
whatever.
Is
there
your
with?
How
many
of
the
elements
are
here
so
assume?
I
create
this.
B
In
terms
of
you
know
like
when
we
use
home
file,
we
have
the
different
folders
for
different
environments
yeah.
How
does
the
different
values
work?
In
that
scenario,
if
you
understand
my
question.
D
Yes,
so
the
different
values
work
by
on
the
basis
of
folder
structure.
So
what
it
does
is
you
need
to
give
it
a
repo
url
and
a
repo
part.
So
what
I've
done
is
I've,
given
it
a
repo
url
and
a
repo
path.
So
the
repo
path
is
cuban
for
a
backup,
so
it
goes
into
backup.
It
goes
into
sees
that
it
has
one
chart
which
is
called
valero,
so
it
applies
only
one
chart.
Now,
in
the
case
of
ci
cd,
I
have
more
than
one
chart,
which
is
argo.
D
Events
are
for
rollouts
and
argo
workflows,
so
it
applies
all
three
of
them.
The
each
of
them
also
have
this
thing
called
production,
argo
events
values,
so
this
can
also
be
similarly.
D
Can
also
be
def
now
the
difference
between
the
two
and
the
advantage
of
having
this
is
now
based
on
the
cluster
name.
So
what
I've
done
is
the
destination
helm
value
files
is
taking
cluster
name,
the
part.base
name,
iphone
values.yaml.
So
what
that
means
is
it's
going
to
say.
Production
hyphen
in
the
case
of
this
particular
chart,
it's
going
to
say,
production,
hyphen,
r,
workflows,
hyphen
values.yaml,
which
is
why
it
picks
up
the
production,
yaml
file,
values
and
that's
how
it
ensures
that
it
kicks
off.
B
D
E
B
B
Okay,
just
just
so
I
understand
sorry,
the
cluster
called
div
will
correspond
to
the
the
prefix
before
the
dash
of
the
thing
correct.
Yes,
so.
D
C
A
C
Yeah
sorry
go
ahead,
so
what
happens
if
dev?
Let's
say
they
forgot
like
there's
no
qa
out
there,
but
you
define
a
qa
cluster.
Does
it
fall?
Does
it
throw
an
error
or
does
it
fall
back
to
like
a
different
like
base
base
name
like
a
default.
D
No
it'll
throw
an
error,
so
what
it
will
do
is
something
similar
to
this.
So
this
is
more
specific
to
our
workflows.
But
what
we'll
do
is
he'll
create
an
app
and
it'll
say
sync
error.
C
Okay,
all
right
so
there's
there's
in
in
this
there's
really
no
default.
You
you're
gonna
have
to
name
out
all
of
your
variables
that
you're
going
to
use
in
your
file
names.
D
D
So
I
just
showed
you
one
of
the
generators
right
which
is
list
now
you
have
two
generators
which
is
there's
another
called
git
generator
now
the
way
the
git
generator
works
is
one
git
generator.
I
showed
you
was
how
it
pulls
up
the
software,
but
there
is
another
type
of
kit
generator
where
you
put
the
cluster
config
within
this
particular
folder
structure.
So
I'm
not
done
that
right
now,
but
if
you
do
that,
what
happens?
D
D
B
What's
the
so,
just
just
for
my
understanding,
what
was
the
decision
like?
Why
did
you
make
the
decision
to
use
this
one
like
like?
Is
there
any
pros
or
cons
or
or
what
were
you
thinking?
D
Just
this
was
just
I
will,
in
the
long
run,
go
down
this
route
because
I
think
it's
better
to
have
a
completely
other
folder,
but
for
now
to
basically
ensure
that
it's
working,
so
I
had
one
of
the
issues
I
faced
was
helm
has
not
updated
its
crds,
so
I
had
to
go
and
update
the
crd
within
helm
separately
independently.
D
C
Cool
yeah
and
then
on
the
if
you
can
go
back
to
the
application
set
now,
if
you
scroll
up
a
little
bit,
let's
say
like
database
was
one
in
one
git
repo
and
the
dns
infra
is
in
a
separate
repo.
How
is
that?
Do
you
how's
that
handled?
