►
From YouTube: The Level Up Hour (E40): Using Serverless to clean URLs
Description
Sometimes marketers have really long, ugly URLs. This is an excellent way of passing data, but sometimes these long URLs are hard to read and share. Wouldn't it be great if we had a quick way to clean those URLs? Join us for another episode of the Level Up Hour as we try to figure out how to build and deploy a serverless function in OpenShift that can strip the tracking and marketing information from a URL.
Follow us:
Langdon White https://twitter.com/1angdon
Chris Short https://twitter.com/ChrisShort
What is the Level Up Hour?
Every Wednesday at 9am ET we stream live demos on how Red Hat Enterprise Linux (RHEL) users and admins can benefit their organizations and improve their careers by learning how to use containers, Kubernetes, and Red Hat OpenShift.
Learn more at:
https://red.ht/leveluphour
http://red.ht/3ayHlHI
How to participate in the The Level Up Hour:
Watch on YouTube every Wednesday at 9amET
Chat with us on Discord https://discord.gg/5VMVGJt
Check out our GitHub repo https://github.com/level-up-hour/
About OpenShift:
Red Hat OpenShift is an open source container application platform based on the Kubernetes container orchestrator for enterprise application development and deployment.
https://www.openshift.com/
https://github.com/openshift/
#RedHat #OpenShift #tluh
A
A
Good
morning,
good
afternoon,
good
evening,
wherever
you're
hailing
from
welcome
to
the
level
up
hour
here
on
openshift
tv,
I
am
chris
short
executive
producer
host
with
the
most
whatever
you
want
to
call
me
of
open
shift
tv.
I
am
joined
by
the
one
and
only
the
illustrious
langdon
white
langdon.
How
are
you
this
morning.
B
A
A
B
B
B
A
B
B
A
B
My
head,
yes,
so
you
can
also
find
us
on
discord.
You
know
join
our
discord
channel,
which
is
in
the
chat
as
well
as
on
the
screen,
and
then,
let's
see
for
today's
show,
we
are
going
to
be
focusing
a
little
bit
on
using
openshift
serverless,
primarily
as
a
like.
I
was
fooling
around
with
something
and
realized
that
this
was
a
great
example
of
a
serverless
function
and
why
a
serverless
function
is
useful,
but
there's
a
lot
more
to
or
arguably
more
to
serverless.
B
B
B
Ready
alrighty,
so
that
is
the
show
notes
you
know,
and
you
can
also
you
should
find
them
on
twitter.
I
think
I
tweeted
out
about
them
last
week.
I
think
so
you
can
find
them
up
there
too,
and
that
is
the
end
of
the
sharing
time
of
slides.
We
do
have
you
know
some
shows
we
tend
to
focus
more
on
interviews.
Some
shows
we
tend
to
focus
more
on
kind
of
actual
code
or
demos
or
whatever
I
have
called
today
is
more
of
one
of
those
coding
episodes.
B
Yes,
so,
first
up,
let's
talk
a
little
bit
about
service
when
many
people
say
the
term
serverless.
What
they
mean
is
this
idea
of
just
basically
a
blob
of
code
that
gets
run
on
demand
in
the
openshift
serverless
product?
I
don't
know
what
do
we
call
like
or
like
operator
like,
whatever
we
call
operators
when
we
add
them.
B
We
have
three
what
you
know.
I
kind
of
think
of
this
like
pillars
to
the
serverless
product
and
so-
and
I
kind
of
mentioned
two
of
them
in
that
sentence
about
how
I
run
the
function.
So
the
first
thing
is,
and
in
arguably
the
most
important
part
is
this
guy
this
idea
of
what
they
call
serving,
which
is
the
idea
of
only
consuming
resources
on
demand
right.
So
in
other
words,
when
you're
not
using
the
function
or
using
the
feature
set
of
the
code,
it
goes
to
zero
right.
B
There's
nothing
running
there
aside
from
kind
of
the
infrastructure
stuff
and
it
just
idles
in
the
background.
Or
does
it
doesn't
sorry
it
doesn't
even
idle.
It
just
turns
off
this,
as
I
discovered
going
through
this
process
can
make
debugging
difficult,
but
you
know
the
this
is
something
one
of
the
things
I
learned
going
through
this
process.
I
have
a
number
of
bugs.
I
would
like
to
file
both
from
a
user
experience
perspective,
but
also
just
actual
bugs
well.
A
If
I
remember
correctly
when,
when
serverless
really
kind
of
like
took
off,
there
was
always
this
problem
of
like
cold,
starts
and
like
having
a
warm
start
and
like
if,
like
for
a
little
bit,
corey
quinn
was
like.
Why
have
a
warm
start
if
you're
like
just
just,
do
it
on
a
server
right
like
if
you
have
a
constantly
running
function
like
there's
no
point.
B
Right,
like
that
kind
of
defeats
a
purpose,
so
I
can't
speak
to
the
actual
kind
of
advent
of
marcus,
but
this
is
one
of
the
major
differences
between
using
corcus
and
a
typical
jvm
engine.
Is
that
and
like
I've
seen
the
numbers
too,
I'm
trying
to
remember
but,
like
you
know,
starting
basically,
the
same
functionality
on
a
typical
jvm
is
like
10
seconds
or
something,
and
starting
in
quartus
is
like
milliseconds
right.
