►
Description
Bisq developer session 30.01.2019
https://github.com/bisq-network/growth/issues/90
A
Yeah,
hello,
everybody,
this
week's
developer
session.
We
want
to
keep
it
quite
short
today
and
I
think
mainly.
We
will
talk
about
the
monitoring
system.
What
Florin
is
working
on
and
we
have
a
discussion
about
testing
frameworks
will
be
put
on
for
maybe
next
week,
as
well
as
the
new
trade
protocol,
as
it
would
be
good.
First,
the
tech
spec
updated.
So
everybody
can
review
the
text
right
on
time.
Every
discussion
about
this.
Okay,
let's
get
started,
maybe
Florine
you
want
to
take
over
and
maybe
show
quickly
the
current
state
of
the
monitor.
Yes,.
A
B
Right,
yes,
okay,
so
I
wanted
to
start
with
a
little
overview
over
the
architecture
of
the
monitoring
system,
as
it
is
planned
as
of
now.
The
second
release
is
about
to
go
out
and
we
have
one
monitoring
node,
like
the
one
in
the
middle
and
one
monitoring
service,
I'm,
not
sure
how
we
can
achieve
multiple
monitoring
nodes
with
one
service,
but
well
we'll
figure
that
out
when
it
is
time.
B
So
we
have
this
one
monitoring
service.
The
monitoring
service
is
the
graph
on
our
open
source
project,
which
is
very,
very
nice.
It's
a
lot.
It
has
been
a
lot
of
work
that
went
into
the
CRO
final
project
and
it's
it's
really
nice,
the
backend
for
the
so
that
occur
final
project
is
the
the
GUI
part,
the
user
interface
part.
B
Is
it's
basically
just
a
database
where
you
can
send
occurrences
of
certain
of
stuff
yeah?
So
you
have
you
have
a
value
and
you
have
a
timestamp
and
this
the
the
the
graphite
instance
does
memorize
the
time
stamp
and
the
value
you
can
add
the
text
to
the
values
so
that
you
know
what
the
value
means.
But
basically
you
have
a
value
at
a
time
step
if.
B
C
B
Believe
by
default
it
is
implemented
system,
whispered,
database.
That's
a
patent
thing,
it's
a
fixed
size
database
where
you
you
have
you
have
you
can
configure,
for
which
time
spans
you
want
the
data
and
and
how
fine-grained
the
data
is,
for
example,
you
can
say
well,
I
want
every
minute
of
data
in
the
last
day,
and
only
data
accumulated
to
half
an
hour
42
years,
two
years,
for
example.
So
this
is
this
is
really
special
to
the
graphite.
B
B
Okay,
so
that
is
the
architecture.
Are
there
any
questions
regarding
the
architecture?
No
okay,
so
the
monitor
note
looks
like
this.
I
I
tried
to
keep
the
monitor
note
as
simple
as
it
possibly
can
be,
meaning
that
the
the
seed
monitor
did
this
this
little
thing
here,
oh
nice,
there
is
something
quite
wrong.
Okay,
this
this
very
monitor
is
starting
up
a
complete
bisque
application
that
the
whole
the
whole
peer-to-peer
network
thing
the
whole
node
it
gets
connected,
it
stays
connected.
It
uses
up
resources
of
the
seed
nodes
and
so
on.
B
What
I
try
to
do
is
limit
the
dependencies
on
the
rest
of
the
big
project.
I
do
use
the
the
peer-to-peer
stuff
I
do
use
the
tour
staff,
of
course,
but
I
do
not
use
anything
like
persistent
data
store
or
something
like
this
or
how
is
it
called
peer-to-peer
manager?
Stuff
I
do
not
use
this
until
now.
Maybe
the
need
arises
someday,
but
I
managed
to
do
to
keep
it
rather
simple
reason
is
I,
have
more
control
over
what
data
I
want,
what
did
I
get?
