How LLD macros can be expended in prototypes ?
Looks like support request, see http://www.zabbix.org/wiki/Docs/bug_reporting_guidelines
I'd close it.

Oleksiy, I don't understand you.
You don't believe me that it works?
It is really good feature and I want that it works in name of graph prototypes, in names of item protypes it work, do you need screenshots ?

The description is a bit misleading (or I'm bad ...)
Would be more clean if talk about lld created graphs but not their prototypes.
Last your comment clarified what you actually meant.

Limitation supposedly is that a graph actually doesn't belong to any host so it's hard to decide how to expand a lld macro when creating a graph.
More details why it's complicated to expand such macros you can find in ZBXNEXT-1967, which asks for host macros support in graph names

isn't this about usermacros instead of lld macros ?

zalex_ua Indeed, reporter actually described LLD macro nested in user macro.
I've adjusted all places where I was wrong.

Have a nice day!
What about this features?
This is good idea, very needed for effective use of monitoring nodes with user items names.

Hi all!
For example, in zabbix v3 macros support contexts, and this very usable for many sub-items in snmp tree (with Discovery).

In my case Discovery rule configured with:
SNMP OID -> discovery[

,x.x.x]
Items name: {$SLA_NAME.{#SNMPINDEX}}

On the host macros redefined:
{$SLA_NAME:1} -> Tunnel name 1
{$SLA_NAME:2} -> Tunnel name 2

Macros on the host {$SLA_NAME:1} work for Items, Triggers, but not for Graphs names.

Looks like I ran into this issue as well.

I tried creating a macro like : {$IF_DESC:ens192}  = WAN

Then using it in the graph prototypes but was unsuccessful. It would be nice to somehow have interface descriptions for UNIX machines.