►
From YouTube: ASP.NET Community Standup - Jan 21, 2020 - Cloud Native Crossover: Glenn Condron & Sourabh Shirhatti
Description
Join members from the ASP.NET teams for our community standup covering great community contributions for ASP.NET, ASP.NET Core, and more.
Community Links: https://www.theurlist.com/aspnet-standup-2020-01-21
A
B
B
B
A
B
Stupid
San
Diego
person,
where
we
don't
actually
use
coats.
Much
and
I
put
all
my
warm
clothes
out
when
I
was
packing
and
as
I
was
running
out,
the
door
I
left
my
coat
and
I
was
actually
going
through
security
at
the
airport.
I'm
like
where's
my
go
and
then
I'm
like.
Oh,
no,
so
I
had
no
coat
for
the
week
and
I
did
okay,
but
anyhow,
don't.
B
B
This
is
an
exciting
crossover
show
with
the
cloud
native
folks,
with
Shane
and
and
friends
so,
and
this
show
is
also
cool
because
it
was
by
request.
We
have
the
thing
that
shows
at
the
end
of
the
shows
where
you
can
send
in
your
request
for
show
suggestions,
and
we
got
you
know,
got
several
for
micro
services
and
containers
and
cloud
innate
everything.
So,
okay,
okay,
well
I,
have
just
a
few
links.
B
B
Do
that
see
if
I
remember
had
a
computer
over
here,
yeah,
okay,
wonderful
and
then
I
think
he
switched
yeah.
You
switched
the
buta
okay,
so
we
had
the
dotnet
can't
focus
on
blazer,
and
that
was
gosh
was
it
was
last
week,
so
that
was
a
we
have
dotnet
confit
in
September's
and
those
are
full
like
three
days
and
tons
of
content
and
community
speakers
from
around
the
world,
and
for
this
one
we
had
a
focus
event.
B
We're
just
focused
on
one
technology,
so
this
was
a
focus
on
blazer
all
kinds
of
new
stuff,
so
you
know
keynote
overview
from
from
Dan.
There
was
also
a
few
cool
announcement,
so
I'll
show
those
in
a
second
but
and
then
there
was
a
lot
of
great.
You
know
kind
of
deep
dive,
how-to
stuff
from
community
speakers
as
well.
You
know
deep
dive
things
on
authoring,
custom
components.
B
Jimmy
angstrom
did
a
really
cool
one
with
components
where
he
showed
showed
some
cool
stuff
with
like
CSS
and
managing
CSS
and
and
stuff
so
great
great
stuff.
There
yeah
two
of
the
two
of
the
kind
of
Microsoft
II
announcements.
Here
one
was
the
experimental
mobile
blazer
bindings.
So
this
is
this
is
one
of
those
crazy
crossover
sort
of
things.
This
is
building
xamarin
forms
applications
using
laser
binding
syntax.
So,
and
you
know,
I
spent
some
time
while
I
was
there.
Also
I
talked
to
James
Mont
Magnum
in
person.
B
You
know
he's
my
go-to
xamarin
forms
person,
and-
and
so
you
know,
the
general
idea
here
is.
If
you
are
a
web
developer,
you're
super
comfortable
with
with
blazer
razor
syntax
and
all
that
and
you
want
to
build
Dameron
forms.
Sam'l
is
not.
You
know
super
familiar
for
you
and
all
that
this
is
a
great
way
so
you're
still
building
it's,
it's
xamarin
forms
underneath
and
but
Blazer
syntax.
So,
and
this
is
totally
experimental,
people
are
asking
winces
chip
can
I,
go
live,
it's
like
we're,
just
thinking
about
it,
but
yeah.
B
All
right,
Steve
Sanderson,
the
Colton
he
showed
off
using
G
RPC
web
with
blazer
webassembly.
So
the
this
is
something
I
ran
into
and
I,
don't
know.
Glen
or
any
of
your
folks
wanted
to
talk
about
this
when
I
I
loved
the
demo
that
you
did
at
a.net,
coffee
and
then
I
was
speaking
somewhere
and
I
was
like.
Oh
I'll
quickly,
set
this
up
and
then
I
realized.
B
Yeah
so
so
this
is
a
neat
thing
here.
Also
another
kind
of
sticking
point
which
ER
PC
is
browser,
support
and
here
Steve's
using
G
RPC
web
with
blazer
webassembly,
so
yeah
walk
through
here
and
then
he
updated
the
flight
finder
app
and
his
code
is
in
here
as
well,
so
yeah.
So
that's
using
that
as
well.
So
here
he
walks
through
kind
of
simple
case,
and
then
he
also
links
over
and
and
he's
updated
his
flight
finder
application.
With
that
too.
A
B
C
B
C
A
C
B
I'm,
just
thinking
back,
there's,
Ajax
and
Jason.
All
you
know
like,
as
we've
gone
over
the
years
yep
all
right.
Let
me
went
through
these
a
little
faster.
This
one
I
actually
got
credentials
to
post
on
the
Visual,
Studio
blog
I'm,
so
proud
of
myself.
So
I
just
want
to
do
a
call
out
here
for
February
24
we're
doing
this
event
kind
of
like
that
blazer
one
with
Visual
Studio
for
Mac,
two
reasons:
I'm
mentioning
this
one
is
that's
that's.
B
My
new
focus
is
Visual
Studio
for
Mac
and
as
part
of
that,
a
good
kind
of
reason.
This
all
fits
together
for
me
is
that
Visual
Studio
for
Mac
has
added
in
a
lot
of
support
for
asp
net
core
lately,
so
dotnet
core
3
or
3/1
scaffolding.
There's
blazars
support,
there's
all
kinds
of
good
stuff
being
added.