So
the
trade-off
right
is
that
corkus
is
much
more
like
a
compiled
language
of
certain
things.
B
You
can't
do
you
know
which
you
might
expect
etc,
but
so,
despite
the
hype
with
serverless,
it
is
not
for
everything.
It
is
not.
The
is
not
the
right
answer
to
all
questions.
Not
all
you
know
not
all
things
look
like
a
nail
when
you
have
a
hair,
so
keep
that
in
mind.
So
that's
kind
of
the
serving
component,
the
other
component,
that
I'll
talk
about
next,
which
we're
not
really
going
to
talk
about
very
much
today.
Is
this
idea
of
venting?
B
And
so,
if
you
you
know
in
the
eventing
part
of
it
is,
I
need
a
way
of
kind
of
doing
the
triggering
right.
So
it's
kind
of
exactly
what
it
sounds
like,
and
so
one
of
the
really
excuse
me
one
of
the
really
popular
you
know,
kind
of
eventing
things
actually
there's
two
at
least
for
us
right.
So
there's
kafka,
which
is
I
what's
funny,
is
that
the
way
it
was
originally
presented
to
me
a
long
time
ago?
Is
it's?
B
Oh
actually,
I'm
sorry,
so
that's,
basically
the
twitter
firehose
right
so
and
one
of
the
things
that,
unless
you've
ever
done
anything
with
the
twitter
api,
you
wouldn't
necessarily
know
it's
actually
very
very
hard
to
get
everything
that
happens
on
twitter.
They
they
limit
it
to
a
very
small
group
of
customers.
You
know
who
I
assume,
pay
twitter,
vast
amounts
of
money
or.
A
A
A
B
Yeah,
maybe
it's
mq
streams,
but
so
basically
you
know
it's
a
piece
of
software.
It's
been
around
for
a
long
time,
it's
kind
of
more
well
known
in
the
java
world,
but
basically
it's
this
idea
of.
Well,
I'm
getting
events
from
all
these
different
kinds
of
things.
I
want
them
all
to
go
to
the
same
place.
I
want
to
do
transforms
across
them,
etcetera,
right,
so
et
cetera.
So,
hey
look
at
that
is
openshift
going
to
do
a
faz.
We
even
got
a
question.
B
Jp
literally,
are
talking
about
it
today.
So
yes,
sort
of
so
the
next
thing.
So
like
many
buzzwords,
faz
means
a
lot
of
different
things
to
a
lot
of
different
people.
Function
as
a
service
is
typically
what
it
stands
for,
but
again
it's
a
little
bit
questionable
like
what
does
function
mean
in
that
sentence.
Right
so
and
that's
the
last
part
of
the
sh
thing
we
ship
as
openshift
serverless,
currently
tech
preview,
so
just
kind
of
keep
that.
B
But
the
other
two
pillars
I
just
mentioned
rga
and
the
expectation
is
that
the
functions
part
is
going
to
go
ga
soon
and
yeah,
so
k
native
is
actually
the
upstream
for
the
serving
and
eventing
part
of
openshift
serverless
and
another
project
called
boson.
B-O-S-O-N
is
the
upstream
for
the
functions
part
they
do
other.
They
have
other
upstream
stuff
too,
but
you
know
think
about
it
as
the
functions
part
right.
So.
A
B
So
so
jp
dave
mentions
functions
of
service
as
being
roughly
equivalent
to
aws
lambda,
and
so
that
is
one
approach.
Azure
functions.
I
think
it
kind
of
does
the
same
kind
of
model
and
I
want
to
say:
google
has
one
too,
but
I've
never
been
able
to
figure
out
which
one
of
their
many
products
it
is,
but
that's
the
idea
of
I
have
some
code
and
I'm
going
to
plop
it
somewhere
and
it
will
get
run
on
demand.
B
That's
one
approach
to
functions.
The
other
approach
to
functions
is,
I
have
a
container
and
I'm
going
to
plop
it
somewhere
and
it's
going
to
get
run
on
demand.
Aws
offers
that
in
fargate
I
think
it's
the
name
of
the
product
azure
has
one
too,
which
is
called
something
something
you
know
run
container
on
demand.
I
don't
know,
and
so
so
basically,
you
kind
of
have
these
two
halves
of
that.
B
You
know
basically
same
goal.
Same
outcome
same
you
know
whatever,
but
there's
at
least
the
debate.
I've
seen
right
is
around
which
one
is
more
functional,
so
you
get
a
bunch
of
feature
set
from
the
fact
that
you
can
just
drop
code
somewhere,
which
is
super
handy,
very
hard
to
modify
extract.
You
know,
keep
track
of
reuse
things
like
that,
because
it's
literally
just
you
know
a
piece
of
code
on
on
the
internet.
B
B
So
you
know,
oh
sorry,
but
the
barrier
to
entry
in
some
ways
is
high
right,
because
now
you
have
to
like
develop
a
container
and
you
have
to
make
sure
it
lands
in
the
right
place.
So
you
know
I
I'm
I'm
kind
of
neutral
on
the
debate,
but
the
serverless
functions
thing
that
we're
offering
in
openshift
serverless
is
the
container
style,
which
is
what
the
boson
funk
project
does
upstream.