B
How
often
I
get
the
data
and
how
big
my
footprint
in
the
internet,
which
can
be
so,
for
example,
I
just
for
example,
I
go
over
to
to
our
tour
matrix
page,
which
is
already
up
and
running
for
a
month
or
so,
for
example,
if
we
take
the
the
diagram
at
the
top
right
here,
it's
the
tour
round
trip
time
and
for
this
graph
to
to
create
this
graph.
I
only
use
one,
for
instance,
no
risk
at
all.
B
So
one
metric
is
the
tour
round-trip
time.
For
example,
another
metric
is
the
what
is
it
the
OpenOffice,
for
example,
there
is
a
common
configuration
system
behind
this.
There
is
a
scheduler
integrated.
So
all,
if
you
have,
if
you
want
a
new
metric,
you
just
have
to
implement
the
metric
interface
and
do
whatever
you
want.
That
means
we
can
also
do.
For
example,
the
exchange
rates,
monitor
exchange
rates,
fee
estimation
or
dual
psql
price
monitoring,
because
we
do
not
need
the
whole
disk
for
a
matrix
adjust
the
data
we
want.
B
Are
there
any
questions?
No
okay.
So
what
do
we
have
already?
If
you
go
to
a
monitor
piece
network,
you
get
to
this
very
page
and
I
just
played
around
a
bit
with
mr.
stuff
and
just
hit
it
five
minutes
ago.
The
system
has
helped
on
the
red,
because
the
Piscopo
need
to
a
starred.
Dashboard
is
kind
of
boring
by
including
only
these
three
buttons,
but
let's
get
let's
get
to
to
each
of
these
areas.
This
area,
for
starters,
we
have
the
biscuit
or
matrix.
B
Maybe
I
can
get
rid
of
the
bisque,
because
it's
just
a
tour
matrix
if
I
click
the
tour
matrix.
We
have
a
bunch
of
diagrams.
Of
course
nothing
of
this
is
is
is
fixed.
We
can
change
it
anytime
if
seven
something
new
arises
or
if
there
is
not
data
that
error
state
that
we
want
to
see
but
do
not
see
at
the
current
state.
B
But
basically
you
can
do
these
fields
of
interests.
One
field
is
like
like
this.
You
can
drag
it
around,
so
this
is
the
the
most
nice
thing
about
crow
fauna.
Each
of
these
fields
can
hold
information.
You
can
configure,
as
you
like
in
this
very
page.
I
have
summed
up
anything
we
collect
regarding
the
toward
in
at
the
underlying
tour
networks.
There
is
nothing,
there's
no
bisque
involved,
it's
just
tour.
B
The
reason
why
we
do
it
is
is
we
there
has
been
a
time
very
where
that
the
network's
was
really
crowded
and
that
the
biskits
network
was
really
crowded
and
tor
had
some
issues
with
denial
of
service
attacks
and
so
on.
So
the
the
reason
for
this
dashboard
is
that
we
can
see
how
well
tor
is
performing
and
tour
only
nothing
to
do
with
bisque.
For
example,
we
see
that
the
average
startup
time
of
tour
is
about
six
seconds.
B
So
if
you,
if
you
download
the
app
the
disk
app,
get
it
running
the
first
time,
the
tour
startup
time
is
six
seconds,
so
that
means
you
run
the
app
the
app
unpacked
the
binaries.
As
far
as
the
binary
starts,
the
binaries
until
the
Tor
binary
is
ready
to
do
request,
so
it
takes
6
seconds
until
you
can
do
a
request
and
ask,
for
example,
to
seed
nodes
for
data.
B
The
second
important
thing
is
the
the
hidden
service
startup
time
the
hidden
service.
Total
time
is
the
time
following
after
the
6
seconds.
So
if
the,
if
the
tour
binaries
are
in
place
and
we
can
use
tor
to
browse
the
net
disk
starts
to
to
publish
its
hidden
service
and
it
is
the
average
for
over
three
hours-
and
we
see
we
see
a
timely
the
data,
how
long
it
takes
to
get
hidden
service
up
and
running,
and
we
see
the
tourists
out
of
time
does
flick
straight
a
little
bit.