B
So,
if
you're
developing
on
a
mac,
developing
asp
net
core
applications-
and
you
haven't
been
using
this
for
mac,
I
kind
of
recommend
it
so
and
that's
part
of
what
that
event
is,
this
will
show
why
at
ok,
cool
what
you
post
from
community
here,
one
is
sure
heads
doing
a
series-
s
peanut
core,
a
through
Z,
so
he's
already
because
we're
you
know
three
weeks
into
the
year,
he's
talking
about
cookies
and
cookies
and
consent.
So
he
talks
about
first
of
all,
the
the
previous
releases
asp
net
core
2x.
B
B
D
You
know
ghost
on
agar
since
before
after
had
resource
groups,
so
for
those
of
us
that
can
remember
how
long
it's
been
it's
well
over
eight
years
and
I
was
getting
to
the
point
where
it
was
just
super
hard
to
update
the
actual
application
and
we
can
kind
of
all
we
gonna
live
that
kind
of
life.
Now
that's
kind
of
the
lift
and
shift
to
containers
story
now
we're
maintaining
legacy.
D
Applications
is
really
difficult
and
I
think
I've
kind
of
jumped
into
this,
probably
I,
don't
know
four
or
five
times
over
the
last
at
least
five
years
going.
I
need
to
update
my
blog
back
in
or
the
version
of
it,
and
it
was
super
hard
for
me
to
kind
of
update
the
data
or
update
the
actual
app
and
I'm
going
every
time.
D
This
thing
updates
it's
a
pain
in
the
butt
and
finally
I
just
I
took
the
holiday
to
go
ahead
and
do
it
and
bite
the
bullet
and
now
kind
of
set
up
for
success
where
I
can
just
iterate
the
container
and
the
actual
application
portion
of
it
runs.
I've
got
my
data
and
my
sequel
that
runs
on
a
service
and
then
I've
got
all
the
everything
else
in
like
storage
and
kind
of
how
you
would
break
up
that
micro-services
or
kind
of
cognitive
architecture.
D
The
way
you
would
hope
it
would
all
work
and
not
break
down.
So
if
my
container
fails
I,
don't
care
I
can
just
you
know,
spin
up
a
new
one
and
everything's
gonna
still
work,
so
it
was
a
fun
experiment
and
it
all
it's
all
running
on
Azure
and
you
know,
got
it
all
see.
Icd
was
with
DevOps
and
now
learning
github
actions
and
all
these
things
to
to
try
and
move
to
that
new.
That
new
thing,
that's
cool,
nice,
cool.
B
All
right
and
then
one
other
thing
before
I
pass
it
over.
So
we're
doing
this
cross
server
show
Shane,
you've
started
up
this
cloud
native
show
this
link
and
all
the
links
will
be
shared
out
in
the
show,
notes
and
and
I'll
pass
them
out
over
the
chat.
And
then
this
short
link
here
to
the
cloud
native
show
you've
got
a
kms
cloud
native
show
right.
Yeah,
so
tell
us
about
the
show.
D
D
And
we
said
that
you
know
when
it
comes
to
cloud
native,
apps
kind
of
dotnet
dotnet
kora,
really
not
in
the
conversation
a
whole
lot.
So
how
do
we
kind
of
bring
focus
to
how
we
can
build
all
the
components
of
cloud
native
applications
and
and
kind
of,
what's
known
as
12
factor
architecture
with
all
other
things?
D
So
we
first
sat
down
with
you
know.
The
very
first
show
was
with
Brendan
burns
and
who
better
to
kind
of
describe.
You
know
what
is
what
is
cloud
native
and-
and
that
was
that
was
episode,
1
and
and
we're
up
to
at
least
8
published,
shows
right
now,
I've
got
a
couple
of
live,
shows
out
there
too
yeah
the
latest
one
we
did
with
with
Sir
Robin
talking
about
G
RPC
and
in
the
work
we're
doing
open
source
with
Google
and
making
it
happen
for
dotnet
core.
D
A
D
Based
on
the
last
couple
of
topics
so
we'll
implement
some
G
RPC
in
there,
we've
got
a
show
on
container
registry
coming
out
and
we'll
throw
that
in
there
as
well
and
just
kind
of
build
on
that,
and
it
really
it's
on
twitch
and
YouTube
and
various
places
where
we
kind
of
restream
all
these
things
and
just
kind
of
build
on
the
questions
and
make
it
interactive
for
a
couple
of
hours.
It's
a
lot
of
fun.
Yeah.
A
So
there
was
a
couple
of
questions
I
noticed
swing
by
in
in
various
places
someone
was
asking
about
when
we're
gonna
start
previewing,
don't
know
five
settle
down.
We
only
just
ship
three
point
one
but
but
the
answer
to
that
is
we
know.
We've
said
we're
gonna
ship,
don't
know
five
in
November,
so
you'll
probably
start
seeing
previews
of
that
dropping
soonish
and
we'll
have
probably
a
very
similar
number
of
previews
as
what
we
had
of
three.
A
So,
if
you're
looking
forward
to
whatever
you're
looking
forward
to
that's
going
to
come
in
two
on
at
five
it'll,
probably
not
be
too
far
away,
but
I,
don't
think
that
we've
actually
nailed
down
any
first
preview
date.
Yet
if
it's
not
I'm
sure
that
someone
in
the
chat
of
us
will
correct
me
instantly
so
that'll
be
fine.
D
C
So
I
have
a
couple
of
demos
to
show
today
they
mainly
involve
around
how
you
can
get
your
dotnet
app
running
in
a
kubernetes
cluster,
including
dabbing,
on
kubernetes
cluster.
What
I've
done
as
the
precursor
to
setup
for
this
show
is
I've
created
a
KS
cluster
I've
created
a
ACR
registry,
that's
attached
to
the
AKS
cluster
and
I've
set
up
ingress
and
nginx
ingress
in
my
cluster.
So
now
that
I
have
the
basic
set
up.
