We’ve been using pbx_lua for years and it works extremely well. However, it hasn’t seen any updates in 2 years. It is possible that pbx_lua simply doesn’t need any patches, however, I have a few concerns.
Generally, AGI is the preferred method for dialplan integrations. It seems very few use pbx_lua. I’ve been encouraged in the past to migrate to AGI.
Will pbx_lua be depreciated? I don’t want to continue to develop against a module that is due for depreciation.
Even though I prefer pbx_lua to res_agi, I would rather migrate to res_agi now if the future for pbx_lua is uncertain. Thank You!
The pbx_lua module is extended support, meaning it receives no support from Sangoma and is community supported. No one has proposed deprecating or removing it.
agi is a big bottle neck freepbx had to switch to fast agi
had a big system 24 cpu bare metal could not get it more then 200 concurrent call using agi
on the other hand lua in freeswitch does a great job i would think twice before moving …
Very interesting. I actually have a smaller deployment using AGI with Java-Asterisk. I load tested it with 100s of SIP Invites per second and it performed really well. I just personally really dislike Java so I’ve avoided working with it again. Other languages (besides C) aren’t fast enough for my liking. The one language that has piqued my interest is Rust, but it’s still in its infancy.
Lua is working very well for us. If we were to start moving away from it, it would purely be a fear based move.
Another note, if it’s freepbx using AGI, then there’s a good chance they’re using agi with ancient PHP code. This is indeed a recipe for disaster. I can see it working well with well structured php 7+ code, but not the old stuff that you find in freepbx.