record captured SIP provider failure

2026-09-17 14:41:20 +08:00
parent 668c12e523
commit 84d86319a3
2 changed files with 93 additions and 0 deletions
@@ -0,0 +1,91 @@
+{
+  "record_version": "1.0",
+  "recorded_at_utc": "2026-09-17T06:37:48Z",
+  "scope": "packet-captured root-cause validation for mka75518625770806",
+  "code_commit": "668c12e",
+  "deployment": {
+    "region": "cn-beijing",
+    "instance_id": "i-2zeimnh6hq53q4srr3o2",
+    "private_ip": "172.16.0.123",
+    "public_ip": "123.56.71.98",
+    "eip_allocation_id": "eip-2zeevfsaxzwuue2szy7xb",
+    "asterisk_network": "host",
+    "asterisk_container": "agent-call-asterisk"
+  },
+  "security_group": {
+    "provider_signaling_source": "160.202.254.79/32",
+    "provider_media_source_observed_before": "111.9.126.254/32",
+    "diagnostic_udp_scope": "1/65535 from the two provider IPs only",
+    "global_udp_open": false
+  },
+  "call": {
+    "original_number": "18625770806",
+    "callee_prefix": "mka755",
+    "effective_sip_user": "mka75518625770806",
+    "sip_destination": "160.202.254.79:5060",
+    "caller_id": "KQ91526",
+    "execution_id": "exec_cell_1789627067521622369_1",
+    "call_id": "call_4c435bdbf3af488884b9ad879b8045b3",
+    "started_at_utc": "2026-09-17T06:37:44Z",
+    "ended_at_utc": "2026-09-17T06:37:48Z",
+    "attempts": 1,
+    "retries": 0,
+    "status": "failed_before_answer",
+    "reason_code": "CALL_NOT_ANSWERED",
+    "media_packets_in": 0,
+    "media_packets_out": 0,
+    "turns": [],
+    "recording_valid": false
+  },
+  "packet_capture": {
+    "pcap": ".local/agent-call-third/capture-call2-20260917.pcap",
+    "pcap_sha256": "09f2707ccab64948a07d06be4150a1cda2469b151efc39ca2be2d574bec222f82",
+    "decoded_summary": ".local/agent-call-third/capture-call2-packets-20260917.txt",
+    "decoded_summary_sha256": "dd5438d142169b8344fa2c2fce1c243aac53506d2e244cd7df722d2895ff11aa",
+    "packet_count": 20,
+    "provider_signaling_packets": 4,
+    "provider_media_ip_packets": 0,
+    "local_rtp_range_packets": 0,
+    "sip_exchange": [
+      "172.16.0.123:5060 -> 160.202.254.79:5060 INVITE sip:mka75518625770806@160.202.254.79:5060",
+      "160.202.254.79:5060 -> 172.16.0.123:5060 100 Trying",
+      "160.202.254.79:5060 -> 172.16.0.123:5060 480 Temporarily Unavailable",
+      "172.16.0.123:5060 -> 160.202.254.79:5060 ACK"
+    ],
+    "timing": "480 arrived about 465 ms after 100 Trying",
+    "sdp_sent": {
+      "public_media_address": "123.56.71.98",
+      "local_rtp_port": 10076,
+      "codec": "PCMA/8000"
+    },
+    "provider_observed_transport": "Via received=123.56.71.98; rport=5060"
+  },
+  "asterisk_observation": {
+    "external_media_channel": "UnicastRTP/127.0.0.1:45667 answered and joined the bridge",
+    "target_channel": "Called mka75518625770806@provider-primary",
+    "target_bridge_state": "never reached answered media state",
+    "host_firewall": "no matching nft rule observed",
+    "final_active_channels": 0,
+    "asterisk_log": ".local/agent-call-third/capture-call2-asterisk-20260917.log",
+    "asterisk_log_sha256": "67954f3a5b12f6fae0f3c637c9cec5607f7a57c4859a1f54daf104c0a7a58ced"
+  },
+  "root_cause": {
+    "confirmed_for_this_attempt": "SIP provider returned 480 Temporarily Unavailable before answering; no 200/SDP answer was sent, therefore no RTP could exist",
+    "our_side": "EIP/NAT, SIP signaling source, outbound SDP address, PCMA offer, security-group source scope, externalMedia initialization, and ACK path were observed working",
+    "provider_side_or_line_contract": "provider route/availability, account/caller authorization, or the required called-number format/prefix mka755 remains the active fault boundary",
+    "not_proven": "The earlier call that reached 200 OK but showed zero inbound RTP was not reproduced in this capture; this pcap cannot determine post-200 RTP behavior"
+  },
+  "actions": {
+    "our_code_fix": "committed as 668c12e",
+    "security_group": "source-limited full UDP rules retained for diagnosis",
+    "automatic_redial": false,
+    "temporary_cell_rabbitmq_capture_cleanup": true,
+    "ecs_and_eip_retained": true
+  },
+  "next_request_to_provider": [
+    "Confirm whether mka755 must be prepended to 18625770806 on this trunk",
+    "Confirm KQ91526 is authorized for this destination and line",
+    "Check why VOS3000 returned 480 Temporarily Unavailable and whether the route was available at 2026-09-17 06:37:47 UTC",
+    "For a call that reaches 200 OK, provide the actual RTP source IP/CIDR and confirm that RTP is sent to 123.56.71.98"
+  ]
+}
@@ -228,6 +228,8 @@ python3 deploy/render_asterisk.py --config .local/asterisk.json

 真实 Cell 的 `externalMedia` 通道创建后，变量可能要等 `StasisStart` 才可查询；必须在通道就绪后重试读取 `UNICASTRTP_LOCAL_ADDRESS` 与 `UNICASTRTP_LOCAL_PORT`，以此建立向 Asterisk 的 RTP 发送目标。`addChannel` 在通道未进入 Stasis 时的 `400/404/409/422` 只能有限重试，成功后才能标记已入桥；收到目标 `Up` 后还必须通过 `GET /bridges/{id}` 确认 externalMedia 与目标都在同一桥内，不能只看 `Up` 就开始 AI。RTP 接收需校验来源 peer、PCMA payload type 和 SSRC。ARI originate 的 `timeout` 单位是秒，不是毫秒。实时录音必须先 `POST /recordings/live/{name}/stop`，再轮询读取 stored file，不能用 DELETE 丢弃录音后伪装成有效结果。

+2026-09-17 对第三组 `mka75518625770806` 做带 PJSIP logger 的 UDP 抓包：EIP/NAT、5060 信令、`From KQ91526`、PCMA SDP 和 ACK 均正常；服务商 `160.202.254.79` 在收到 INVITE 后返回 `100 Trying`，约 465ms 后返回 `480 Temporarily Unavailable`，未发送 200/SDP，因此该次没有 RTP 是“呼叫未建立”的必然结果，不应继续从端口范围或 Cell 音频链路猜测。抓包记录见 `docs/evidence/third-sip-capture-root-cause-20260917.json`；后续需服务商确认线路可用性、主叫授权及 `mka755` 被叫格式。
+
 ```bash
 # 默认为检查，不启动容器。
 bash deploy/asterisk.sh