B
A
B
A
B
B
B
So
of
course,
there
is
no
way
to
tell
the
difference
between
these
two
windows,
this
one
all
right.
So
there
you
go
when
you're
doing
something
with
openshift
and
it's
like
demo
right.
My
I
usually
have
to
have
two
logins
or
often
have
to
have
two
logins,
because
I
want
to
have
like
the
cluster
admin
login
right,
but
then
I
want
a
real
user
login.
B
So
the
way
I
do,
the
real
user
login
is
as
a
private
window,
so
like
the
only
difference,
is
the
private
icon
in
firefox
all
right,
so
I'm
not
actually
going
to
show
you
kind
of
the
full-on
install
I'm
just
going
to
kind
of
show
you
the
steps,
but
basically
you
go
here
and
you
know,
and
if
you
don't
have
cluster
admin
access,
you
go
and
petition
your
your
sysadmin
gods
for
to
install
red
hat
openshift
serverless.
I've
obviously
already
installed
it.
B
B
Right,
so
exactly
so,
so
I
filed
a
bug
about
this,
but
basically
that
the
answer
I
got
back
was
because
it's
the
4
6
version,
because
that's
the
oldest
version,
it's
supported
on
you,
know
kind
of
throw
that
out
there
and
and
take
that
you
know
the
way
you
like,
but
keep
it
in
mind.
You
know
if
you're
actually
running
four
seven
you
might
want
to
switch
and
it's
pretty
easy
to
switch
right.
You
theoretically
can
just
switch
to
four
seven.
A
B
B
The
big
old
install
button
wait
for
a
bit
then,
like
some
operators,
you
kind
of
have
to
go
to
one
more
step,
and
actually
I
will
point
out
here,
potentially
two
steps
so
because
we're
only
talking
about
k
native
serving
I'm
only
installing
k
native
server.
So
one
thing
and
not
if
you
notice,
k
native
event.
So,
even
though
I
have
like
the
capability
of
it,
I
don't
actually
have
k
native
eventing
just
like
when
I
install
the
operator,
I
have
the
up.
B
You
know
the
ability
for
k-native
serving,
but
not
the
actual
implementation,
so
the
k-native
serving
when
we're
talking
about
serving
and
venting
and
functions
this
this
candidate
of
serving
is
the
serving
pillar.
But
it's
what
kind
of
holds
up
the
function
right.
So
you
you
need
both,
but
you
don't
need
a
cadence
of
eventing
unless
you
want
to
use
that
too.
B
One
thing
to
keep
in
mind
is
that
when
you
go
click
this
create
candidate
of
serving
you
need
to
make
sure
you're
in
the
key
native
serving
project
which
you
won't
be
by
default.
When
you
do
the
install.
So
just
keep
that
in
mind,
this
is
all
covered
in
the
installation
guide.
So
follow
the
steps
in
the
install
guide.
You
know,
I
definitely
recommend
it.
I
didn't
run
into
any
problems
there.
B
B
B
So
now
I
will
go
back
to
my
terminal
to
kind
of
show
you
okay,
so
now
what
I
do
with
it.
Some
of
this
you
can
do
in
the
ui.
Some
of
it.
You
can't
you
know
the
the
tool
chain
for
doing
this
on
the
command
line
is
so
like
simple
that
I
didn't
find
it
terribly
usable
like.
I
didn't
even
really
try
to
pull
this
off
in
the
ui,
because
it's
literally
like
two
commands,
so
I
just
kind
of
you
know
I
just
kind
of
did
it
this
way.
B
A
B
God
so
this
is,
where
kind
of
I
want
to
file
some
bugs
again,
one
of
those
things
where
it's.
It
looks
to
me,
like
the
engineers
expected
users
to
use
a
thing
one
way
and
the
docs
authors
expect
to
think
of
people
to
use
things
a
different
way
and
like
a
lot
of
software,
you
can
use
them
two
different
ways.
B
B
However,
I
can
also
go
get
funk,
which
is
the
boson
funk
command
and
yeah,
so
I
can
have
them
separate,
so
I
can
go
get
kn,
which
I
can
get
from
here.
Funk
is
not
here.
I
have
to
go
download
that
separately,
but
you
can
use
them
independently,
but
funk
can
also
be
a
plug-in
to
kn,
so
open
shift
somewhere.
B
A
That
are
tech
preview.
If
you
do
this
in
your
own
cluster
and
you
call
support,
you
might
find
that
they
will
push
back.
If
you
do
this
in
your
cluster
and
if
you
do
anything
that
you
see
on
this
channel
in
your
cluster
and
you
call
support
tell
them-
you
saw
it
on
openshift
tv,
so
they
can
come
yell
at
me
for
not
doing.
A
B
B
So
I
am
putting
it
in
the
chat
as
well,
but
basically,
if
you
go
download
the
serverless
clients
from
the
basically
the
build
infrastructure
right
rather
than
getting
it
from
kind
of
the
direct
link
out
of
openshift,
this
k-native
install
includes
the
plug-in
okay
cool.
So
to
be
honest,
this
may
not
have
been
necessary
right.
