►
From YouTube: Grafana Community Call 2021-01-21
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).
A
A
I
saw
that
you've,
you've
added
a
couple
of
questions
already.
A
Definitely
not
so
let
me
see
if
I
can
read
it
just
real
quick,
hello.
B
A
A
You
can
come
back
if
you
want
to
try
something
out,
and
you
know,
come
back
and
see,
ask
follow-up
questions
if
you'd
like
so
we
do
have
a
few
graffinistas
in
this
call
with
us.
A
We
have
myself,
we
have
brian
gann,
I
see
we
have
daniel
lee
and
diana
payton
as
well.
So.
C
D
C
C
There
is
a
believe,
it's
called
query,
support,
help
in
the
developer,
plugin
documents
now
and
so
yeah,
you
can
add,
you
know
all
sorts
of
notes
and
you
know
pre-built
queries
and
any
other
gotchas
and
stuff
for
your
personal
plugin
into
into
your
data
source.
Plugin
awesome:
do
you
do
you
have
a
link
you
can
share?
A
So
if
you
do
have
a
question,
a
plug-in,
related
question
or
if
we
run
out
of
them
other
grafana
related
questions
as
well.
Please
add
it
to
the
agenda.
I
will
you
know,
go
through
the
the
questions
in
the
list
here
and
you
know
if
we
run
out
of
questions
there,
just
you
know
throw
one
out
there
and
we'll
get
to
it.
So
the
first
question.
A
The
first
question
is:
is
it
possible
to
somehow
access
secure,
json
data
to
be
able
to
send,
for
example,
api
keys
in
the
body
of
the
request
using
backend
srv.fetch,
and
I
don't
believe
you
send
brian
correct
me
if
I'm
wrong,
but
I
don't
believe
you're
sending
the
secure
adjacent
data
in
the
fetch
request,
you're,
muted,.
E
B
Or
is
it
possible
to
do
it
like
this?
Typescript
are.
F
E
I've
only
done
it
with
the
back
end,
not
not
with
the
front
end.
F
B
Yeah,
so
I
I
I
thought,
if
it
maybe
is
there
some
construct
that
you
can,
you
can
use
keys
to
to
so
to
give
us
better.
Back-End
knowledge
know
that
that
I
I
I
want
that
key-
that
I've
saved
in
my
configurator
config
config
editor
that
I
want
this
key
to
be
sent
like
a
like.
B
B
B
B
So
I
I
I
can
send,
I
can
send
it
like
hardcoded,
so
it
works,
but
but
I
I
can
see
also
password
and
and
user
information,
and
also
I
can.
I
cannot
deny
dynamically
change
it
in
configure
config,
configure
data.
F
Yes,
it's
the
point.
It
is
like
a
security
flaw
if
you
were
able
to
access
it
from
accessing
from
your
your
javascript
code.
So
if
you
like
decrypted
the
password
and
add
it
into
your
body,
then
basically
anybody
using
your
plugin
can
can
see
see
that
the
decrypted
password,
but
I
am
thinking
brian,
you
know
this
with
the
routes
and
and
the
the
proxy
should
be.
F
E
E
F
Did
you
give
me
one
minute,
I'm
just
going
to
look
at
the
dig
into
the
plugins
I've
written
and
see
if
I
did
have,
I
ever
done
anything
similar,
but
why
are
you
sending
the
password
in
the
body
that
seems
strange?
What's
what's
the
api
that
takes
a
password
in
the
body.
B
So
I
have
like
apis
that
use
basic
authorization
and
also
user
cs
clients,
a
client
id
and
client
secret.
So
I
done
it
on
my
front
end
and
then
save
it
as
a
encrypted
basic
header,
and
then
I
get
it
in
rows.
B
This
is
the
graph
roles,
so
like
a
header
name
yeah,
and
the
problem
is
that
I
need
to
get
to
get
my
ap
to
get
my
data
from
my
p,
I
mean
also
to
give
my
user
and
password
and
it
is
not
possible
to
do
it
with
rows,
because
we
there
is
no
field
like
body
to
to
to
write
like
point
secure,
json
data
point
password,
for
example,.
