Module yeti

Module yeti implements SBC functions for Yeti-switch project

module "yeti" {
    # pop_id = 0

    # db_refresh_interval = 300

    # msg_logger_dir = /var/spool/sems/dump

    # audio_recorder_dir = /var/spool/sems/record
    # audio_recorder_compress = true
    # audio_recorder_http_destination =

    # core_options_handling = true
    # pcap_memory_logger = false

    # ip_auth_header = X-ORIG-IP
    # ip_auth_reject_if_no_matched = false

    # http_events_destination =

    # write_internal_disconnect_code = false

    # postgresql_debug = false

    # supported_tags = {}
    # allowed_methods = {INVITE, ACK, BYE, CANCEL, OPTIONS, NOTIFY, INFO, UPDATE, PRACK}

    /*
    routing {
        # schema = switch21
        # function = route_release
        # init =

        # new_codec_groups = true
        # pass_input_interface_name = false
        # connection_lifetime = 0

        /*
        master_pool {
            # host = 127.0.0.1
            # port = 5432
            # name = yeti
            # user = yeti
            # pass = yeti

            # size = 10
            # check_interval = 25
            # statement_timeout = 5
        }
        */

        # failover_to_slave = false
        /*
        slave_pool {
            # host = 127.0.0.1
            # port = 5432
            # name = yeti
            # user = yeti
            # pass = yeti

            # size = 10
            # check_interval = 25
            # statement_timeout = 5
        }
        */

        /*
        headers {
            # header(header_name[, sql_type = varchar, [format[, format_param]])
        }
        */
    }
    */

    /*
    cdr {
        # pool_size = 10
        # batch_size = 50
        # batch_timeout = 10000

        # auth_pool_size = pool_size
        # auth_batch_size = batch_size
        # auth_batch_timeout = batch_timeout

        # check_interval = 25

        # connection_lifetime = 0

        # schema = switch
        # function = writecdr

        /*
        master {
            # host = 127.0.0.1
            # port = 5432
            # name = cdr
            # user = cdr
            # pass = cdr
        }
        */

        # failover_to_slave = true

        /*
        slave {
            # host = 127.0.0.1
            # port = 5432
            # name = cdr
            # user = cdr
            # pass = cdr
        }
        */
    }
    */

    /*
    resources {
        # reject_on_error = false
        # reduce_operations = false
        # scripts_dir = /usr/lib/sems/scripts/yeti
        /*
        write {
            # hosts = 127.0.0.1:6379
            # timeout = 5000
            # username =
            # password =
        }
        */
        /*
        read {
            # hosts = 127.0.0.1:6379
            # timeout = 5000
            # username =
            # password =
        }
        */
    }
    */

    /*
    lega_cdr_headers {
        # header(header_name, none|string|array|smallint|integer [, active_call_key, String])
        # add_sip_reason = false
        # add_q850_reason = false
    }
    */

    /*
    legb_cdr_headers {
        # header(header_name, string|array|smallint|integer)
    }
    */

    /*
    legb_response_cdr_headers {
        # header(header_name, string|array|smallint|integer)

        # add_sip_reason = false
        # add_q850_reason = false
    }
    */

    /*
    rpc {
        # calls_show_limit = 100
    }
    */

    /*
    statistics {
        /*
        active-calls {
            # period = 60
            /*
            clickhouse {
                # table = active_calls
                # destinations = {}
                # buffering = false
                # allowed_fields = { }
            }
            */
        }
        */
    }
    */

    /*
    auth {
        # realm = hostname
        # skip_logging_invite_success = false
        # skip_logging_invite_challenge = false
    }
    */

}

pop_id (default: 0)

Point of presence id

db_refresh_interval (default: 300)

Database refresh timer in seconds. Sets period between check_states() DB calls to check entities versions. see: Entities caching

msg_logger_dir (default: /var/spool/sems/dump)

Directory to store *.pcap files

audio_recorder_dir (default: /var/spool/sems/record)

Directory to store *.mp3 or *.rsr files. Works for audio_recorder_compress = true only.

audio_recorder_compress (default: true)

Sets recordings storing format.