B
I
might
have
been
totally
fine
with
funk,
but
I
wanted
to
kind
of
follow
the
directions
coming
from
the
docks
you
know
kind
of
as
closely
as
possible,
basically
so
that
I
can
go
and
file
bugs
if
there's
any
bugs
and-
and
so
you
know
so,
I
didn't
actually
experiment
all
that
much
with
just
using
straight
funk.
Your
mileage
may
vary
feel
free
to
do
whatever
you
write,
fun
yeah.
B
B
A
B
B
That
said,
you
know
so
so
as
a
result,
many
of
the
go
application
world
you
know
of
which
most
you
know
now
we
have,
I
don't
know
what
are
there
like
50
different,
like
kubernetes,
go
tools
right,
there's,
k
native
there's
like
I
have
them
in
my
whole,
like
os
clients
thing
that
I
was
working
on
on
an
episode
a
while
ago.
You
know-
and
I
I
only
have
like
10
in
there.
You
know,
there's
argos,
you
know
all
of
these
are
statically
linked.
B
All
of
them
are
unrelated
to
each
other.
None
of
them
are
packaged,
and
so,
as
a
result,
things
like
this.
In
my
mind,
things
like
this
happen
right,
so
you
have
now
multiple
kns
in
the
wild
that
you
can
have
issues
with
or
whatever
you
can
say.
Oh
you
know,
you've
got
the
wrong
one.
You
know
you
can't
just
say
get
up
to
date,
because
you
you're
still
wrong.
B
So
there
it
is,
you
know,
take
it
or
leave
it,
but
this
is.
This
is
one
of
the
arguments
for
why
packaging
is
important,
and
this
is
one
of
the
examples
of
where
it
drops
the
ball
when
you
don't
use
kind
of
proper
packaging.
One
problem,
so,
okay,
rants
aside,
so
now
what
we
do
is
we
take
a
look.
What
funk
can
do
for
us
and
the
first
thing
we
can
do
is
it
has
a
nice
scaffolding
ability,
especially
when
you're
getting
started?
B
This
is
super
handy,
so
you
just
give
it
create
you
give
it
a
language.
You
want
you
give
it.
Of
course
you
can't
remember
the
syntax,
it's
language
and
then
we
need
to
give
it
a
trigger.
That's
a
t,
so
you
say
python
python
and
then,
as
it's
been
a
while,
since
I
did
a
lot
of
typing,
so
this
should
go.
This
should
be
fun.
B
So
if
we,
I
also
like
to
try
to
push
push
it
up
towards
the
top
of
the
screen.
Oh
and
now
we
have
to
give
it
a
name.
So
what
we're
gonna
do?
Okay,
so
to
back
up
a
little
bit.
I
actually
mentioned
it
on
an
episode,
a
couple
episodes
back
one
of
the
things
that
so
I
use
telegram
to
communicate
with
my
family,
along
with
a
lot
of
other
people,
and
when
I
want
to
share
a
link
with
my
family.
A
B
Usually,
you
know
you
can
go
and
like
edit
it,
but
doing
that
on
my
phone
or
whatever
is
a
lot
harder
than
you
know
like
doing
it
on
a
laptop
or
whatever
is
fine
but
doing
it
on
my
phone's
pane.
So
what
I
wanted
to
do
was
create
a
little
telegram
bot
that
I
could
send
the
url
to,
and
it
would
send
me
back
the
you
know
the
cleaned
up
version
so
so
we'll
just
call
it
clean,
url
and.
A
B
A
B
B
Kind
of
the
idea
behind,
or
the
argument
for
and
against,
ad
blockers
right,
it's
like
right,
a
lot
of
the
internet
is
free
because
they
surface
advertising.
If
you
use
ad
blockers
you're,
you
know
you're
you're,
not
paying
them,
or
at
least
you're,
not
paying
them.
That
particular
thing,
so
you
know
I
use
ad
blockers,
sometimes
basically,
where
you
know
where
I
want
to
kind
of
you
know.
B
If
I
want
to
give
it
back
right
or
I
you
know,
I
will
continue
to
use
like
fully
loaded
marketing
links
and
everything
else
when
I'm
kind
of
browsing
around
the
internet
myself.
But
when
I
pass
it
to
my
kids,
you
know
this
is
a
different
person
now
and
their
kids.
So
it
is
definitely
a
trade-off
and
something
to
keep
in
mind
right
that
you
don't.
You
know
every
time
you
kind
of
clean,
a
url.
Every
time
you
use
an
ad
blocker,
your
changing
your
relationship
with,
whatever
you
know,
source
you're
using.
A
And
I
would
you
know,
I
would
even
say
they're,
not
ad
blockers
anymore
they're,
actually
like
content
blockers
right,
like
they
will
block
whole
js
libraries
from
loading.
You
know
the
whole
nine
yards
and-
and
that's
great,
if
you
like
you,
know
a
lightweight
web
right
like
right.
I
think
we
all
would
appreciate
a
lighter
weight
web.
You
know,
given
the
something.
A
A
Yeah
I
mean-
and
these
are
things
that,
like
I
use
judiciously
right
like
because
like
for
example,
I
use
the
browser
brave,
which
has
a
lot
of
ad
in
content
blocking
bits
in
it.