D
F
B
Just
like
we
have
authorization
zone
like
out
to
point
mode
and
then
we
need
to.
We
have
some
pet
proxy.
That's
why
we
need
one
more
access
key
and
I
cannot
use
this
authorization.
The
out.
B
Because
I
need
one
more
header
and
so
it's
like
two
requests,
so
I
I
need
two
requests
to
do
it.
First,
one
is
client
secret,
client
id
and
grant
so
I
should
add
and
then-
and
then
I
I
get
this
access
token
and
with
access
token,
I
can
get
my
data
from
api,
something
like
that.
F
Okay,
there
is
my
access
token
stuff
built
into
the
routes,
but
it's
all
done
with
headers.
B
Yeah
because
we
have
like,
like
it's,
it's
minus
out
minus
token
header
and
I
need
to
add
this
header
to
like.
B
F
An
oauth
endpoint
yeah,
it
was,
it
is
okay,
but
then
there
is
support
for
that.
Then
I
will
send
you
an
example,
so
that
is
supported.
There
is
a
a
a
a
fields
called
auth
on
the
routes
and
we
have
a
documentation
for
this.
I
think
yep.
B
B
That's
there,
the
manual
manually
added
in
our
proxy.
B
F
End
this
is
like
a
one
of
the
reasons
to
to
write
like
a
back-end
like
a
go.
Back-End
for
your
plug-in
is
just
with
custom
authentication,
there's
like
two
two
main
reasons,
and
one
is
custom,
authentication
and
the
other
one
is
alerting.
If
you
want
a
learning,
support
and
grifana
then
asks
on
the
backend.
F
I
don't
know.
This
is
some
way
to
wrangle
our
route
system.
Don't
know
marcus,
o'brien
any
thoughts,
so
the
the
the
token
auth
one
that
that
marcus
link
there
in
the
chat,
like
everything,
happens
from
the
back
end.
So
you
don't
you
can't
get
access
to
the
token
from
the
front
end
afterwards,
it's
all
done
sort
of
invisibly
on
the
back
end.
B
But
then
I
show,
but
then
I
should
write
all
my
code
and
go
so.
I
cannot
use
typescript
as
a
logic
for
queries.
A
No,
so
the
the
front-end
code
for
when
you're
using
a
back-end
plug-in
is
very
minimal.
So
most
of
the
querying
logic
happens
in
the
back
end
and
then
you
forward
that
to
to
the
the
typescript
code.
So
the
typescript
component
is
very
minimal.
B
Yeah,
that
was
my
point,
because
I
not
really
the
guy
that
write
code
and
go
so.
I
need
to
understand
how
goal
is
working
then.
Maybe
that's
why
I
asked
is,
if
it
possible
or
is
it,
is
it
possible
to
make
like
a
body
in
a
epi
proxy?
That's
a
grafana
implemented,
and
then
you
can
something
like
this
as
I
can.
I
I
send
it
in
chat,
so
we
have
pass.
B
E
E
So
instead
of
calling
query
to
the
back
end,
you're
calling
get
resource,
and
that's
where
you
do
your
query
to
your
custom
api
with
all
the
headers
and
everything
encrypted,
and
you
just
send
that
raw
payload
back
as
whatever
is
json
or
whatever
content
it
could
just
be
csvs
or
any
raw
data
format
back
to
your
typescript
and
then
process
it
in
the
front
end.
A
Yeah,
we
don't
have
documentation
for
the
resources
yet,
but
if
brian
doesn't
have
it,
I
I
can
give
you
an
example
of
how
you
could
add
a
resource.
A
A
So
you
understand
the
rest
of
the
code,
but
the
difference
here,
as
brian
pointed
out,
is
that
instead
of
of
registering
a
a
handler
for
your
queries,
you
only
register
a
resource
handler
and
that
becomes
sort
of
a
as
a
web
server
on
the
back
inside
that
you
can
query
for
any
endpoint.