B
The
hidden
service
start
of
time
on
the
bottom
right
is
rather
stable.
It
takes
32
seconds
and
bear
in
mind.
This
whole
system
runs
on
a
ditch,
loschen
server,
so
there
is
no
internet
service
provider
in
between
which
can
maybe
slow
things
down.
This
is
a
public
IP
and
the
host
does
nothing
else,
but
this
matrix.
C
C
B
C
B
B
B
There
should
not
be
any
denial
of
service
or
something
like
this
that
will
slow
the
the
rent
return
down.
Only
reason
I
do
this
is
because
you
can
see
tor
is
working
at
or
is
in
a
scientifical
has
been
some
difficulty.
We
see
there
is
a
half
a
second,
but
there
has
been
a
time
where
one
little
request
took
43
seconds.
B
So
we
get
an
idea
of
how
to
open
forms
and
and
pea
based
on
from
this
information
we
can,
we
can
say
well
if
there
is
some
network
issue
you're
in
trouble,
we
can
say
well.
Tor
is
one
of
the
major
reasons
that
is,
that
basically
is
in
trouble
or
bisque
is
in
trouble
because
of
Biscay
and
not
because
of
tor.
B
There
are
always
some
some
there's
some
data
missing,
because
I
restarted
percent
and
just
to
get
to
get
rid
of
all
the
runtime
issues
and
so
on,
but
okay,
the
first
first
thing
I
want
to
talk
about
is
the
top
right.
The
peer-to-peer
networks
load,
the
beautopia
network
loads.
If
you
know
I
mean
there,
is
there
some?
There
is
data
that
is
not
displayed
there,
but
I
can
I
come
to
this
later.
B
Measurement
so
that
stuff's
got
added
same
here,
one
tray
statistic
and
one
of
a
payload
is
new
in
the
network
if
I
have,
if
I
have,
if
I
take
them
right
before
north.
Before
known,
there
is
no
data
at
all
new
in
the
network,
so,
for
example,
if
if
there
is
some
denial
of
service
attacks
ongoing
or
something
like
this
and
or
someone
does
at
100
offers
something
like
this,
we
would
see
in
this
time
slot
there
has
been
there
are,
there
have
been
100
offers
edit.
A
A
Secret,
wouldn't
be
probably
not
possibly
most
days
from
the
broadcast
messages,
and
when
you
receive
a
message
from
the
network,
you
don't
know,
we
store
originator
in
the
offering
some
type
of
messages
there
is
the
onion
address
inside
like
an
offer,
but
another
message
is
not
like
a
witness,
and
this
data
you
received
from
one
see
note
but
proceed
not
received
it
from
somewhere
from
all
his
peers
were
what
is
headed.
You.
B
B
A
A
B
B
B
B
B
B
B
We
had
400
milliseconds
22
seconds
and
one
second,
so
we
get
an
idea
how
how
well
the
signals
are
connected
and
if
they
are
there,
okay
I
said
this
is
the
the
maximal
value
so
that
the
law,
each
each
measurement
I,
think
I,
take
five
measurements
for
each
value
and
then
I
calculate,
and
that
we
can
see
here.
I
do
have
a
couple
of
different
values
available.
There
is
a
maximum
value,
a
minimum
value.
B
There
is
the
p50,
so
the
median
value
or
I
do
have
a
25
if
necessary,
and
the
P
75
and
average
is
already
is
also
there.
If
we
need
it.
However,
I
believe
that
using
the
the
maximum
round-trip
time
is
the
best
instrument
to
judge,
because
maybe
subsequent
requests
to
the
same
host
get
there
faster,
because
the
network
is
already
primed
for
this
memory
request,
so
it
would
not
reflect
a
user
that
only
started
it.