That's
pretty
much
all
I
need
to
deploy
a
web
app
and
start
routing
traffic
to
it.
C
So
I
want
to
show
you
my
local
workflow.
You
know
they're
they're,
multiple
open
source
tools
out
there
that
you
can
use
to
achieve
this
I
happen
to
have
a
personal
preference,
so
I'm
gonna
show
my
demos
predicated
on
using
that.
But
the
takeaway
from
this
theme
should
be
that
you
can
use
dartnet
in
be
productive
in
kubernetes,
and
there
are
a
lot
of
tools
out
there.
C
Already
that
help
you
be
productive,
so
I
have
D
is
like
a
design
doc
from
scaffold
up
on
my
screen
right
now,
because
scaffold
is
the
tool
of
choice
that
I
use.
The
reason
why
I
use
the
tool
here
is
when
you
go
from
source
control
and
building
and
making
changes
all
the
way
to
deploying
to
your
cluster.
There
are
a
couple
of
steps
that
everyone
goes
through.
You
start
with.
C
C
That's
why
tools
like
this
exist
so
I
had
this
architectural
diagram,
and
this
is
what
I'm
going
to
be
using
today,
but
you
know
I'm
using
scaffolds
or
other
tools,
I
think
tilt.
That
Deb
is
another
one
which
people
can
try
out.
So
let
me
go
ahead
and
jump
over
to
some
code
and
show
you
what
I
can
actually
do
so.
I
have
a
my
kubernetes
cluster
deployed
and
if
I
go
hit
it
right
now,
I
get
the
I,
don't
know
folks,
you
can
see
that
it
just
says
default
back-end
for
forms.
A
A
C
C
But
let
me
show
you
what
my
default
experience
looks
like
right
now,
so
I
drop
into
my
command
line
and
I
go
scaffold,
dev
all
right,
and
so
since
there
was
no
change
to
my
application,
it
already
found
the
image
locally,
and
so
now
it's
deploying
a
couple
of
things.
It's
deploying
my
dotnet
service
all
right
and
creating
the
associated
kubernetes
deployment
with
that
and
in
a
second
I
will
show
you
my
kubernetes
Hamel
to
say.
What's
going
on
and
I
also
have
a
ingress,
my
ingress
resources
being
reconfigured
without
traffic
to
that
endpoint.
A
C
A
C
Want
to
point
out
that
I,
if
you
see
to
come
out
at
an
error
and
scaffold
dev,
this
is
the
dev
workflow,
so
it's
also
actually
watching
for
changes.
So,
first
let
me
go
ahead
and
show
you
that
my
app
is
deployed,
and
just
so
you
believe
me
cool
right.
My
API
is
deployed,
wasn't
much
right.
Of
course,
there
were
some
config
files
dropped
in,
but
that
would
be
a
reusable
artefact
that
you
can
use
if
I
go
in
here
and,
let's
say
where's
my
controller,
my
weather
forecast
controller.
Let's
just
make
some
change.
C
A
A
A
A
C
So
so,
right
now,
let's
go
back
to
this.
I
made
a
change.
It
picked
up
my
changes
and
I
D
deployed
it
right.
So
what
exactly
did
I
deploy?
In
addition
to
my
app
and
my
app
artifacts,
there
were
some
infrastructures
code
that
I
mentioned.
So
let's
go
look
at
the
actual
kubernetes,
manifest
that
I
deployed
so
I
deployed.
C
C
There's
not
much
else
and
I
said:
hey
map
to
port
its
target
board,
880
inside
the
container
5,000
outside
for
the
load,
balancer
and
then
I
also
deployed
this
hello
world
ingress,
which
says:
hey
route
traffic,
add
slash
star
to
dotnet
service
at
port
5000.
That's
how
my
traffic's
going
doubt
it
and
you
see
if
I
go,
make
a
request
to
like
just
sanity
check
right.
If
I
make
the
request
to
route
the
app
get
404
because
there's
no
route
configured
there,
that's
pretty
straightforward!
C
Now,
once
you've
deployed
your
application,
there
are
a
couple
of
common
observability
concerns
that
you
want
to
solve,
because
there's
no
fun
having
an
application.
If
you
have
no
idea
what's
going
on
in
my
contrived
example,
there
was
there
was
nothing
wrong,
so
I
was
lucky
there,
but
you
probably
want
logs
and
metrics
right.
Those
are
the
like
the
two
pillars
of
observability.
So
let's
look
at
how
we'd
observe
that,
so
this
application,
as
as
you
can
see
from
my
command
line,
is
actually
spewing
a
bunch
of
Jason
like
if
I
go,
make
another
request.
C
There
is
a
log
collection
service,
so
there's
a
IKS
plugin
that
actually
runs
as
kubernetes
daemon
said,
that
runs
fluid
bed
and
it
actually
harvest
all
the
the
STD
out
logs
from
your
container
and
makes
it
available
in
Azure
monitor.
So,
if
I
jump
over
to
Azure
Monitor,
and
so
let
me
just
make
this
I
had
the
query
already
in.
So
let's
give
it
a
second
or
I
guess
6.8
seconds
to
be
percent,
you
can
see.
C
C
C
The
timestamp
extract
the
message
path
method
when
they're
available
stacktrace,
if
I
have
an
exception,
so
I've
been
able
to
get
this
structure
logging
available
in
Azure,
monitor,
which
is
great
now,
if
I
have,
if
I
have
like
errors
or
other
issues,
they're
clearly
visible
here
and
they're
structured
and
easily
searchable
in
the
way
that
I
would
want
my
logs
to
be
the
other
thing.
I
just
want
to
bring
some
attention
to
is.
There
are
also
these
span
ID
and
trace
ID
elements
right
now
they
are
unused.
C
But
since
we
support
distributed
tracing
built
into
the
box,
especially
the
w3c
format
for
distributed
tracing,
we
have
a
span
ID
in
Italy
s,
ID
associated
with
every
request.