Do
you
have
multiple
of
the
get
generators
then
like
on
line
10.
D
So
you
mean
no.
In
that
case,
I
would
create
multiple
application
sets
and
what
you
can
do
is
have
one
application
set
which
configures
a
particular
type
of
application.
So,
for
example,
I
would
have
a
separate
application
set
which
only
installs
database
a
separate
application
set,
which
only
sets
up
anything
else.
D
Lie
in
one
single
application
set.
C
Right,
but
coming
from
different
repos,
where
so
you
know
in
the
the
trend,
is
to
go
to
poly
repos
instead
of
mono
repo,
so
the
helm
charts
for
the
database
are
going
to
be
in
the
the
database.
Repo
and
the
security
in
this
case,
like
the
dns
infra,
will
have
its
hom
tur
over
and
it's
repo
you
know,
along
with
the
source
code.
So
you
know
you
have
the
source
code
plus
your
helm,
chart
in
in
a
repo
and
then
you
know,
and
you
have
one
for
each
microserver
set
up
that
way.
D
Can
use
multiple
git,
but
then
matrix
generator
gets
disabled,
so
matrix
is
enabled
only
for
one
kit
with
one
list,
so
you
can
do
two
things
you
can
have.
Two
application
sets
one
mainly
for
your
kubernetes
infrastructure
and
you
can
have
a
second
application
set
which
will
be
multi-repo
setup,
but
then
you
won't
be
able
to
use
this
particular
format
where
it
multiplies
the
application
generator
with
the
elements.
B
D
No,
it's
it's
the
same
thing.
The
issue
is
the
same
with
matrix,
so
currently
matrix
has
enabled
only
one
to
one
multiplication,
but
in
the
future,
most
probably
they
would
enable
multiple.
D
C
Yeah,
it
could
be
interesting
because,
basically,
the
way
it's
defined
now
you
still
have
an
app
of
apps
repo.
C
You
have
all
your
manifold,
all
your
home
charts
and
a
single
repo
and
that
you're
just
doing
a
matrix
against
that
you're
minimizing
the
you
know
how
you're
driving
it
through
the
matrix,
but
you
still
have
a
lot
of
files
that
you
have
to
manage
if
you,
okay,
so
it'd,
be
interesting
to
see
if
you're
able
to
pull
that
from
multiple
repositories
at
the
matrix
level,.
D
C
That'd
be
awesome
if
you
could
do
that,
because
you
know
just
from
what
I'm
seeing
is
we're
still
going
to
end
up
with
an
app
of
apps,
and
then
you
end
up
having
if
you
have
shared
microservices
across
multiple
applications
or
products.
If
you
have
a
assured
microservice
across
multiple
products
that
you
know
being
able
to
have
that
chart
in
one
place
will
definitely
simplify
things
instead
of
having
a
the
same
chart
and
multiple
app
of
app
repos.
D
G
B
D
D
C
C
C
So
right
in
here
I
just
want
to
give
if
you
can
cut
and
paste
in
the
example
from
the
application.
D
B
Can
I
can
copy.
C
All
right,
perfect,
and
from
last
time
we
there
was
some
discussion
on
a
slack
channel
on
on
the
discord
channel
about
argo
workflow.
I
do
you
foresee
argo
workflow
coming
into
any
of
the
integration
that
we
need
to
do
with
ortelius,
or
is
it
really
just
like
setting
up
a
pipeline?
You
know
run
once
you
get
into
the
whole
get
ops
world
and
you
hit
go
it's
going
to
go
off
and
perform
a
series
of
steps.
That's
my
understanding
is
that
correct.
D
So,
yes,
it
does
go
and
perform
steps
in
terms
of
installation,
but
arco
workflow
is
more
like
running
pipelines.
So,
for
example,
if
you
want
to
call
during
your
whole
process,
you
also
want
to
go
and
call
an
sp
bucket
or
say
it's
a
workflow
where
you
are
calling
a
cloud
function
which
then
in
turn
triggers
an
another
cloud
function
and
that
generates
a
file.
And
you
want
to
upload
that
into
an
s3
bucket
or
something
to
that.