It's
basically
application
because
there
is,
there
is
no,
no
no
history
to
his
basic
application.
B
A
B
Yes,
although
I
observed
it,
it
is
not
always
the
first
thing
that
takes
the
longest
time,
so
I'm
not
quite
sure
of
what
what
the
mechanisms
are
behind
this
data,
but
I
believe
that
the
maximum
time
is
you
put
a
starting
point.
We
can,
of
course,
use
the
the
median
or
the
average
or
something
like
this
yeah
open
for
discussion
has
said
before
nothing
of
this
is
set
in
so.
B
So
I
I
take
over
a
reference
source.
So
this
it's
a
particularly
I,
think
it's
the
seven-to-three
seat,
note
and
say:
well,
this
seat
note
is
my
reference,
so
this
signal
has
always
feel
different.
However,
if
another
seat
note,
for
example,
here
if
it
yes,
then
I
see
well
the
seventh
or
two
three
years.
B
Each
node
is
is,
of
course,
the
reference
peanuts
or
there
is
a
different
to
different
to
itself,
but
the
other
two
seat
notes
in
this
diagram,
one
is
to
message
short
and
one
is
three
messages
short
and
we
have
the
same
thing
on
the
other
plus
side.
Of
course
there
another
seat,
though
this
has
two
messages
more
than
than
the
reference
signals
and.
A
B
The
main
idea
behind
this
very
diagram
is
that
if
you
see
that
there
is
a
difference,
so
this
is
this
is
the
number
of
messages
there
is
no
percentage
in
kivett.
Is
it
just
the
number
of
managers
we
can
do
and-
and
that
is
the
the
next
thing
I
want
to
talk
about-
we
can
set
alerts
to
the
whole
thing.
A
B
B
Munford,
you
said
that
arbitrator
messages
may
be
more
important
than
trade
statistic
messages
you
see.
I
have
added
an
alert
to
this
diagram
as
well.
So
if
you,
if
you
look
there,
we
have
this
in
the
very
first
line
is
the
alert
the
number
of
alert
messages,
the
number
of
arbitration
arbitrator,
messages,
filter
messages
and
so
on,
and
this
very
alert
is
configured
to
trigger.
B
Of
course,
this
data
is
just
a
single
seed
note
for
now:
I'm,
not
quite
sure
how
we
can
do
a
diagram
for
each
and
every
seat
note,
but
we
will
conquered
it
as
well,
but
we
have
the
possibility
to
say
well,
if
you
see
now,
it
has
less
than
aesthetic
number
of
certain
kind
of
message
we
want
to
have
an
alert.
This
is
also
possible.
B
At
last
but
not
least,
there
is
this:
this
diagram
here
on
the
right
in
the
middle
that
does
this
nice
slow
flash
and
has
a
broken
heart
up
here,
I
figured.
We
can
use
the
round-trip
time
to
figure
out
if
a
signal
is
alive
or
not,
and
you
see,
I
did
have
this
free
feed
nodes
life
at
11:10
today
and
then
I
did
some
reconfiguring,
also
visible
data.
The
data
here
below
is
not
existing.
I
switched
another
seed,
node
and
then
the
alert
said
well.
B
I
do
have
five
seed
nodes,
but
I
only
receive
data
from
two
and
that
for
over
two
hours,
says
here
that
the
red
red
line
here
is
the
time
when
the
element
is
triggered.
So
at
this
time
the
alert
triggered
because
one
or
more
of
the
signal
values
that
have
been
not
there
if
I
do
add
the
seed.
Now
it's
I
configured
before
the
alert
should
finish
so.
Basically
we
have.
We
have
the
means
to
see
if
the
seed
note
is.
A
B
It
is
now
we
see
if
the
seed
nodes,
if
one
of
the
seed
nodes
is,
is
come
or
we
can
obtain.
If
there
is
one
that
is
in
fourth
orders
multipling,
what
we
can
just
say,
one
of
the
cenotes
or
more
is:
is
it
fort?