Now,
if
I
were
to
have
multiple
applications
talking
to
each
other,
all
spewing
their
logs
into
the
same
log
analytics
workspace
I
could
track
as
a
quest
with
its
span
as
it
spans
across
multiple
apps
and
see
a
true
distributor,
trace.
Yeah.
A
C
C
B
C
C
So
so
I
did
show
you
logs.
The
other
thing
that
I
do
want
to
show.
You
is
metrics.
Cuz
logs
are
great
if
you
know
exactly
what
you're
looking
for,
but
if
you
have
spew
from
multiple
applications
in
a
single
log
analytics
workspace,
it's
a
little
hard
to
make
method
of
madness
right.
You
could
have
way
too
much
information,
so
metrics
are
great
as
they
give
you
early
visibility
into
what
is
going
on
into
your
application,
like
you
know,
as
darknet
of
developers
might
be
aware.
C
They're,
like
you
know
the
you,
the
perf
counters,
from
the
good
old
days
that
you
knew
if
those
counters
went
out
of
some
standard
deviation,
you
had
a
problem
on
your
hands
and
it
was
coming
soon.
So
the
other
thing
we'd
really
want
to
do
is
make
sure
that
counters
are
available
so
before
I
show
you
the
demo
just
a
little.
Like
brief
background,
so
we've
made
some
improvements
in
dotnet
over
three,
oh
and
three
one.
C
Subsequently,
where
we've
exposed
diagnostic
information
from
the
runtime
like
metric
based
information
available,
we
our
new
parameters,
called
event
counters.
Now
these
event
counters
can
be
consumed
both
in
process
via
the
event
listener,
API,
which
I'll
get
to
in
a
second.
They
can
also
be
consumed
out
of
process
and
that's
how
tools
like
darknet
counters
works,
for
example.
So
let
me
switch
back
over
to
my
application
and
tell
you
how
I've
actually
consumed
event
counters.
So
what
I
have
here
is
I
want.
C
I
want
everyone
looking
at
this
to
squint
a
little
hard
right
now
and
pretend
like
this
is
the
ideal
world.
Obviously
this
is
demo-
and
you
know
things
are
obviously
subject
to
change,
but
this
is
how
it
perhaps
could
be.
So
if
I
show
you
my
startup
file
and
had
added
something
to
configure
services,
it
says
services,
dot,
add
matrix,
so
I've,
actually,
you
know
add
my
own
librarian
here
that
does
some
stuff
in
it
like,
which
is
exposed.
Matrix
is
a
service
in
your
application.
The
other
thing
that
I've
done
in
my.
C
If
you
look
at
what
I've
configured
endpoint
routing
is
I've,
said:
endpoint
start
map
metrics
endpoint.
So
what
I've
also
done
is
this
thing
isn't
configured
right
now,
but
add
a
known
location,
which
is
slash.
Metrics
I've
exposed
the
metrics
produced
by
that
application.
Okay,
so,
given
that
my
app
is
already
deployed,
if
I
go
back
over
here
and
I,
make
a
request
to
the
slash
metrics
endpoint,
any
guesses,
for
what
you
see
logs
no
way
glad
these
metrics.
C
If
you
keep
refreshing
it,
so
there's
numbers
can
change
every
time,
so
these
so
I
I
cashed
them
every
time,
they're,
synthesized
and
they're
synthesized
once
a
second,
because
that's
how
I've
hard
coded
in
my
app
right
now
so
every
time
I
do
make
a
request.
You
will
see
that
there
are
new
values
available,
yep
I
can
see.
The
allocation
changes
is.
B
C
C
A
D
C
D
C
That
is
a
great
question
and
that's
something
we
have
to
figure
out,
because
there
are
existing
libraries
in
the
ecosystem.
There
is
a
Prometheus
dotnet
library
that
has
significant
usage,
and
you
know
it
does
offer
some
things
that
this
POC
does,
and
it
has
stuff
like
histogram
based
counters
I'll.
Give
an
example
of
a
metric
like
request,
latency
and
average,
or
instantaneous
value,
is
kind
of
a
useless
measure.
What
you
really
want
is
like
P,
75,
P,
95,
p
99
right,
because
you
want
to
find
outliers
in
your
request
latency.
C
So
there
are
some
things
we're
looking
to
do
even
up
like
including
changes
to
the
runtime
to
explores
this
richer
fidelity
of
information,
but
the
Prometheus
dartnet
library,
as
an
example,
uses
I,
believe
uses
diagnostic
source
or
or
event
source
to
synthesize
it
sorry
using
that
source.
So
they
can
expose
a
richer
fidelity
of
information
that
my
PFC
doesn't
II.
The
compelling
thing
about
what
I
showed
right
now
is:
it
has
no
external
dependencies.
It's
all.
It
just
uses
things
in
the
BCL
or
in
SP
net
core
yeah.
A
A
D
This
is
a
gives
us
a
simple
demo
right.
This
is
probably
the
equivalent
of
a
hello
world
example
when
it
comes
to
these
types
of
this
type
of
data,
I'm
curious
and
have
you
thought
maybe
consider
I
forget
what
the
actual
term
is,
but
when
we
start
to
watch
something
it
starts
to
slow
it
down.
Like
you
know,
how
are
we
can
be
considered?
D
C
So
we're
working
our
way
from
the
bottom
up
right,
so
you
want
to
make
sure
synthesizing
the
metrics
and
the
runtime
aren't
expensive,
because
you
want
to
start
from
the
bottom.
Then
work
your
way
up
into
the
web,
app
stack
and
what
the
cost
of
serving
this
is.
So
we
have
looked
at
the
runtime
fun
fact:
II
the
tech
and
power
like
benchmarks
that
we
run
internally
to
track.