Okay,
that
you
would
use
workflow
for.
D
Okay,
perfect
by
next
week,
I
would
probably
implement
argo
buffalo.
I've
already
started
working
on
argo,
workflow,
arco
events
and
the
our
rollouts
so
rollouts,
most
probably
people
might
not
use
if
they
go
down
the
service
mesh
route,
but
workflow,
I
foresee
a
high
chance
of
using
it
unless
they
just
prefer
to
stay
to
whatever
csd
pipeline
they're.
Currently
using
yeah.
D
C
So
there
are
some
extra
features
around
rollouts
that
are
take
it
to
the
the
like
helm
to
them
is
like
helm
on
steroids
is
what
I
could
gather.
What
rollouts
are.
D
So
the
difference
is:
if
you're
going
down
the
service
mesh
route,
though
you
don't
need
rollouts,
because
rollouts
does
pretty
much
the
same
thing
your
service
mesh
below,
which
is
blue,
green
canary
deployments.
All
those
things
are
service,
mesh
capable
your
service
mesh
will
take
care
of
it.
So
if
you
go
down
that
route,
then
you
won't
really
need
rollouts,
but
if
you
don't
want
to
service
mesh
or
you
don't
need
a
service
match,
rollouts
works
perfectly
well
and
it
adds
a
lot
of
benefits.
As
you
said,.
C
Got
it
and
then,
on
the
you
know,
events
over
workflow.
I
think
events
would
be
my
preference
on
the
doing
that
one
over
workflow,
just
because
ortelius
is
gonna
hook
into
the
event
engine
there.
I'm
getting
notified
that
a
basically
on
the
web
hook
side
getting
notified
that
a
deployment
has
completed.
D
C
Yeah
it
and
the
ones
that
we'll
be
setting
up
will
be
around
a
web
hook,
so
we
would
provide
a
web
hook
from
the
artillia
side
that
will
expose
over
to
argo
events,
and
then
what
we'll
really
need
to
see
is
what
the
payload
looks
like
that
the
events
are
throwing
over
to
us
and
see.
If
that
has
the
data
that
we
need
to,
you
know,
perform
actions
upon.
B
B
Yeah,
this
is
great
work.
I
imagine
you
had
some
late
nights,
I'm
working
this
out,
especially
how
it's
just
coming
out
as
well.
It
would
be
a
little
bit
more
difficult.
D
Yeah
so
like,
I
think
this
is
the
only
time
that
I
realized
the
value
of
customize
versus
helm,
because
the
helm,
files
or
artifacts
were
not
updated
up
to
the
latest,
the
same
way
how
the
manifest
files
work.
D
C
D
It's
not
based
on
templates
is
directly
based
on
kubernetes
manifest,
so
you
can
directly
reference
your
latest
manifest
without
really
any
modification
versus
the
charts.
I
would
have
to
create
my
own
chart,
which
is
what
I
did.
I
updated
the
crd
and
most
press.
I
can
so
I
I
had
to
tell
that
before
you
run
the
helm,
you
you
basically
update
the
chart.
I
can
show
you
what
I
did
most
specifically.
D
D
B
D
B
Yeah,
it
wasn't
enough
okay,
so
we're
saying
hooks
on
our
applications.
C
D
Yeah,
so
this
is
a
helm
file.
I
I
normally
use
the
initial
installation,
I
use
hand,
file
and
post
that
I
use
are
go
for
everything,
so
helm.
File
at
the
beginning
requires
a
couple
of
tools
which
is
external
dns,
so
that
arco
has
an
external
url
sort
manager,
because
without
certificates
it
doesn't
let
you
spin
up
a
load
balancer.
That
is
a
google
issue
along
with
that
application
set
and
argo
cd.
D
B
B
B
Azure
does
a
little
bit
different
to
gcp,
so
we
worked
out
how
to
do
it's
called
internal
advancer.
Do
you
have
that
one
in
there
emma
or
is
that
in
my
repo.
B
D
Is
the
one
that
it's
called
yeah
add-on.
B
D
Yeah
so
we
first
tried
to
use
the
internal
one,
but
the
internal
required
api
kp
and
we
were
like
no
that's
too
expensive.