A
B
A
B
A
So
my
my
first
instinct
is
a
streaming
background
or
a
streaming
data
source
where
you
can
set
up
depending
on
where
you
wanna
subscribe
to
the
data
from.
If
it's
a
a
another
api
that
has
support
for
websockets,
you
can
hook
up
a
websocket
and
then
just
stream
the
data
from
there.
And
then
the
panel
will
update
as
the
new
data
comes
in.
A
But
it
depends
on
where
you
get
the
data
from.
Alternatively,
you
could
write
a
back
end
again.
If
you
have
a
you
know,
let's
say
that
you
have
something
like
a
you
know,
a
message,
bus
or
something
like
that
that
doesn't
support
an
http
api
for
example.
Then
you
would
want
something
like
a
back-end
service
or
a
back-end
plug-in.
A
So
I
guess
the
question
to
you,
then,
is
what
what
type
of
api
does
the
service
have.
B
It's
like
rest,
so
I
I
do
some
subscription
on
my
web
server
and
I
give
url
from
grafana
and
then,
if
something
happens,
then
this
web
server
pushed
the
notification
so.
B
A
It's
medical,
I
think,
yeah,
you!
I
think
you
could
use
the
the
resources
again
for
this,
because
I
think,
assuming
that
you
get
you
have
the
permission
for
for
accessing
the
the
the
resource
endpoint
for
that
plugin,
you
could
implement
the
post
endpoint
for
it.
Does
that
sound,
reasonable,
daniel
and
brian?
A
But
but
the
second
part
of
it
is
to
get
grafana
to
you,
know:
asynchronously
update,
yeah,
and
that
would
be
you
know,
I
think,
do
we
have
you
know
daniel
or
brian.
Do
you
know
if
we've
implemented
circular
data
frames
in
the
back
end?
A
I
haven't
heard
about
it,
yet
I
haven't
had
it
okay,
so
in
the
front
end
it's
it's.
There
is
a
specific
data
frame
that
is
like
you
know
circular.
So
if
you
add
something
to
the
the
the
end
it
kind
of
pushes,
so
it
keeps
a
buffer
and
it's
very
useful
if
you
have
something
like
you
know,
streaming
data
source,
but
I'm
not
sure
that
we
have
that
in
the
back
end.
A
So
I'm
not
sure
if
you
can
get
the
you
know
the
as
async
refresh
from
a
back-end
plug-in
at
this
point.
It's
definitely
something
that
we
want
to
support
at
some
point,
but
I'm
not
sure
if
we
got
it
right
now,
but.
E
You
know,
I
know
I
know
you
could
do
you
can
do
a
streaming
response
over
a
resource.
Also,
oh,
you
can
yeah
yeah
you
can.
That
was
that
came
out
in
seven
one.
I
think.
E
C
Okay,
because
there's
an
example
of
that
in
the
7.4
play
that
I
dashboard
that
I
linked
in
the
agenda.
A
As
far
as
I
know,
that's
that's
that's
like
alpha,
though
I'm
not
sure
that
is
even
exported
yet
is
it.
C
Oh,
it's!
It's
live
yeah,
it's
just
kind
of
buried
in
the
test
data
db
thing.
Let
me
get
that
too
come.
A
But
the
question
is
whether
that
api
has
been
exposed
in
graph
in
the
grafana
plug-in
apis.
I'm
not
sure
it
has.
Yet
it
might
be
an
internal.
It's
something
that.
D
F
Going
on
so
yeah,
I
think
before
all
the
bits
come
together,
it'll
probably
be
8.0
in
in
may.
D
Actually,
with
streaming
the
way
how
it's
done
right
now
in
grafana,
it's
you,
I
mean
you,
you
can
request
the
data
from
a
backend
data
source
and
just
show
it
on
the
panel,
but
for
some
of
my
panels,
which
we
are
doing
for
redis,
it's
actually
because
the
backend
is
a
stateless
breaking
data
source.