Our
progress
actually
collects
these
metrics
for
every
app
run
that
we
do
and
we
do
not.
C
It
was
within
observation
error
for
the
impact
of
synthesizing,
these
metrics
okay,
but
those
were
synthesized
and
egressed
using
the
out
of
process.
Egress
mechanism
not
be
an
endpoint
of
the
application.
So
to
answer
your
question,
we
absolutely
are
thinking
about
it,
but
we're
starting
at
the
bottom,
we're
starting
in
the
runtime
and
we'll
work
our
way
up
to
see
the
impacts
there.
A
Were
also
otherwise
to
expose
these
types
of
your
eyes
that
are
potentially
interesting
right,
like
you
could
imagine
sidecar
containers
that
have
this
endpoint
on
them
that
aren't
actually
part
of
your
application,
for
example,
as
a
completely
alternate
way
of
doing
a
very
similar
thing,
but
a
very,
very
different
implementation
right,
so
we're
still
talking
about
it,
we'll
see
we'll
see
how
it
lands.
Yes,
sir.
C
So,
jumping
back
to
the
demo,
so
I've
shown
you
how
my
app
actually
exposes
this,
but
it's
really
no
fun
unless
it's
persisted
somewhere
right.
This
is
just
an
endpoint
that
you
have
to
hit.
So
the
thing
that
John's
question
did
lead
to,
and
I
did
mention
before
is
this
format
is
the
same
format
that
Prometheus
expects
so
funnily
enough,
eks
has
a
way
of
configuring
it
to
scrape
Prometheus
style
metrics.
C
So
if
I
jump
back
over
to
my
application
and
I'm
going
to
show
you
the
the
kubernetes
config
for
the
I
believe
this
is
the
azure
monitoring
agent
config.
So
I
apologize
to
add
your
monitor
folks,
if
I'm,
showing
the
old
and
ugly
and
outdated
way
of
doing
this,
and
they
might
be
a
cleaner
way,
but
right
now,
I've
done
the
the
infallible
way
which
is
drop
a
config
map,
and
this
is
he
can
take
map
used
by
the
scraping
agent,
so
we'll
skip
over
stuff
that
doesn't
really
matter.
C
The
interesting
theme
would
be
this
line
of
configuration.
Glen
will
yell
if
it's
not
large
enough
for
people
to
read
mm-hmm
but
I
said
so.
It's
just
an
array
of
kubernetes
services
to
script
metrics
from
and
it's
my
app
my
service
name
in
kubernetes
landed
service
name.
Dot.
Namespace
is
how
you
can
route
to
anything
else.
C
So
even
now
that
we've
done
this
something
I
did
not
mention
is
I
showed
you
when
I
deployed
my
app
there
was
a
deployment
file
in
the
ingress
file.
The
other
thing,
of
course,
that
I
did
deploy
was
this
config
map
for
the
scraping
agent.
So
if
I
switch
back
over
to
my
terminal
and
see
the
logs,
you
might
have
seen
there's
a
lot
of
spew
on
my
screen
and
that
is
coming
from
these
slash
metrics
HTTP
GET.
So
what
I've
done
so
since
I've
deployed
to
config
map
the
the
azure
Monitor
agent
is
scraping.
B
A
C
C
So
the
if
you
see
here
and
I
know
it's
not
the
best
thing.
The
metric
of
queried
for
is
requests
per
second,
and
you
see
this
nice
spike
in
the
chart,
because
right
before
I
walked
into
this
room.
I
pointed
a
Lourdes
generator
at
it.
So
I
knew
this
graph
would
have
a
right
like
a
nice
spike,
but
the
the
other
thing
that's
available
is
is
not
just
requests
per
second,
if
I.
In
fact,
if
I
go.
C
C
I'm
basically
I'm
collecting
all
these
metrics
I
had
shown
you
one
visualization,
but
you
can
do
fun
stuff
in
as
your
monitor,
like
you
know,
obviously
one
is
observed
the
values
but
set
alerting
on
it.
You
can
set
yourself,
you
know,
send
yourself
an
email
if
your
thread
pool
like
Q
lakhs
becomes
too
high,
or
something
like
that
right
and
dig
reactive
measures
based
on
okay.
A
So,
let's
walk
back
a
little
bit
from
this.
We.
What
we
did
here
is
we
put
we
added
to
the
application,
a
Diagnostics
endpoint,
an
endpoint
that
just
has
the
metrics
in
there
right,
and
it
is
the
format
that
Prometheus
expects
to
get
is
the
Prometheus
format.
So
if
you
were
to
point
at
Prometheus
instance
at
that
URI,
it
would
work
the
same
that
we
did
the
same
as
what
this
demo
did.
A
Then
we
did
some
basically
some
aks
stuff
that
anybody
has
not
been
using
aks
before
I,
probably
way,
no
or
some
community
stuff
I
should
say
to
get
the
to
get
an
agent
to
hit
that
endpoint
and
put
it
into
Azure.
Monitor
are
the
same
system
that
has
all
of
our
logs
in
it.
That's
what
all
that
demo
was
so
all
of
our
stuff
in
the
middle,
that's
kind
of
some
people
were
like
well,
I,
don't
really
understand.
A
A
lot
of
this
was
really
just
a
long
way
of
like
adding
the
line
of
code
to
kubernetes.
I
said
hit
this
URL
and
grab
the
metrics
from
it
and
put
them
into
our
monitor.
For
me,
that
all
makes
sense
is
that
sound
right?
Yes,
and
so,
if
you
were
trying
to
do
this
cross-platform
because
I
see
some
people
in
chat
who
were
like
okay,
this
is
like
this
is
super
specific
to
Asia.
A
The
everything
that
you
did
is
exactly
the
same
if
you
would
have
just
be
using
Prometheus
and
various
other
like
common
common
technologies,
because
that
is
the
format
like
the
metric.