What's
what's
the
other
option
and
that's
when
we
found
this
because
api
gateway
gets
very
expensive
very
fast
from
what
I've
noticed.
B
D
So
internally,
it
is
already
https
it's
pretty
much
similar
to
what
we
I've
shown
you.
It
just
shows
hdpu,
because
once
it
hits
the
load
balancer,
it
basically
goes
down
the
hpp
route,
but
that
does
post
the
load
balancer.
So
all
the
way
to
the
load
balancer.
It's
all
https.
B
C
Yeah,
I
I
I
think,
that's
a
a
good
thing
to
do.
D
Okay,
so
we
will
go
ahead
and
implement
this,
so
I
so
the
reason
I'm
doing
it
separately
is
first.
I
want
to
ensure
that
what
I'm
doing
is
the
right
direction
and
post
that
I
would
implement
within
artelias
perfect-
and
this
is
what
I
was
talking
about.
So
argo
has
its
own
health
charts
and
basically
the
crd,
as
you
can
see,
is
two
months
ago
versus.
B
D
Reference
use
this
separate
crd
for
this
thing
and
I
had
to
go
and
dump
this
particular
value.
Under
my
space,
yeah.
D
Yeah
yeah,
so
this
is
the
whole
crd.
It's
why
to
use
file.
It's
basically
replacing
this
current
crd,
which
is
there
got
it.
G
C
That
one
came
off
of
the
dev
branch
of
argo.
D
D
It's
as
I
said,
it's
not
yet
released
and
we're
going
to
be
releasing
it
by
end
of
this
month.
I
was
actually
chatting
with
a
couple
of
guys
on
argo,
so
till
then
we'll
have
to
wait,
and
even
after
then
I
might
post
probably
have
to
create
a
pull
request
to
this
particular
help
file,
so
that
I
update
the
crds.
Otherwise,
I'm
not
sure
when
it'll
get
done,
because,
most
probably
since
it's
a
community
based.
E
C
We
wrap
up
the
investigation,
the
next
part's
going
to
be.
You
know,
where
are
we
going
to
hook
in
ortelius
to
since
ortillius?
Has
this
metadata
that
it
can
push
into
the
get
repos
you
know?
Are
we
going
to
be
interacting
with
a
single
repo,
the
app
of
apps
repo,
or
are
we
going
to
be
interacting
with
multiple
poly
repo
structure
that
we
need
to
deal
with?
C
You
know:
that's
that's
going
to
be
kind
of
the
driving
forces
on
where
we
need
to
interact
with
the
get
repos,
and
then
you
know
where
what
in
those
repos
are
we
going
to
going
to
be
updating?
You
know
when
we
go
to
deploy
a
microservice,
we
know
what
the
the
microservice
got
built
over
on
this
on
this
poly
repo,
so
it
has
its
own
build
and
push
out
to
the
registry.
C
Now
we
have
a
new
image
tags
that
we
want
to
deploy.
Are
we
going
to
apply
that
to
a
branch
in
that
repo,
like
a
dev
branch
or
a
ci
branch,
or
are
we
going
to
take
and
apply
that
new
image
tag
over
to
the
app
of
app
repo
in
the
values
file
there?
So
those
are
types,
some
of
the
the
things
that
we
need
to
that
will
come
out
once
we
finish
up,
the
the
investigation
is
how
how
this
is
going
to
drive
and
then
we
get
into
also
the
pipeline
process.
C
You
know
as
we
go,
you
know
if
you
have
a
new
image
tag,
I
can
see
applying
that
to
the
dev
values
file,
but
you
know
how
do
we
handle
you
know
qa,
staging
and
production?
Do
we
come
back
around
and
is
that
an
automated
pipeline
process
using
the
argo
workflow
you
know,
so
we
have
a
little
bit
more
work
to
do.
D
D
If
failure
then
skip
because
otherwise,
there's
always
going
to
be
someone
who
has
to
manually
go
and
change
those
values.
C
Yeah,
that's
the
thing
we
want
to
avoid.