But
it
breaks
websites
sometimes,
and
they
have
a
report
broken
website
button
and
just
about
every
tool
that
I
use
nowadays.
A
B
A
A
B
B
B
B
A
B
In
three
three,
so
source
to
image
is
an
approach
to
this.
Basically,
this
is
the
kind
of
thing
that
turns
an
open
shift
into
a
pass
like
so
a
platform
as
a
service
rather
than
a
you
know
something
that's
closer
to
an
eye
as
or
an
infrastructure
as
a
service,
because
what
it
does
is
it
says:
okay,
I
know
how
to
build
python
projects
for
this
kind
of
scenario,
for
all
that
generic
stuff,
that
you
as
the
end
developer,
don't
care
about.
B
Okay,
you
care
about
it
sometimes
or
a
little
bit,
but
mostly
about
how
it
works,
not
about
replacing
it
right.
So
what
this
is
doing
is
essentially
providing
a
path
for
sorry,
I'm
very
coffee.
Today
I
just
had
enough
coffee
by
the
way.
Oh,
I
am
also,
I
think,
out
of
coffee,
so
we
should
be
wrapping
this
up
soon.
I
think,
but.
B
All
right,
so
this
is
kind
of
the
pause
part
of
it.
The
nice
thing
about
that
little
crate
statement
is,
it
knows
what
to
go.
Look
for.
One
thing
I
noticed
in
here
is
that
while
I
was
working
on
this
project,
they
actually
released,
I
believe,
released
a
new
version
of
some
of
the
paths
components
so
like.
B
I
think
it
was
this
one
here
so
and-
and
I
don't
know
how,
but
it
decided
to
drop
in
a
tag
here,
so
I
wasn't
getting
the
latest
version
anymore,
and
so
then
it
started
to
break
in
mysterious
ways.
So
keep
that
in
mind,
if
you
know,
keep
an
eye
on
this
yaml
file,
even
though
you
don't
really
need
it
very
much,
it
does
get
changed
by
the
various
operations
we're
going
to
do,
and
sometimes
not
for
the
better.
B
B
But
I
have
an
even
more
fun
example
of
that
in
a
second,
so
that's
one
of
those,
so
the
one
that
I
think
may
also
be
a
thing
is
that
it
also
tags
the
images
digest
when
you
update
it.
So
if,
like
on
the
list
of
when
things
are
behaving
oddly,
I
go
clear
that
I
make
sure
this
doesn't
have
a
tag
on
it
and
that
seemed
to
make
things
better.
So
just
kind
of
some
tips
of
you
know
how
to
avoid.
B
B
So
then
it
generates
you,
this
template,
func
file,
and
so
the
keyword
is
funk
dot
pi.
So
the
the
file
has
to
be
called
that,
or
maybe
that's
settable
somewhere,
but
I
haven't
found
it
yet,
but
that's
the
default.
Basically,
you
end
up
with
this
context.
Object.
One
thing
to
keep
in
mind
is
that
there
is
another
python
project
called
parliament.
That
is
not
the
one
we're
talking
about
here.
This
is
actually
parliament
dash
functions.
B
And
so
yeah
keep
that
in
mind.
It
basically
just
shoves
a
flask
request
object
and
this
cloud
event
object.
If
you
have
one.
This
is
where
this
is
part
of
where
the
eventing
stuff
comes
in.
So
if
it
comes
in
as
a
an
event,
it'll
actually
come
in
as
a
cloud
event.
But
imagine
it's
basically
just
you
know
a
two
item
array.
One
is
a
flash
request
object
and
one
is
a
cloud
event
except
everything
that
I've
done
so
far
has
not
had
a
cloud
event.
B
B
B
Does
that
work?
That's
good!
Oh
yeah,
that
one
does
now.
I
can
read
it
yeah,
so
so
what
I'm
trying
to
do
is
remember
a
few
episodes
ago
I
found
like
a
profile
switcher
for
vs
code,
so
I
finally
got
it
working
and
all
set
and
can
switch
now
between
my,
like
normal
view
and
my
video
view,
so
I
just
gotta
remember
to
go
tweak
it
a
little
bit
to
maybe
go
one
more
notch
up.
B
That's
all
if
you
notice,
I
have
already
deployed
this
thing
or
whatever
I'm
trying
to
think
how
to
explain
this.
Okay.
So
this
ad
added
by
my
build
infrastructure
right.
This
was
also
added
by
the
build
infrastructure
and
then,
as
you
notice,
I
don't
have
those
tags
that
I
was
worried
about
before
right,
but
we
have
the
actual
code
here.
So
what
lands
in
here
right
is
so
I
get
this
context
object
just
like
before,
and
I
probably
should
have
updated
the
the
comment
I
didn't
get
around
to
it.
B
One
thing
to
keep
in
mind
is:
I
was
having
a
lot
of
trouble
getting
logging
info
out
of
the
python
library
and
into
some
place.
I
could
read
it
so
if
I
tried
a
print
statement
that
did
not
work,
even
though
that's
typically
the
container
debugging
model.
The
only
thing
that
I
found
that
I
was
getting
having
good
luck
with
getting
data
out
of
it
was
in
my
actual
box.