The
metric
endpoint
is
the
same
format
as
the
Prometheus,
the
Prometheus
expects,
and
as
such,
it's
all
the
same,
except
in
the
bit
where
we
had
it
had
one
that
one
line
in
a
config
file
that
said
hit.
This
endpoint
would
be
in
a
Prometheus
specific
configuration
rather
than
somewhere
else,
actually
to
speak
to
that,
I
also
showed.
C
A
C
A
The
goal
the
goal
at
the
end
of
a
lot
of
work
would
be
the
first
part
where
he
just
runs
the
runs,
the
scaffold
command
and
the
app
is
running,
and
then
he
can.
He
just
hit
this
and
see
all
the
things
without
changing
is
happening
further.
That
would
be
the
old
ideal
in
state
and
we'll
see
how
close
we
can
get
to
that
over
time.
A
C
I
do
want
it
sure
I
just
want
to
point
to
this
repo.
It's
get
help
slash
my
name,
slash
metrics,
because,
specifically
because
the
agile
question
was
asked,
this
is
just
a
dart
net
app
that
exposes
a
matrix
endpoint
with
Prometheus
configured
to
run
in
the
cluster,
and
this
is
the
Prometheus
UI
showing
the
value
of
my
counters.
C
D
We
see
this
as
a
glance,
or
do
we
see
this
as
a
kind
of
an
inner
loop
development
cycle
for
just
any
project,
or
do
you
see
this
somewhere?
Where
do
you
see
this
in
the
developers
line
cycle
I'm,
looking
through
the
scaffold
tool
and
I'm
like
hey?
This
is
something
I
have
to
have
like
in
my
my
tool
chain.
Now
you
know
outside
of
BS
or
BS
code.
D
C
So
I
can
take
a
stab
at
answering
this,
but
I
will
preface
this
with.
This
is
my
opinion
and
not
team's
opinion.
What
I
do
believe
that
these
tools,
you
can
be
productive
with
them
today,
but
the
concept
count
is
way
too
high
and
something
unreasonable
through
a
developer
whose
passion
is
in
kubernetes.
If
your
passion
is
just
like,
hey
I
want
to
finish
my
project.
Kubernetes
is
a
means
to
the
end.
I
do
not
care
about
all
the
stuff,
so
I.
B
D
Yeah,
that's
a
good
point.
I
think
that
in
thank
you
for
answering
it
that
way,
because
I
think
we
find
that
there
are
on
the
spawn,
honest,
respectable
developers
there.
There
are
developers
who
want
to
just
check
in
code
and
then
there's
developers
like
you're
mentioning
there
were
I
want
to
know
all
the
nuts
and
bolts
and
I
don't
care.
If
there's
a
hundred
tools
to
get
my
job
done,
I'll
figure
out
how
to
chain
them
together.
So
this
probably
is
closer
towards
that
end,
as
opposed
to
check
it
in
just
check
it.
A
I
only
want
to
focus
on
my
code,
but
I
know
that
I
have
this
other
service
or
somebody
else
builds
or
maybe
I
need
this
database
running
or
I
need
this
thing
over
here
and
once
you
start
operating
in
that
ecosystem.
How
are
you
doing
it?
Are
you
doing
it
on
you?
Are
you
doing
that
via
having
your
own
IKS
cluster?
Where
you
stand
everything
up
and
then
you
Dev
on
it?
A
Well
then,
something
like
scaffold
sounds
super
amazing,
so
you
leave
all
the
other
stuff
running
and
you
just
dev
on
the
bits
that
you
wanted
and
you
work
on
the
bits
you
want
to
work
on
or
are
you
gonna
run
them
on
your
local
machine
or
are
you
gonna
do
like
a
bunch
of
things
right?
So
what
you
want
to
do
to
manage
all
those
dependencies
is
different
for
everybody
as
well,
and
it
can
be
hard
set
up
and
it's
not
unique
to
kubernetes
or
to
micro
services.
A
They
just
exacerbate
a
problem
that
can
exist
I've
seen
other
plenty
of
other
organizations
using
you
know,
don't
have
framework
on
IAS
on
Windows
servers
with
asp.net
and
have
the
same
problem,
because
they
have
multiple
teams
and
perhaps
multiple
websites
that
make
up
one
like
solution
and
they
have
the
same
problem.
Sure
sure.
B
So
you
know
there's
some
some
chat
reaction,
that's
like
wow!
This
is
difficult
to
debug
and
you
know
to
understand,
like
you
need
a
lot
of
knowledge
to
be
able
to
run
and
handle
this.
Is
this
something
where,
over
time,
this
gets
like
with
the
kind
of
work
you're
doing
with
metrics
and
just
kind
of
in
general,
you
think
things
will
get
simpler
for
for
dotnet
developers
or
is
it
always
just
gonna?
Be
you
need
to
know
all
the
stuff?
Yes,.
D
Specific
I
think
the
language
agnostic,
even
if
we
did
this
whole
example
and
go
or
notor
or
whatever
it
was
I
think
that
the
complexity
of
getting
this
all
set
up
and
working
is
probably
about
the
same
yeah.
Yes,
so
I
think
tooling
in
general,
across
kubernetes
development,
we
know
as
an
industry
is
not
it's
not
easy.
It's
it's
certainly
not
there
yet
and
knowing
that
when
stuff
breaks,
that
is
always
our
biggest
challenge.
You
know,
how
do
we
do
bug?
D
A
The
I
think
of
them
overall
I
mean,
if
we're
talking
about
predictions
that
could
be
entirely
and
wildly
inaccurate
over
time.
If
you
talk
it
like
humanities,
Kampf
or
something
like
that,
you
talked
about
it
was
there
was
a
lot
a
couple
of
years
ago
about
running
communities
and
doing
everything
in
kubernetes
and
how
kubernetes
works
then
like
last,
then
there
was
a
year
or
so
ago.