D
So
when
you
want
to
preserve
some
data
from
a
first
query
and
then
show
second
quest
additional
data
on
the
second
query
right
using
this
previous
data,
you
have
to
do
it
on
the
front
end,
which
is
why
we
just
did
a
set
timeout
on
the
front
end
and
do
our
own
kind
of
streaming
station
and
show
the
data
with
the
custom
panels
on
the
graph
or
on
the
in
the.
D
F
Okay-
I
I
don't
actually
don't
know
enough
about
this,
so
I
suppose
ryan
who's,
the
the
guy
who's
working
on
on
on
the
streaming.
He
has
a
lot
planned
for
eight
zero,
I'm
not
sure
exactly
what's
going
to
make
him
date
or
zero,
so
that
could
be
something
we
could
get
him
to
talk
about.
One
of
these
meetings
about
the
the
plans
going
forward
there.
D
A
So
sorry,
what
kind
of
documentation
would
you
like
to
see
like
how
to
work
with
panels
and.
D
Yeah
with
the
panels
graphs,
I
mean
when
you
create
a
panel
right,
you
want
to
put
your
components
on
it
right,
the
graph,
for
example,
or
the
new
time
series
you
have
a
table.
So
what
we?
What
we
did
something
I
was
talking
before
we
had.
We
use
a
back-end
data
source.
We
get
one
data
once
then,
after
period
of
time
we
get
the
second
data
for
data,
then
we
aggregate
it
and
I
wanted
to
show
it
on
the
graph.
So
we
had
a
really
hard
time
to
understand.
D
Oh,
we
need
to
put
it
y
access
to
display
this
data,
and
then
we
put
how
the
way
how
we
do
the
legend.
So
we
added
the
legend.
Okay,
that's
great
in
the
list.
After
that,
we
see
on
our
axis
in
y-axis,
we
see
the
numbers,
but
I
don't
see
milliseconds
there.
B
D
D
A
Yeah
right
and
so
from
what
I
know
the
you
know
that
that
graph
component
that
you're
using
it's
it's
kind
of
static
and
limited-
and
I
mean
that's-
the
big
part
of
the
new
graph
panel-
is
to
make
it
into
a
reusable
component.
So
we
are.
We
are
very
much
aware
of
that
limitation.
A
Well,
what
I've
heard
is
8.0,
but
that
that
might
be
subject
to
change
depending
on
their
development.
Is
it?
Is
it
more
of
like,
what's
available
to
the
metrics.
E
Is
that
that
documentation
is
sparse
or
because
that's
kind
of
like
what
are
the
things
that
are
that
come
with
come
with
the
class
that
you're,
extending
or
in
react
something
that
you're
inheriting.
D
D
So
this
is
the
way
how
we
do
that
we
create
with
components
with
access
labels,
visible
time
frames,
values
and
use
the
graphics
legend.
We
specified
units
in
a
couple
of
places,
but
still
it's
not
displaying
that
properly
or
it's
really
hard
to
navigate
with
code
to
display
milliseconds
on
the
graph.
A
I
think
that's
yeah,
that's
the
best.
I
haven't
personally
used
the
graph
graph
component
at
all,
but
we
can
definitely
forward
it
to
the
the
the
team
but
even
better
open
eye
issue,
and
I
can
I
can
send
it
to
the
the
people
interested.
A
A
A
C
C
C
A
I'm
seeing
a
jennifer
in
here
did
you
have
a
question
or
just
hear
out
of
curiosity.
G
Hey
yeah
nice
to
meet
everyone
yeah,
I'm
new
to
grafana
product
manager,
we'll
be
working
with
the
back
end
enterprise
team.
So
I
just
wanted
to
drop
in
a
little
bit
and
kind
of
see
how
things
things
worked.
A
Cool
well
happy
you're,
here
excited
to
be
here
too
thanks,
say:
hi
yeah
these
these
calls
they
they,
you
know
every
each
one
of
them
are
a
bit
different
depends
on
what
we
have
to
share
and
what
the
pla
the
community
has
to
share.
We've
done
some
show
and
tells
we've.
A
B
A
So
one
interesting
thing
that
I
found
out
about
just
a
couple
of
hours
ago
is
the
the
frame
meta
in
the
plug-in
apis.