So
the
way
this
is
designed
is,
I
have
a
function
as
a
service
that
is
a
url
cleaner.
B
A
B
Right,
okay,
so,
but
I
think
this
is
it
so
check
it
out
clear
urls,
that's
the
browser
plugin,
but
so
it
has
a
community
around
it
and
stuff
like
that,
so
the
the
url's
cleaning
gets
better
over
time.
Basically,
so
what
do
I
do
here?
Oh
so,
what
I
wanted
to
show
you
was
with
the
logging
oops
this
bot
that
I
was
trying
to
get
the
actual
bot
side
of
it,
which
is
the
part
where
you
know
telegram.
B
So
you
massage
the
inbound
request
into
the
telegram
object.
Then
you
can
kind
of
interact
with
it
and
say:
okay,
let
me
look
down
here.
I'll
show
you
where
it
does
it,
but
basically
it
says
up
there
here:
go
construct
the
url
then
do
the
url
request
right
then,
once
I
get
a
response
back,
then
you
basically
send
the
bot
a
message
back
with
the
clean
url
right,
so
pretty
straightforward
right,
like
basically,
you
have
two
serverless
functions,
one
that
does
url
cleaning,
one
that
responds
to
bot
commands.
B
B
So
I
wanted
to
separate
those
two
functions,
but
in
the
headache
of
trying
to
figure
out
because
the
problem
with
doing
a
lot
of
this
stuff-
and
this
goes
to
where
it
might
have
been
actually
useful-
to
go
and
install
service
mesh,
because
this
is
where
the
problem
of
distributed
tracing
really
shows
its
head.
You
know
shows
its
face
right,
which
is
that
I
have
all
these
different
services
right.
B
Yes,
and
so
you
can
see,
it
has
a
bunch
of
different
commands
here,
but
then
I
have
this
bot
called
url,
cleaner
bot
and,
as
you
can
see,
I'm
sending
in
a
number
of
messages
trying
to
figure
out
what
exactly
is
getting
thrown
over
the
wall
from
telegram
into
my
serverless
function
and
then
what
you
know
damage
is
being
done
to
the
object
that
it
received,
and
how
can
I
then
massage
that
into
something
else,
and
so
what
do
you
need
for
that?
You
need
logging
right.
B
B
The
only
messages
I
could
get
it
to
print
were
warning
and
error,
and
so
I
stuck
with
warning
you
know
just
because
sure-
and
I
also
was
trying
to
get
it
to
whatever
like,
like
pretty
print-
essentially
objects
rather
than
getting
them
directly.
You
know,
because
I
don't
I
don't
care
about
the
object,
handle
what
I
care
about.
It's
like
what's
in
the
object,
so
that's
kind
of
the
the
bot
itself,
but
going
back
to
my
little
clean
url
thing.
B
So
this
thing
works
really
quite
well,
because
I'm
getting
a
very
simple
get
request
which
I'm
basically
getting
in
the
url.
Then
I
can
pull
the
flask
object
out
of
the
context.
Get
this
rec
object,
then
I
can
dig
into
that
and
there
I
have
one
get
parameter
called
url
and
so
basically
make
sure
that
the
url
actually
has
some
content
to
it.
You
know
et
cetera.
I
do
a
little
bit
of
error
handling,
it
probably
can
use
more,
but
then
go
and
do
the
actual
cleaning
and
then
give
it
back.
B
B
B
Level
power,
okay,
so
now
I
have
a
project
there
and
that's
the
bulk
of
what
I
kind
of
need
to
do
regarding
openshift,
because
now
what
I
can
actually
do,
which
is
super
handy,
is
just
want
to
make
sure
this.
You
know
work
yeah,
it
should
so
I
do
the
you
know.
I
use
the
special
k
native
that
has
funk
as
a
plug-in
and
then
I
do
what's
called
a
build
and
I
can
do
a
minus
v,
so
you
can
kind
of
see
all
of
it.
B
There,
it
is,
of
course,
control
shift
the
bar
there.
We
go
okay,
so
we
actually
covered
this
on
the
show.
Some
couple
of
episodes
ago,
podman
recently
released
the
ability
to
act
as
a
rest.
Endpoint
that
looks
like
the
docker
rest
endpoint.
Yes,
however,
you
have
to
turn
it
on
yeah,
and
I
think
I
actually
can
do
this
at
the
like
system
level,
but
I
just
didn't
get
to
it.
So
the
simple
version
is
basically
just
run
this
as
a
system
service
listing
on
port
one.
Two,
three
four.
B
Yeah,
so
to
lost
algorithms
point
yeah,
I'm
not
sure
like.
I
don't
know
where
that
object
is
landing
or
what
exactly
is
going
on
there.
It's
kind
of
weird
okay,
so
I
clearly
already
have
a
system
service
running
it.
It's
just
that
I
lost
my
environment.
Variable
right,
so
docker
host
just
needs
to
be
told
that
this
is
where
it
is
right.
So
that
should
work
now,
theoretically,
okay,
cool,
so
then
it
goes
and
pulls
you
know
basically
the
builder
component,
so
the
the
platform.
B
B
B
Humans
that
happen
to
be
on
both
projects
but
they're,
they're,
independent
things.
Buildpack
has
its
own
mission,
so
build
packs.