• true. store MP3 audio file for each leg:
  • {audio_recorder_dir}/{global_tag}_{node_id}_lega.mp3
  • {audio_recorder_dir}/{global_tag}_{node_id}_legb.mp3
• false. store single multichannel RSR file
  • {general.rsr_path}/{global_tag}.rsr

audio_recorder_http_destination (default: empty)

Destination name in the http_client module to automatically upload RSR files. Works for audio_recorder_compress = false only.

core_options_handling (default: true)

Handle SIP OPTIONS requests on SEMS core level, do not forward such requests to yeti application

pcap_memory_logger (default: false)

Enable workaround to write 100 Trying sent before session creation to the PCAP file

ip_auth_header (default: X-ORIG-IP)

Set tech SIP header name to lookup for source IP in the origination pre-auth routine

ip_auth_reject_if_no_matched (default: false)

Whether to pass origination requests not matched by pre-auth to the DB routing

http_events_destination (default: empty)

Destination name in the http_client module to call HTTP hooks for started, connected, disconnected call events

write_internal_disconnect_code (default: false)

Add field disconnect_code_id with internal call disconnect reason after the disconnect_rewrited_reason

postgresql_debug (default: false)

Log getprofile, cdr, auth_log requests and getprofile replies at the debug loglevel

supported_tags (default: {})

Additional tags to add to the Supported header (timer and 100rel will be added automatically by the related handlers)

allowed_methods (default: {INVITE, ACK, BYE, CANCEL, OPTIONS, NOTIFY, INFO, UPDATE, PRACK})

Methods list to declare in the Allowed header

Section routing

schema (default: switch21)

Database schema to be added to the search_path in addition to the public for the routing connections

function (default: route_release)

Database function to be called to get call profile

init (default: empty)

Function to run for routing connection initialization. Usually is used to initialize yeti pg extension to process LNP requests

new_codecs_groups (default: true)

What function to use for codecs preloading

• true. use load_codec_groups()
• false. use load_codecs()

pass_input_interface_name (default: false)

Add input interface name to the DB routing requests

connection_lifetime (default: 0)

Force reconnect each [connection_lifetime]{.title-ref} seconds for routing connections. Used as workaround on memory leaks in PG backends

header(header_name[, sql_type = varchar, [format[, format_param]])

Add SIP header from initial incoming SIP request to be used for routing DB requests. Function can be used multiple times. Overrides load_interface_in() usage

parameter	description
header_name	SIP header name to search in the initial incoming SIP request. Name will be normalized: converted to lowercase, - replaced with _
sql_type (default: varchar)	SQL type name for the value in the routing request
format (optional)	Parse header and use part specified by format as value. allowed values: uri_user, uri_domain, uri_port, uri_json, uri_param. Each format value can be suffixed by _array to get an array of the headers values instead of the first one, so we have uri_user_array ... uri_param_array in addition
format_param (required for uri_param, uri_param_array)	uri param name to extract from parsed URI

failover_to_slave (default: false)

Enable failover for routing DB requests. Slave connections pool is configured in the [routing.slave_pool]{.title-ref} subsection

Sub-sections routing.master_pool, routing.slave_pool

Connection pools parameters for routing

host (default: 127.0.0.1)

Database host

port (default: 5432)

Database port

name (default: yeti)

Database name

user (default: yeti)

Database user

pass (default: yeti)

Database password

size (default: 10)

Connections pool size

check_interval (default: 25)

Reconnect delay on connection errors in seconds

statement_timeout (default: 5)

Routing transaction timeout in seconds

Section cdr

pool_size (default: 10)

Connections pool size

batch_size (default: 50)

cdr count for single transaction. We collect batch_size CDRs before writing to optimize database usage

batch_timeout (default: 10000)

cdr batch timeout in milliseconds. Write CDRs after the batch_timeout even if batch_size is not reached

auth_pool_size (default: 10)

Connections pool size

auth_batch_size (default: 50)

auth_log count for single transaction

auth_batch_timeout (default: 10000)

auth_log batch timeout in milliseconds

check_interval (default: 25)

Reconnect and retransmit delay in seconds

connection_lifetime (default: 0)

Force reconnect each [connection_lifetime]{.title-ref} seconds for cdr, auth_log connections. Used as workaround on memory leaks in PG backends

schema (default: switch)

Database schema to be added to the search_path in addition to the public for the cdr and auth_log connections

function (default: writecdr)

Database function to be called to write CDRs (function for auth_log is hard-coded to be write_auth_log)

failover_to_slave (default: false)

Enable failover for cdr and auth_log DB requests. Slave connections pool is configured in the [cdr.slave]{.title-ref} subsection