There
was
a
lot
about
our
managed.
Kubernetes
will
have
managed
companies
instances.
You
won't
have
to
worry
about
machines,
about
clusters
as
much
I.
A
Think
more
and
more
as
we
go
forwards,
we'll
start
to
see.
Kubernetes
fade
more
and
more
into
it
as
a
detail
and
implement
detail.
An
implementation
detail
of
the
background
and
you're
starting
to
see
that
more
and
more
and
it'll
be
more
like
serverless
was
more
like
functions
was
something
more
like
a
more
like
an
app
service
that
runs,
and
that
happens,
to
use
the
power
of
communities
to
give
you
a
more
powerful
app
service.
A
A
C
I
do
want
to
answer
John's
question
with
actually
a
Sam
I
mean
a
demo,
because
it
is
an
example
of
something
of
like
debugging,
specifically
like.
Oh,
my
god,
debugging
the
cluster
is
really
hard
right,
but
with
the
right
tool
it
could
become
like
commoditized
and
as
easy
as
just
attaching
to
debugger
anywhere
else.
I
want
to
show
you
a
demo
off
there
working
with
existing,
tooling
and
right.
So
that
might
help
answer
the.
C
We're
heading
so
so,
switching
back
over
to
my
screen
we
switch
over.
Did
this
me
just
I
should
have
cleared
out
my
app
before
I
asked
you
to
switch
line
right
so
I
have
another
application
running
here.
All
right,
I'm,
just
gonna
show
you
this.
So
this
is
a
little
more
than
that
example.
It
has
a
back
end
and
the
front
end
and
I
believe
has
it
at
its
cluster
as
well.
So
this
is
great.
You
I
deploy
this
to
kubernetes,
but
then
I
want
to
debug,
and
that
becomes
a
little
challenging
right.
So
how?
C
How
would
I
do
bugs
I'm
gonna
show
you
how
I
actually
went
through
this
manually?
I
went
on
a
self-discovery
journey
and
then
I
realized.
Oh,
wait,
there's
a
tool
that
exists.
That
lets
me
do
this
with
one
click
so
to
to
actually
debug
you're
down
in
an
application.
It's
pretty
similar
to
the
remote
debugging
experience
to
be
as
or
anywhere
else
where
we
actually
have
a
process
running
alongside
the
target
process
that
exposes
like
either
a
web
endpoint
or
like
TTY
that
we
can
connect
to
and
send
commands.
C
So
if
I
go
back
over
my
scaffold,
Yama
file,
which
I
don't
think,
I
showed
in
much
detail
so
I,
don't
this
custom
profile
that
says
hey
instead
of
building
this
darker
file
and
tagging
with
this
image
use
this
alternative
darker
file,
which
I've
called
debugger
doctor
file,
so
I
can
always
run
like
scaffold.
Dev
and
I
can
specify
a
profile
and
I
call
this
the
debug
profile.
Let
me
show
you
what
it
does.
So,
let's
go
back
over
into
my
app
and
I.
C
Have
this
debug
docker
file,
so
in
addition
to
building
my
app,
which
is
the
fun
stuff,
let's
actually
look
at
the
last
couple
of
lines
here,
so
there's
the
install
dependencies
for
visual
studio,
remote,
debugger
and
then
actually
add
the
add
vs
dbg
inside
my
container
alright.
So
now,
when
you
run
your
application,
there's
another
VSD
in
the
container
that
I
can
connect.
So
and
now
let
me
show
you
my
so.
This
is
perhaps
uninteresting
in
if
I
go
too
much
in
the
weeds,
I.
A
C
The
only
difference
was
I'm
installing
the
dependency
yeah
yep.
So
if
you
have
this
in
your
doctor,
file,
I
did
a
separate
profile.
If
you
think
it's
benign
to
bake
it
into
your
darker
image,
that's
also
possibility
so
I'm
gonna
jump
over
to
my
other,
a
please.
This
is
manual
and
convoluted.
So
there
is
a
visual
studio,
cord
extension
by
the
fine
folks
over
at
Google
called
cloud
cord
right,
and
it
works
well
with
scaffold.
So
I
have
another
gap.
This
is
the
application
I
was
talking
about,
which
was
front
and
back
end
in
Redis.
C
So
if
I
look
at
my
scaffold
file,
it
has
all
my
manifest
in
here
right,
so
I'm
gonna
go
over
here
and
instead
of
running
scaffold
from
the
command
line
like
I've
been
showing
you
thus
far,
I'm
going
to
use
the
vs
core
tooling
in
here.
So
let
me
just
make
sure
my
kubernetes
cluster
is
in
the
right
state.
If
I
get
that
the
default,
404
I
know,
I
can
proceed
actually
believe
my
app
is
still
running
somewhere.
I.
C
D
A
Yeah,
when
you'd
use
the
docker
tools
in
Visual
Studio
to
kind
of
f5
into
docker,
to
run
in
your
app
in
the
doctor
container,
it
does
it
it
volume
and
stir
the
bugger
up
from
the
local
machine
into
the
container
to
get
the
same
effect
as
what
Seurat
was
showing
by
building
an
image
and
as
a
debugger
in
it.
Yeah.
A
C
Thing
I
was
trying
to
do
is
run
two
different
demos
in
the
same
cluster
and
I
forgot
to
tear
down
my
resources,
so
I
didn't
what
I
like
have
my
next
deployment
stomp
over
my
old
one,
so
got
that
done.
So,
if
you
see
down
here
in
the
corner,
I
have
this
cloud
code.
So
I'm
gonna
go
ahead.
Click
on
this
and
it
says:
there's
like
a
new
application
and
there's
a
deploy.
C
Application
in
fly,
application
continuously,
that's
equivalent
of
typing
scat
for
dab
in
the
command
line,
right,
side,
hips
and
then
I
choose
the
default
profile.
