What is this
Attempt to provide access to Asterisk's AMI via Postgres/SQL
Features
- You can have access to the asterisk's ami from any programming languages which supports access to the postgresql.
- You can employ all power of sql language: sorting, filtering by where clauses, grouping, join resulting output of ami commands (like ordinary database tables) with the real tables.
- You can use your Asterisk as source of data for ETL processes. For example, to collect some stats.
Require
- Postgresql server with installed plpython
- Asterisk server with enabled and configured AMI
- Python and pip package manager
- Pyst python module (https://github.com/al-the-x/pyst)
On CentOS 6.3
- install repo from http://yum.postgresql.org/repopackages.php
- refresh metainfo: yum update
- yum install postgresql92-server postgresql92-plpython
Installation
-
Install dependencies:
sudo pip install https://github.com/al-the-x/pyst/tarball/master
-
Install python module:
sudo pip install https://github.com/frutik/asterisk-ami-pgsql/tarball/master
-
Create database (if not exists) and add plpython (if not exists):
createdb -U postgres asterisk
createlang plpythonu -U postgres asterisk
-
Install sql part:
psql -U postgres asterisk -f [TODO : add path to this file]install.sql
-
Add your server and required credentials into asterisk.managers table:
insert into asterisk.managers values ('192.168.1.1', 'me', 'mysecret');
Upgrade
- Download new version
- Perform install procedure again
Usage
Call White House's contact center
asterisk=# select asterisk.originate_async('localhost', 'SIP/117', '92024561111', 'default', 1);
originate_async
-----------------
t
(1 row)
Show 3 online peers
asterisk=# select * from asterisk.sippeers('localhost') where "IPport" <> '0' limit 3;
ipport | status | chanobjecttype | natsupport | objectname | realtimedevice | dynamic | acl | videosupport | channeltype | ipaddress | event
--------+------------+----------------+------------+------------+----------------+---------+-----+--------------+-------------+----------------+-----------
5060 | OK (5 ms) | peer | yes | 216 | no | yes | no | no | SIP | 10.1.10.4 | PeerEntry
2254 | OK (12 ms) | peer | yes | 215 | no | yes | no | no | SIP | 10.1.10.3 | PeerEntry
2253 | OK (12 ms) | peer | yes | 214 | no | yes | no | no | SIP | 10.1.10.3 | PeerEntry
(3 rows)
Check peer
asterisk=# select "Address-Port", "Status", "CodecOrder", "SIP-Useragent", "MaxCallBR", "Busy-level" from asterisk.sipshowpeer('127.0.0.1', '1117');
Address-Port | Status | CodecOrder | SIP-Useragent | MaxCallBR | Busy-level
--------------+-------------+--------------------+---------------+-----------+------------
5060 | OK (234 ms) | ulaw,alaw,g729,gsm | NetDial | 384 kbps |
(1 row)
Add agent to queue
select asterisk.queue_add('localhost', 'incomming_queue', 'AGENT/007');
Remove agent from queue
select asterisk.queue_remove('localhost', 'incomming_queue', 'AGENT/007');
Show queue status
asterisk=# select * from asterisk.queue_members('localhost', 'test_te') where "Name" = 'SIP/1113';
Queue | Name | Location | Membership | Penalty | CallsTaken | LastCall | Status | Paused
---------+----------+----------+------------+---------+------------+----------+--------+--------
test_te | SIP/1113 | SIP/1113 | dynamic | 1 | 0 | 0 | 5 | 0
(1 row)
Check customers waiting in queue
test=# select * from asterisk.queue_entries('127.0.0.1', 'test_te');
Queue | Position | Channel | CallerID | CallerIDName | Wait
---------+----------+---------------------+----------+--------------+------
test_te | 1 | SIP/430913-02c3aef0 | 430913 | 430913 | 69
(1 row)
Check general stats of queue
test=# select * from asterisk.queue_params('127.0.0.1', 'test_te');
Queue | Max | Calls | Holdtime | Completed | Abandoned | ServiceLevel | ServicelevelPerf | Weight
---------+-----+-------+----------+-----------+-----------+--------------+------------------+--------
test_te | 0 | 1 | 65 | 693 | 284 | 60 | 71.6 | 0
(1 row)
Join with the real table
You can do this. But examples will be provided later...
Limitations
- This solution is not for scaling of access to AMI - you still need use some kind of ami proxy.
- This solution "just works for me". This is not best example of python code.
- Keep in mind - some of those commands could be slow.