Not
sure
mikhail.
Did
you
know
about
it
or
maybe
I'm
just
slow
to
the?
What
is
it
it's
it's
see
if
I
can
get
it,
I
was
hoping
to
to
look
into
it
a
little
today.
It's
it's
a
meta
property
on
on
data
frames
that
can
provide
a
little
extra
debugging
information
for
users,
statistics
and
you
know,
output
yeah.
I
just
learned
about
it,
so
I'm
kind
of
curious
about
it.
A
Send
you
a
link
here
so
this
was
that
was
just
highlighted
or
sent
to
me.
So
there's
a
on
the
on
the
frame.
There
is
a
field
called
meta,
and
I
I've
been
told
that
if
you,
if
you
supply
the
the
frame
meta,
then
you
get
some
additional
nice
features.
You
know
for
just
seeing
things
like
performance
statistics
on
the
data
returned
and,
for
example,
what's
the
preferred
visualization.
A
So
if
you
have
a
let's
say
that
you
have
a
data
source
that
returns
a
data
frame,
you
could
pass
a
hint
to
the
explorer
mode
on
how
you
want
it
to
be
presented.
A
So
you
could
pass
a
visualization
type,
for
example,
so
you
can
hint
ex
give
a
hint
to
explore
on
whether
you
want
to
visualize
this
as
a
graph
table
or
logs,
for
example,
and
apparently
you
can
set
some
kind
of
hierarchy
here
as
well.
That,
hopefully,
is
used
for
for
some
kind
of
debugging
purposes.
A
E
It's
exactly
like
you
said
you
can
spend
being
able
to
specify
a
preferred.
You.
D
D
E
Yeah,
so
it's
well,
let's
see
with
the
field
no
for
the
field.
Configuration
would
could
be
part
of
the.
A
Questions
might
as
well
learn
a
little
bit
together,
so
it
does
have
a
a
custom
property
here
that
I'm
I'm
I'm
very
curious
about
data
source,
specific
values
and
it's
just
a
it's
just
an
interface.
D
Right,
yes,
I
see
there
is
a
preferred,
visualization
type
and
executed
question
by
the
way.
If
you
have,
if
you
have
a
custom
panel,
can
you
update
property
of
the
data
source
from
the
panel?
A
D
D
I
have
some
specific
panels
which
want
to
do
a
specific
task
from
my
data
source.
So
the
way
how
I
do
that,
normally,
when
you
have
your
data
source
right,
it's
a
return
table
or
graph,
and
you
display
it.
With
your
panel,
I
have
a
special
panel
which
have
to
execute
specific
the
queries
from
the
data
source.
So
this
is
why,
instead
of
a
user
who
have
to
put
the
panel,
I
have
to
configure
the
data
sources
the
way
I
need
the
put
the
right
command
there.
D
A
I
I
no
I'm,
I
admit,
I'm
struggling
a
little
bit.
Do
you
have
an
example
of.
D
D
D
You
see
my
grafana
right
yep,
so
what
what
we
do?
Let
me
start
from
the
beginning,
so
I
have
multiple
panels
biggest
keys
panels,
cli
and
latency.
So
for
the
latency
panel.
I
I
have
a
specific
query
in
my
on
my
mind,
which
I'm
using
for
backend
data
source.
So
I
don't
actually
feel
that,
and
data
and
panel
itself
knows
what
kind
of
data
it
should
get
and
execute
from
query.
D
Think
this
would
have
said
that
I
can
do
graph
and
show
me
the
axis
the
numbers,
but
I
don't
know
that's
milliseconds.
Actually
this
is
something
we
have
to
fix.
Still
same
with
another
panel,
I
have
my
panel
and
it's
kind
of
do
the
scanning
process
of
the
keys
in
the
database
so
to
make
it
work.
Actually
I
have
to
choose
radius,
then
I
have
to
use
the
scan
command
and
there
are
multiple
parameters
I
have
to
specify
and
then,
when
I
do
that,
I
can
do
like
a
start
scanning.