I'll
speak
to
the
other
one
in
a
second.
So
it's
just
ignore
this
for
a
second.
So
right
now
it's
doing
the
same
thing
as
running
scaffold
in
the
terminal,
but
there
are
a
couple
of
neat
things
that
it
does
expose.
So
you
see
the
0
out
of
5
I
had
5
different
services
to
deploy.
So
it's
all
in
the
process
of
deploying.
So
let's
give
it
a
second.
It.
C
C
A
C
C
A
A
B
A
The
question
right
and
different
organizations
have
different
answers.
I
will
assert
that
if
you're
running,
like
don't
have
framework
on
Windows,
you've
always
got
a
bunch
of
stuff
like
that
available
on
the
machine.
There
are
probable.
There
are
potentially
policies.
You've
got
that.
Have
luck,
then
lock
them
down
on
those
machines
but
they're
kind
of
there
and
yeah.
A
That's
a
question
that
you
have
to
ask
yourself:
it's
super
like
situational
and
specific
to
different
people
on
whether
or
not
they're,
whether
or
not
the
benefit
of
being
able
to
randomly
debug
is
useful
and
beneficial
to
them.
In
general,
most
places
I've
worked
in
the
past
would
say
no
and
don't
debug
in
production,
but
you
know
if
you're,
using
the
cluster
as
ad.
If
this
is
a
dev
cluster,
then
absolutely
so
well,.
B
A
A
C
A
Yeah
and
one
by
guess,
the
answer
that
I
gave
was
more
about
once
you
do,
that.
How
do
you
get
traffic
there
and
also
use
the
other
instances?
And
that's
where
that's
where
things
like
those
spaces
come
in,
which
is
like
how
do
I
spin
up
a
new
dev
container
in
my
production
cluster
and
have
just
my
traffic
from
my
test
flow?
Go
to
it
and
everybody
else's
production
traffic
go
to
the
production
instances,
but.
C
A
B
C
Alright,
so
I
start
my
deployment
and
it
did
deploy,
took
awhile.
There
were
five
services
right.
This
is
up
and
I
can
hit
my
app.
This
is
great
right,
but
you
guys
want
to
see
the
the
magic
moment,
which
is
how
can
I
be
bug
in
there.
So
let
me
actually
first
jump
over
to
my
cord,
actually
set
a
breakpoint
somewhere
Oh
funny
enough
I
had
this
breakpoint
and
I'm
talking
on
okay,
so
I'm
going
to
f5
I
should
probably
show
you
what
I
hit
up
on
that
five.
C
A
A
A
A
C
A
A
I,
go
ahead,
attach
the
pin
one
so
there's
a
lot
of
magic
here
then.
So
this
thing
knows
how
to
find
its
way
to
his
cluster,
find
the
service
find
the
front
end
pod
and
to
get
they
get
the
information
of
the
processes
that
are
running
it
and
then
ask
you
which
one
you
want
to
attach
to
and
then
let
you
attach
and
then
hit
a
breakpoint
in
a
cluster
running
on
Azure
from
your
local
machine
when
traffic
hits
it
so.
C
My
wake
point
should
be
hit
it's
taking
a
while.
So
I
guess
that's
a
good
thing.
If
not,
you
just
have
to
believe
me
on
this
to
Ali.
When
you
add
a
comment,
it
hits
there
right.
So
we
did
I
didn't
say
you
did
no
I
I
just
saw
I've
made
a
request.
Oh
there
you
go
and
my
big
point
got
hit
so
so
Mike
source
code
is
local
deployed
it
into
an
aqueous
cluster
set
a
breakpoint
in
my
local
dev
experience
went
and
made
a
request
to
my
public
endpoint
right.
C
A
C
A
A
Broader
thanks
for
out
that
extension
that
drew
Jared
just
showed
at
the
end
was
from
Google
and,
as
we
saw
syrup,
went
all
through
all
this
effort
of
not
awful
lots
of
effort.
He
went.
He
went
through
the
effort
of
creating
scaffold,
creating
a
debug
docker
profile
may
having
this
magic
command.
They
could
let
him
attach
and
then
realize
that,
oh
so
much
from
Google
it
solved
this
in
an
interesting
way
with
a
BS
code
extension.
Let
me
just
go
use
that
right.
Sometimes
moments
like
that
happen.
C
A
For
kubernetes,
so
first
the
first
I
will
answer
a
question
that
you
didn't
ask
which
is
cloud
native
in
general.
You
should
go
wash
out,
show
it'll
get
into
it.
It'll
get
into
more
communities
and
you'll
also
see
what
the
person
who
helped
invent
kubernetes
looks
like
has
Brendan
burns
episode,
4
kubernetes
in
general,
the
AKS.
A
D
D
In
the
show,
but
on
on
the
cloud
native
show,
Channel
Brendan
Burns
has
got
a
lot
of
kind
of
five
to
eight
minute
stuff,
so
it's
kind
of
101
on
what
is
kubernetes.
How
does
it
work
and
kind
of
just
general
knowledge
about
kubernetes
to
that
he's
done
some
cool
things
on
a
lightboard.
That
is
just
good
general
knowledge
on
what
is
a
cluster?
What's
a
pod,
what
is
you
know,
orchestration
things
like
that,
so
just
really
good
just
base
knowledge,
not
just
the
the
deep
dive
topics
but
in
general
stuff
estate
is
good.
D
A
And
I
mean
we
kind
of
shotgunned
it
here
we
jumped
around
the
place.
We
showed
a
lot
of
important
things,
some
details
that
people
will
be
over
people's
heads
I
guess
we
could
do
a
more
101
style
community
stand-up.
If
people
want
that
they
should
tow
Galloway,
they
made
the
community
managed
to
make
this
episode
happen.
So
yeah.