There's
lots
of
different,
build
packs,
different
use
cases,
build
packs,
maybe
we'll
cover
build
packs
of
different
show,
but
that's
different
okay.
So
now
what
I
did
was
I
built
my
container
so
now.
If
I
look
in
podman
images,
I
have
a
new
container,
except
this
is
something
I
haven't
figured
out,
yet
I'm
not
sure
where
we're
getting
41
years
ago,
because
I
know
I
just
built
this.
B
B
So
yeah,
I
don't,
I
don't
know
it's
weird
and
like
if
you
look
at
like
yeah
see
if
you
look
at
the
boson
container,
it's
fine
oops
back
to
the
t-mux
selection
problem,
but
yeah.
If
you
look
at
this
container,
it's
fine,
I
it's
something
about
how
the
can
funk
build
thing
is
building
that
it's
like
injecting
the
wrong
date.
Somehow
it's
very
weird
I
haven't
figured
that
one
out
yet
okay,
so
so
I
built
it.
B
Basically,
I
have
the
image
it's
tagged
in
here
right
with
the
remote
target.
So
one
of
the
things
that
I
kind
of
glossed
over
a
little
bit
is
that
I
would
have
been
prompted
for.
Where
do
I
want
this
container
to
land
right?
Ultimately,
because,
obviously,
if
I
build
a
container
just
on
my
laptop,
it's
not
going
to
be
useful
to
anyone,
because
you
you
can't
get
to
my
laptop
from
the
public
internet,
which
is
where
the
open
shift
is
from.
B
So
you
know
if
you
have
a
data
center
or
whatever,
with
a
shared
registry,
you
know
or
even
using
your
openshift
server
as
a
shared
registry.
You
can
use
that
as
your
target,
but
for
this
I
knew
it
was
all
public,
and
so
I
just
used
it
on
quay,
and
so
that
was
fine.
One
thing
I
wanted
to
kind
of
pause
and
mention
here,
though,
because
this
burned
me
a
little
bit
is
that
by
default.
B
The
problem
is,
it
will
create
it
private,
because
by
default
way,
created
is
private
by
default,
yeah,
which
then,
if
you
go
and
try
to
deploy
it
to
openshift,
it
says
I
can't
find
it
so.
You
just
gotta
have
to
keep
in
mind
when
you
do
that,
build
or
whatever
it
will
create
it,
and
then
you
need
to
go
log
into
clay,
make
it
public
then
come
back
and
do
your
deploy
and
then
things
will
go
less
path,
all
right.
B
B
B
Oh
right
yeah,
so
one
thing
I
want
to
point
out
quickly:
if
I
can
is,
we
have
actually
created
a
whole
project,
not
that
one
this
one
and
if
you
notice
it
created
a
whole
project
or
sorry,
I
did
a
new
project
to
get
to
the
project
itself.
Sorry,
it
created
a
whole
new.
You
know
like
service
and
topology,
and
then
it
deployed
this
container
as
a
pod
there
and
gave
it
a
route
and
all
those
things.
So
all
that
stuff
happened
behind
the
scenes.
B
However,
if
you
run
out
of
space
in
temp,
which
you
know
I
got
four
gigs
there,
it
takes
a
little
bit,
but
not
forever.
If
you
try
to
build,
it
will
fail
and
fail,
like
it'll,
fail
with
a
message,
but
it
won't
error
out.
So,
for
example,
if
I
do
something
like
this,
which
is
what
I
usually
get
to
when
I'm
doing
a
lot
of
debugging
right
is
like
okay.
Let
me
try
this.
You
know
rebuild
redeploy
this
book.
B
So
it
creates
a
new
image,
util
thing
like
we're
talking
about,
but
and
it
will
redeploy
it
and
that's
all
good.
That's
exactly
what
I
want
right.
However,
eventually,
when
the
temp
runs
out
of
disk,
build
will
fail,
but
not
error,
so
it
won't
trigger
the
and,
and
it
will
proceed
to
deploy
the
deploy
doesn't
fail
because
the
container
is
there
and
it
is
just
expecting
it
just
thinks
you
want
to
do
a
redeployment
of
the
same
thing.
B
B
B
B
Because
we
auto
scale
the
zero
after
a
timeout
nice,
so
this
is
cool
because
it
will
also
auto
scale
the
other
way
and
we
should
see
it
yeah.
So
now
it's
blue
again
and
I
got
basically
my
error
message
and
if
we
do
url,
I
do
have
an
ipad
url
equals.
I
don't
know
google
and
I
can
remember
how
urls
work.
B
Right
question:
mark
yeah,
why?
Oh,
oh
so
this
is
the
other.
This
is
the
problem
that
is,
makes
it
hard
to
test
it's
just
you
have
to
you
encode
any
thing
you
put
into
it
right
because
otherwise
the
browser
tries
to
interpret
it
whatever.
So
I
just
whoa
wait,
I
don't
mean
you,
you
encode,
I
mean
url
in
code.
B
B
B
B
B
Man,
exactly
with
six
thousand
points,
nettle,
attack
them
with
5600
points.
No
friction,
I
think,
with
no
change
and
maybe
joe
fuss,
both
with
no
change
at
4,
000
and
2300
check,
conan
kudo
up
another
episode
or
so
as
well
as
bacon
fork.