Maybe
I
find
some
way
to
to
do
a
better
display
and
in
the
alert
list,
for
example,
but
that
is
the
status
I
have
planned
for
the
upcoming
release.
B
A
B
B
C
B
A
A
A
B
B
Take
this
from
every
average
yeah.
Well,
you
see
it
creates
pockets.
There
is
one
count,
for
example,
and
it's
one
count
in
the
time
range
of
one.
Second
half
a
second
to
one
second,
and
there
is
one
and
half
one
and
a
half
second
to
two
seconds
and
it
gets
lighter
I,
don't
know
what
what
what
what
behind
after
white,
but
the
white
thingy
says.
Well,
there
is
a
count
of
two
and
the
more
occurrences
in
this
pocket.
There
is.
B
A
Probably
when
this
message
maps
naturalist,
separated
in
two
diagrams,
where
one
is
the
priority
data
like
the
traitors
and
so
and
the
other
is
the
trace
36
and
they
counted.
We
must
take
terrorists
link
and
it's
from
a
good
enough,
because
then
you
see
that
basically
with
arbitrate
and
those
messages
should
be
always
the
same.
There
shouldn't
be
any
change
and
the
other
can
talk
to
it.
A
B
This
whole
diagram
thingy,
if
it's
mainly
that
we
can
have
a
quick
look,
how
how
the
network,
how
bisk
how
the
peer-to-peer
networks
is
performing.
If
there
is
some
issue
we
have
to,
we
configure
the
alerts
accordingly,
if
or
hand,
and
we
get
a
notification
that
there
is
something
wrong,
and
if
there
is
something
wrong,
we
can
take
a
look.
B
A
B
B
Yes,
I
believe
me:
we
should
discuss
the
backup
stuff
separately
because,
as
I
said
before,
the
database
is
a
fixed
database.
So
if
the
data
gets
older
measurement
points,
data
points
are
dropped,
so
we
only
have
every
half
an
hour
for
a
year
old
diagram,
for
example
there
we
now
have
every
five
minutes,
so
maybe
we
will
really
want
to
break
up
the
data.
There
is
only
the
data
so
that
we've
made,
maybe
where
exactly
we
have
to
discuss
this.
A
C
C
C
C
Something
like
that
yeah
and
just
so
so
if
they
see
everything
is
screen,
then
everything
is
fine,
but
if
they
have,
for
example,
problems
contacting
an
arbitrator
and
they
see
that
there
are
at
least
one
arbitrator
is
not
connected.
If
that's
that
might
be
an
issue,
they
are,
and
they
don't
spam
support
forums
or
so
and
just
wait
until
everything
is
green
and
then
try
again
and.
B
A
B
B
Okay,
so
the
last
thing
I
just
added
a
few
minutes
ago
well
an
hour
ago,
is
the
start
starting
point
for
trading
statistics.
For
example,
we
can
have
own
dashboard
for
this
and
I
just
pick
the
data
I
already
have
we
see
well,
we
actually
do
right
now
or
at
12:18
something
we
had
240
office
and
we
can.
We
can
add
other
trading
statistics
here
as
well.
C
One
more
question
of
luring
dimensions
in
the
beginning,
with
the
offer
payload
at
the
moment.
We
cannot
cannot
except,
for
example,
create
separate
charts
force,
for
example,
for
separate
markets
like
an
Arab
market
or
a
Pitkin
euro
market
to
have
his
Open
Office,
not
just
only
for
the
total
office
in
some
for
certain
markets,
a.
B
B
C
Yeah
I
think
from
a
kadhi
will
be
interesting
to
see
if
you
have
a
decrease
in
volume
on
certain
markets,
if
it's
directly
related
to
the
office
that
are
available
for.
If,
if
we
have,
for
example,
a
number
of
offers
triggers
and
organic
growth
also
on
the
trading
volume.