A
So
I
see
okay,
so
you
want
to
be
able
to
like
select
from
presets,
possibly
in
the
panel,
and
that
will
configure
the
the
query
editor
for
you.
D
D
And
we
are
working
actually
on
a
very
cool
pr
for
sdk
go.
Maybe
you
saw
that
not
that
when
your
backend
data
sources
they
are
blocking
sometimes
so,
when
you
have
multiple
data
sources
of
the
same
kind
and
one
of
them
time
out
and
for
some
reason
it's
blocking
other
data
sources
of
this
type.
This
is
the
way
how
it
works.
So
now
one
of
my
developers
here
submitted
pr
request
when
it
uses
special
caching
technology
in
mutex,
so
one
data
source
of
the
same
type
one
will
be
independent
of
another
one.
D
Because
it
was
a
big
problem
that
you
have
from
in
my
case
array
is
data
source
and
well
one
of
the
databases
it
just
time
out
it.
It's
not
available
right
and
all
data
sources
came
out,
so
it's
applicable
for
data
source
as
well.
D
Right
how
it
changed,
how
instance
manager
managed
data
source
installation
they
get
back
and
forth,
adding
some
table
some
changes
and
tests
the
way
how
we
should
probably
test
it,
because
it's
changed
behavior
of
the
data
sources,
but
I'm
excited
if
it's
going
to
be
done.
A
Well
I'll
leave
that
to
marcus.
It
seems
like
he's
he's
on
that.
A
D
Do
you
have
any
adaptation
metrics
of
grafana,
how
many
users
still
use
version
6,
who
is
migrated
to
version
7
7.3,
because
this
is
always
right
now
in
in
my
case,
in
radius
data
source?
I
have
a
limit
of
version
7.2,
because
this
is
where
I
have
variables,
which
is
available
right
in
data
source.
But
if
I
want
to
use
live
streaming,
which
is
available
in
version
7.4,
I
have
to
say
this
is
the
version
I
need.
A
So
you
would
like
to
be
able
to
see
how
much
how
many
users
will
be
impacted
by
a
change.
Basically
right,
yeah.
D
A
Yeah,
I
know
that
we
have
some
of
this
information,
so
I
wonder
how
we
can
best.
You
know
make
that
make
that
available,
because
I
am
I
hear
you
because
I
even
in
my
own
plugins
I've
been
thinking
you
know
how
should
I
move
to
this
new
api
because
it's
only
available
in
the
absolute
latest
minor
version-
and
you
know-
that's
probably
gonna
break
it
for
a
lot
of
users.
So
I
agree
that
it
would
be
helpful
to
just
get
us
some
indication
of
the
users
I'm
targeting
with
this
release.
A
A
So
I
know
that
in
7.4
I
believe
there
is
a
new,
a
new
api
for
calculating
min
and
max
for
for
fields
and
it's
a
very
helpful
function,
but
it
will
only
be
available
to
users
7.4
and
on,
and
that
means
that
you
can't
just
update
your
your
your
plugin
overnight.
To
make
use
of
that.
You
need
to
have
some
kind
of
fallback
for
users
that
don't
have
that
and
then
you
know
maybe
have
some
indication
of
when
you
can
drop
that
fallback
so
yeah.
I
definitely
see
the
use
case.
A
A
E
E
D
D
D
The
way
how
it
works
remember,
we
talked
when
you
run
with
workflow
you,
you
create
a
new
version,
yeah
yeah
yeah.
This
doesn't
get
published
right.
I.
F
F
A
A
Yeah,
exactly
my
my
thought
was
to
you
know,
construct
the
the
the
future
url.
It's
not
gonna,
be
you
know
it's
not
gonna
be
valid
yet,
but
once
you
make
that
release
it's
going
to
be
so
so
yeah
I've.
A
It's
thank
you
for
reminding
me
I've
I've
gotten
a
couple
of
pull
requests
with
or
urls
like
this
to
the
plug-in
repository.
So
it's
definitely
something
I
need
to
look
into,
but
yeah.