You
know
thank
you
again
so
often
for
participating
in
the
sweet,
sweet
internet
points.
We're
always
excited
to
give
them
out
and
those
should
be
the
correct
codes.
B
Let
me
just
see
if
I
can
grab
a
url
for
you
so
that
I
can
drop
that
in
the
chat.
Thank
you
waiting
for
it
to
load.
Sorry,
usually,
I
have
a
cheat
sheet,
but
I
forgot
to
do
it
today,
all
right,
so
there
is
the
code
for
today's
episode.
You
know
please
go
and
you
know,
file
the
form
and
collect
your
internet
points
and
we're
always
very
excited
to
get
them
out
and
you
know
someday
we'll
help
swipe.
B
B
B
B
Of
weird
guys
on
it,
okay,
so
let's
see
oc
logs,
it's
probably
going
to
be
gone
by
the
time
I
actually
get
to
it.
The
other
thing
that
I
may
also
file
a
bug
about,
but
this
one's
tougher
is,
if
I'm
doing
something
like
this.
Oh
actually,
I
don't
know
if
this
thing's
gonna
have
any
logs,
because
I
don't
think
I
actually
write
anything
to
the
log,
but
the
other
one.
The
other
one
sends
out
a
lot
of
messages
because
it
still
doesn't
work
right.
B
So
I
haven't
cleaned
out
the
debug
messages.
One
of
the
things
I
was
just
going
to
say
something
else
about
this,
so
one
of
the
other
things
that
I
may
kind
of
go
follow
about
is
this
construct
here
is
also
kind
of
annoying
and
or
maybe
I'll
like.
I
need
to
find
a
way
to
write:
it's
not
an
alias
but
a
bash
function,
because
what
you
also
can't
do
is
I
can't
deconstruct
this
and
get
this
because
user
container
has
to
come
after
this,
because
otherwise,
I
could
say
something:
fancy
like.
B
One
and
then
I
could
do
like
xrs
oc
logs
and
I
could
even
put
a
user
container
here,
but
I
can't
put
it
after
particularly
easier
easily,
so
yeah,
so
that's
kind
of
annoying.
This
is
this
is
one
of
those
huge
complaints
about
systemd
when
it
came
out
and
and
even
today
right.
So
if
you
do
system
control,
you
know
you
know,
restart
I'm
going
to
intentionally
make
a
mistake.
There,
network
manager.
B
It's
hard
to
do
that
in
a
bash
script,
because
you
can't
because,
because
the
keyword,
the
the
changing
thing
is
in
the
center
of
the
command,
which
is
much
harder
to
script
around
than
when
it's
on
the
end.
So
if
you
remember
how
service
worked,
it
would
be:
oh,
my
goodness,
network
manager
restart.
B
A
B
A
B
B
That
time
it
was
blank
because
I
bet
it's
sleepy
now,
yeah
see
it's
gone
again,
so
so
yeah.
So
this
is
a
little
bit
of
the
annoying
headache.
You
know
I
could
obviously
wake
it
back
up
and
you
know
kind
of
all
that
stuff,
but
long
story
short.
You
know
it's
a
little
bit
of
pain.
You
start
to
get
the
hang
of
it
and
I
think
you
know
a
nice.
B
You
know
I
have
a
lot
of
whatever
they
call
bash
functions
right,
so
you
know
writing
a
new
one
with
you
know.
Basically
you
can
you
know
you
can
have
full
commands
in
there.
So
doing
that
to
get
to
the
logs
is
not
too
big
a
deal.
It
was
kind
of
annoying.
So
you
know
your
mileage
may
vary,
but
so
hopefully
this
was
a
good
little
intro
to
how
to
do
some
serverless
stuff
completely
like
from
zero
right,
and
you
know
you
should
totally
go
check
it
out.
B
B
The
all
these
bots
aren't
out
there
running
that
anybody
can
use
necessarily,
but
you
can
go
run
it
yourself
on
your
favorite,
open
shift
instance,
or
you
know
one
thing
that
I
haven't
done
a
lot
of
experimenting
with,
but
I
think
you
might
be
able
to
do
is
you
could
actually
run
it
in
pod
man?
So
I
don't
know
it
would
be
interesting,
but.
B
Doesn't
give
you
any
of
the
capabilities
that
openshift
would,
in
this
case,
like
the
auto
scaling
right
right
right
off
the
bat?
So
it's
not
quite
the
same,
but
at
least
the
same,
it's
one
of
the
other
things
so
kind
of
going
back
to
the
original
arguments
about
the
serverless.
You
know
being
just
a
blob
of
code
versus
serverless
being
a
container.
B
B
B
A
Well,
you
know
coming
up
here
at
11
we
have
asked
and
openshift
admin
we're
going
to
be
talking
about
the
new
bits
and
openshift
4.8
for
administrators
and
then
that's
it
basically
45
minutes
from
now
and
then
at
p.m.
Eastern
1800,
utc.
We
are
going
to
talk
about
troubleshooting
rel
and
I
know
that
is
a
wide
and
open
topic.
But
we're
like
we're
going
to
break
some
stuff
and
we're
going
to
show
you
how
to
fix
it.