So
just
to
maybe
to
use
this
information
in
a
visual
way
and
it's
easier
to
find
correlation.
A
A
We
have
this
this
server,
which
was
earlier
pitch
itself,
and
now
it's
only
API,
which
is
used
by
the
markets
page
on
this
webpage,
and
it
it
just
right
here
that
run
say
you
know
the
headless
note
similar
like
this
he's
not
there
and
that
writes
the
data
is
chasing
funds
to
discount
the
PHP.
Application
is
using
this
data
so
to
end
and
terrorize
HTTP
API
for
requesting
some
data
and
basically
the
number
of
offers
for
currencies
and
so
on.
That
should
be
a
little
there
and
turn
of
trades
as
well.
A
I'm,
not
sure
how
much
we
should
overlap.
Those
two
areas,
I
think
at
the
moment.
Maybe
it's
better
to
keep
the
molitor,
more
low
level
and
technical
without
basically
deep
packet
inspection
that
they
look.
What
in
the
data
is
inside
but
of
course,
maybe
at
some
point
it
makes
sense
or
we
need
it
end
or
we
extend
it.
But
then
we
would
have
basically
a
second
parallel
statistics
about
this
date.
A
A
Used
by
certain
artists-
and
you
could
get
some
kind
of
like
Ito's
risk,
so
probably
it's
not
good
to
mix
this
together.
At
least
yeah
I
mean
a
monitor
server.
We
probably,
but
this
monitor
system
will
probably
not
have
an
open,
API
or
whatever,
because
it's
important
for
us
to
really
notice
that
it
doesn't
get
that
doesn't,
has
a
risk
to
get
details
to
whatever
I'm
with
the
marketer
or
you
know
happen
when
open
API,
but
it's
also
relevant
data.
A
B
We
can
have
different
monitoring
nodes
because
well
that
the
name
monitoring
node
is
maybe
not
very
correct,
because
this
monitoring
node
does
not
have,
must
not
have
Manor
or
may
actually
not
not
have
any
connection
to
the
peer-to-peer
network.
So
we
could
configure
a
second
node,
a
second
monitor
to
only
do
the
market
stuff
and
get
the
market
data
and
then,
but
then
send
it
to
the
to
the
monitoring
today
to
the
Crafar
now
thingy.
So
that
is
possible.
So
we
would
have
separation
and
and.
A
Somebody
it
was
when
the
second
node
would
have
an
API,
also
that
people
can
request
and
when
it
would
be
attacked
in
the
worst
case.
This
goes
down,
but
you
know
this
exactly
correct
as
well
to
see
because
the
other,
the
markets
really
it's
written
in
PHP
and
we
don't
have
an
active
developer,
who
is
maintaining
it
and
so
and
yeah.
It
would
be
probably
more
future-proof
solution
to
maybe
move
over
in
that
direction.
A
I
just
thought
about
to
get
lost
too
much
because
yeah
we
shouldn't
work
non-current
for
the
next
year,
we're
busy
it's
the
tool
to
improve
the
peer-to-peer
system
and
optimize
it
so,
but
midterm
are
us
with.
Maybe
somebody
else
can
can
work
on
distant
or
that's
not
so
related
at
the
peer-to-peer
system
anymore.
But
it's
interesting
concept
yet
who
have
talents?
Parallel
notes
was
different
for
different
tasks
and
different
risk,
exposures,
and
so.
B
A
B
B
Steps
or
tower
tasks
in
order
to
I
had
a
discussion
with
with
Manfred
today,
I
think
and
we
rearranged.
We
discussed
a
little
and
I
rearranged
the
tasks
according
to
this
discussion
it
and
now
there
is,
there-
is
I,
wanted
to
ask
you
guys
what
what
should
we
do
in
the
next
couple
of
months?
Let's
say:
what
should
we
do
for
the
settles
release
if
there
are
another
release
behind
and
what
should
be
the
contents
of
this
release?