Sub-sections cdr.master, cdr.slave

Connection pools parameters for cdr and auth_log

host (default: 127.0.0.1)

Database host

port (default: 5432)

Database port

name (default: cdr)

Database name

user (default: cdr)

Database user

pass (default: cdr)

Database password

Section resources

reject_on_error (default: false)

Whether to reject calls on resources accounting errors (e.g redis availability)

reduce_operations (default: false)

Merge increments/decrements for the same resources to decrease requests count to redis. Recommended value: true. Default is false historically because of testing

scripts_dir (default: /usr/lib/sems/scripts/yeti)

Directory to lookup for resources management redis LUA scripts

Sub-sections resources.read, resources.write

Redis connections configuration for resources

hosts (default: 127.0.0.1:6379)

Endpoint for redis connections

timeout (default: 5000)

Timeout for redis requests in milliseconds

username (default: empty)

username for redis AUTH command

password (default: empty)

password for redis AUTH command

Section lega_cdr_headers

header(header_name, serialization_type: none|string|array|smallint|integer [, active_call_key, active_call_key_type = String])

Add SIP header value from initial incoming SIP request to be saved in CDR. Function can be used multiple times

parameter	description
header_name	SIP header name to search in the initial incoming SIP request. Name will be normalized: converted to lowercase, - replaced with _.
serialization_type	Type name for the value in CDR writing request. none means skip saving to CDR and used together with active_call_key
active_call_key (optional)	field name to add header value to the active calls and statistics snapshots (see: statistics section)
active_call_key_type (required for active_call_key)	active call key type name. only String is allowed for now

add_sip_reason (default: false)

Parse Reason header from legA BYE request. Save serialized SIP cause, text, params as value for 'reason' field in legA headers CDR field

add_q850_reason (default: false)

Parse Reason header from legA BYE request. Save serialized Q.850 cause, text, params as value for 'reason' field in legA headers CDR field

Section legb_cdr_headers

header(header_name, serialization_type: string|array|smallint|integer)

Add SIP header value from legB outgoing SIP request to be saved in CDR. Function can be used multiple times

parameter	description
header_name	SIP header name to search in the SIP requests sent from legB. Name will be normalized: converted to lowercase, - replaced with _
serialization_type	Type name for the value in CDR writing request.

Section legb_response_cdr_headers

header(header_name, serialization_type: string|array|smallint|integer)

Add SIP header value from legB 200 OK reply to be saved in CDR. Function can be used multiple times

parameter	description
header_name	SIP header name to search in the Bleg 200 OK SIP reply. Name will be normalized: converted to lowercase, - replaced with _
serialization_type	Type name for the value in CDR writing request

add_sip_reason (default: false)

Parse Reason header from legB BYE requests and error/redirect replies (code >= 300). Save serialized SIP cause, text, params as value for 'reason' field in legB headers CDR field

add_q850_reason (default: false)

Parse Reason header from legB BYE requests and error/redirect replies (code >= 300). Save serialized Q.850 cause, text, params as value for 'reason' field in legB headers CDR field

Section rpc

calls_show_limit (default: 100)

Active calls limit for yeti.show.calls and yeti.show.calls.filtered jsonRPC methods output

Section statistics

Configure periodic reporting of the active calls to the clickhouse. Set both clickhouse.table and clickhouse.destinations to enable functionality

Sub-section statistics.active-calls

period (default: 60)

Interval between active calls snapshots sending in seconds

Sub-section statistics.active-calls.clickhouse

table (default: active_calls)

Clickhouse table name for inserting. Resulting query first line will be: INSERT INTO {table} FORMAT JSONEachRow

destinations (default: empty)

Destinations names list for the http_client module. Destinations should be configured with mode=post and content_type='application/vnd.api+json'

buffering (default: false)

Buffer small calls which were started and finished between snapshots and merge them into the nearest snapshot

allowed_fields (default: {})

Reduce fields for active calls data to the specified ones. Send all available fields if empty

Section auth

Configure authorization for incoming SIP requests

realm (default: hostname)

Set realm to match for SIP digest authorization. Will use hostname if not specified

skip_logging_invite_success (default: false)

Do not write auth_log for successfully authorized requests

skip_logging_invite_challenge (default: false)

Do not write auth_log for incoming requests without Authorization header that were challenged with 401 reply