So
I'm
thinking
just
instead
of
because
what
the
workflow
does
right
now
is
that
it
pulls
it
from
one
of
the
github
actions.
A
It's
a
it's
an
output
for
from
the
the
creep
release
action
I
believe,
and
instead
I
just
you,
know,
construct
it
from
the
values
I
have
so
that
that
could
be
it's
good.
It
could
be
an
alternative.
D
Okay
also,
I
found
I
I
changed
the
change
because
the
change
lock
was
fixed
in
there
with
one
of
the
recent
versions.
So
I
I
updated
my
change
log
and
updated
changelog
for
the
plugin
same
way
the
change
logged
on
for
grafana
it,
which
has
one
hash
for
the
version.
D
A
D
Okay,
because
grafana
project
has
one
hash
for
the
version
and
three
hashes
for
features.
Yes,.
D
I
was
surprised
I
mean
I
kind
of
updated
my
changelog
and
then
I
run
with
so
release,
or
at
least
it's
not
something
you
can
test
and
run
all
the
time
right.
So
I
run
it
and
found
okay.
This
is
a
version
which
is
untucked
which
something
has
to
be
changed,
updated
and
the
change
lock
it
didn't
pick
up.
So
I
tried
to
follow
because
I
can
find
the
plugin.
I
try
to
follow
the
changelog
format
of
the
grafana
itself
and
didn't
pick
up,
so
you
have
to.
A
I
think
the
reason
for
this
is
on
grafana.com
plugins.
If
you
look
at
under
the
tab
change
log
yeah,
I
think
it's
just
you
know
prints
the
the
contents
of
the
changelog.md,
so
you
you
probably
want
you
know
a
a
header
and
I
think
it's
just
to
make
the
the
output
on
grafana
the
calm
a
little
nicer.
D
A
I
I
get
to
me,
the
the
the
documentarian
in
me
is
kind
of
hurting
when,
when
I
see
multiple
single
hashes
in
a
in
a
markdown
file,
so
so
I
would
rather
see
the
the
grafana
changelog
for
my
change.
A
A
Yes,
yes,
so
there,
ideally,
you
know
it.
Ideally,
you
should
only
have
one
one
top
level
header
in
a
markdown
file
and
that's
the
single
hash.
D
C
I'm
here
I'm
just
listening
and
being
a
fly
on
the
wall
when
it
comes
to
the
plug-in
documentation.
Marcus
is
the
is
the
king
there,
I'm
kind
of.
A
C
C
For
as
many
as
you
need
to,
you
should
start
with
an
h1
and
then
the
next
down
should
be
h2s
and
then
the
next
down
should
be
h3s.
You
should
use
headings
to
break
up
your
document
into
coherent
chunks.
There
isn't
really
a
rule
for
it,
but
I
guess,
if
you
had
to,
if
I
had
to
make
one
up
on
the
fly,
it
would
be
add
a
heading
where
every
someone
might
want
to
link
to,
but
really
headings
are
to
help.
C
The
google
tech
writing
for
engineers
course,
which
I'll
link
right
here,
has
some
really
good
guidance
on
how
to
use
headings
and,
if
you're
interested
in
improving
your
technical
writing
skills
that
I
highly
recommend
going
through
their
content.
It's
actually
written
for
engineers
and
it
it
covers
most
of
what
I
am
correcting
when
I
go
through
pr's.
So.
A
Yeah,
I
so
because
mikhail
he
noted
that
in
our
changelog
for
the
grafana
project,
there
is
multiple
h1s.
D
A
I
highly
recommend
it
it's
it's
specifically
made
for
engineers,
so
that
aren't
you
know,
tech
writers
by
profession,
but
it's
like
if
you're
a
developer
and
you
want
to
start
writing
better
documentation.
That's
the
the
target
audience
and
it's
a
it's
a
really
well
written
course
in
class.
A
So
we
have
one
minute
left.
Thank
you
all
for
joining
and
thank
you
for
all
the
great
questions.
I
have
a
ton
of
feedback
here
that
I
will
get
back
to
to
the
team
with
and
yeah
hope
you
got
some
answers.