Is
you
know
having
people
edit
all
these
files
and
I
can
see
where
the
application
sets
are
going
to
definitely
help.
You
know
with
the
matrixing
capabilities,
but
we
want
to
minimize
human
interaction,
because
if
you
have
150
microservices-
and
you
know
you
have
10
teams
moving
them,
you
know
doing
updates
and
changes
multiple
times
a
day.
You
we
can't
have
a
there's.
You
can't
have
a
process,
that's
manual
yeah,
it's
just
not
feasible
to
scale
it.
D
Yeah,
I
completely
agree
the
one
slightly
simpler
solution
to
some
of
it
is
you
have
two
value
files?
One
is
a
common
value
file
and
one
is
environment
specific,
so
anytime,
it
succeeds
in
dev
post
that
that
value
should
move
into
comment
where
post
that
it
doesn't
matter
like
it
can
run
manually
into
everything
and
that's
fine.
D
C
Yeah,
the
funny
part
is
the
you
end
up
multiplying.
You
know
you
have
your
common
one.
Then
you
have
one
for
each
stage
of
the
life
cycle.
You
just
keep
on
multiplying
these
files.
C
E
C
That's
kind
of
the
thought
process
I
have
from
my
side,
like
I
said,
awesome
job.
The
next
couple
to
do's
should
get
us
pretty
close
to
be
able
to
start
pulling
together
and
posing
the
questions
of
how
we
can
integrate.
B
C
It
would
probably
be
from
both
sides
so
probably
come
from
the
us
and
and
your
side
as
well,
and
that
one's
gonna
be
a
little
bit
trickier
it'll
be
more
a
collaborative
design
document
that
we'll
need
to
pull
together
and
kind
of
route
it
around
a
bunch
of
different
eyes
to
make
you
know
in
front
of
a
bunch
of
people,
so
we
don't
miss
anything
in
the
different
use
cases.
That's
why
I'm
glad
you
you're
able
to
get
a
hold
of
sasha
and
sergio
with
with
their
use
cases.
C
B
C
Yeah
so,
and
what
we
probably
would
want
to
do
is
after
we
do
the
last
you
know
on
near
the
next
couple
weeks:
it's
probably
just
record
just
the
layout,
you
know
kind
of
just
something,
so
we
can
post
it
out
on
youtube
and
we
can
have
everybody
in
the
group
be
able
to
review
what
you
guys
have
done
and
just
give
them
a
level
set
through
a
video
yep
sure
that
sounds
good,
so
awesome
all
righty.
I
have
to
actually
run
off
to
another
meeting
today,
so
I.
B
C
It's
not
a
big
deal.
I
just
happen
to
have
a
I'm,
the
chief
of
our
volunteer
fire
department
and
in
the
village
that
I
am
is
that's.
In
my
district
one
of
the
villages
there
we're
going
to
build
a
new
senior
center
and
fire
the
station
there.
So
we're
going
to
retire
the
old
station
and
build
a
new
one.
Now
we
have
to
do
all
this
community
input.
It's
just
like
this.
C
It's
open
source
everybody
gets
put
there
and
put
in
yeah,
so
I
have
that
meeting
coming
up
for
everybody
to
provide
their
input
on
what
what's
going
to
happen
with
that
station.
So
it's
I
think
it's
gonna
be
like
a
three
million
dollar
building,
so
it's
gonna
be
it'll,
be
nice
to
have.
E
C
B
B
Pretty
we're
pretty
isolated,
I
mean
for
me,
it
doesn't
matter.
I
mean
I'll,
quite
happily,
just
sit
on
a
keyboard
for
you
know
two
weeks,
but
my
wife
is
a
different
story.
She
yeah
with
the
baby
as
well
so
yeah
that
they
they
let
you
go
once
we
get
a
blue
band.
It's
called
because
we've
passed
that
first
private
test.
They
have
one
on
day
one
one
on
day:
three,
then
they
give
you
a
blue
band
and
you
have
two
sessions
a
day
where
you
can
go
up
onto
the
roof.
C
C
B
B
Yeah
I'm
trying
to
get
I'm
trying
to
get
vaccinated
yeah,
I'm
over
here,
yeah
yeah.