Time trend · VPN state
Same structure as Overall · every line split by VPN state (top 5) · Last 7 days
Basic info
Call quality
World map
Country corridors
Breakdown
Root cause
✨ AI insight
Explorer
Total calls
1.24M
last 7d · ▲ 3.2%
Unanswered
287K 23.1%
last 7d · ▼ 1.4%
Answered
953K 76.9%
last 7d · ▲ 5.1%
Active users
89.4K
last 7d · ▲ 2.7%
Total calls · by VPN state
Unanswered · by VPN state
Answered · by VPN state
Active users · by VPN state
TotalCall share · by VPN state
Level 0
Total bad call · by VPN state
Total = Not ring + Bad quality · 5 lines, top 2 VPN state
Total bad call rate · by VPN state
Overall (all VPN state)
22.4%
Level 1
Split: Not ring vs Bad quality · by VPN state
Call not ring rate · by VPN state 11.2%
Bad quality rate · by VPN state 11.1%
Level 2
Breakdown by issue family · all multi-line by VPN state
Delayed ring (S02) · by VPN state 8.1%
Silence (S01) · by VPN state 3.2%
Video bad · by VPN state 13.8%
Audio bad · by VPN state 10.5%
Connection bad · by VPN state 5.9%
Level 3
Quality detail · all multi-line by VPN state
Video / Video calls
Video freeze > 5%
Black screen
Low res (≤240p)
Audio / Answer call
Audio freeze > 5%
No audio
Connecting / Answer call
RTT > 500ms
Connecting > 5%
Connecting fail
Heatmap metric
Hover a country for details
📞 Call count
⚠️ Bad call rate
Call count by country · top-down
0
421K
data via mock · 36 countries
Top 10 countries
| Country | Calls | Bad % |
|---|
Origin
UAE outbound · 9 destinations · scroll for detail
Priority matrix · UAE outbound
X = call volume · Y = bad call rate · bubble = destination · top-right = priority to fix
Outbound destinations
| Dest | Calls | Bad % | vs avg |
|---|
UAE → all destinations · arc width = volume · arc color = bad rate tier
Each pie shows that dimension's share of total calls in the selected window (last 7 days)
Country
City
OS
App version
SDK type
Network type
Device temperature
Call mode
Call star rating
VPN state
Roaming state
User tier
Cohort
VIP
Wallet user TBD
KYC status TBD
A/B group
Top 8 issues account for 56% of bad calls in the selected window · root cause is from server-side attribution + client SDK logs
A02No audio638 calls · 12.2% of bad
Backend audio stream drop
40%
Client mic permission denied
28%
Network packet loss
18%
SDK encoder crash
10%
Other / unknown
4%
V01Video freeze > 5%520 calls · 9.9% of bad
Bandwidth drop
48%
Device CPU throttle
28%
WiFi → cellular handoff
16%
Other
8%
C01RTT > 500ms487 calls · 9.3% of bad
Cross-region routing
42%
Cell tower congestion
31%
VPN added latency
18%
Other
9%
S02Delayed ring412 calls · 7.9% of bad
FCM / APNS push delay
52%
Server queue backlog
26%
Client wake-up slow
14%
Other
8%
V02Black screen288 calls · 5.5% of bad
Video decoder init fail
45%
GPU driver issue
24%
Camera permission denied
20%
Other
11%
S01Silence / never ring245 calls · 4.7% of bad
Push token expired
38%
Carrier delivery fail
29%
User offline
22%
Other
11%
A01Audio freeze > 5%198 calls · 3.8% of bad
Audio packet loss
44%
Echo cancellation buffer
25%
SDK audio worker hang
20%
Other
11%
C02Connecting > 5%143 calls · 2.7% of bad
SIP signaling slow
47%
Edge server cold start
28%
DNS resolution slow
16%
Other
9%
Executive summary
Bad call rate is 11.8% over the past 7 days — slightly below the prior 7-day baseline of 12.2%, but two new regressions are pulling the trend back up in the last 48 hours.
The Pakistan and Nigeria corridors remain the highest pain points (14.2% and 18.5% respectively); both are driven primarily by audio issues.
India outbound has been stable on the volume side, but iOS 4.10.0 users on 4G are seeing more video-freeze events week-over-week, likely related to the recent WebRTC SDK upgrade rollout.
Key findings 3 detected
REGRESSION
Pakistan bad-rate spike · +23% over 3 days
AE → PK corridor climbed sharply Tue–Thu, concentrated in Android 4.10.0 users on whatrtc_v4. 78% of the increase attributes to "Backend audio stream drop" — coincides with the SGBOT nodejs service migration deployed on Tue 4/26.
EMERGING
iOS 4.10.0 + India + 4G video-freeze regression · +18% WoW
Video-freeze events grew week-over-week in a tight cohort: iOS 4.10.0 / India / 4G. Likely related to the WebRTC SDK upgrade that shipped this version. ~142 affected calls, with 67% of those experiencing freeze duration > 5s (vs 41% baseline).
→ View affected cohort in Explorer
IMPROVEMENT
Call setup success improved · +0.6pp
Delayed-ring and Silence rates both dropped after the FCM/APNS push-token refresh shipped on 4/24. ~290 fewer not-ring sessions in the last 7 days. The improvement is concentrated in Android users in UAE and Saudi Arabia.
Recommended actions
- Priority — investigate SGBOT migration impact on PK Android audio path. Coordinate with engineering (@Peiyong-Yang for Open Platform). Backend audio stream drop accounts for 78% of the Pakistan spike. Suggested SQL filter:
country=PK AND os=Android AND appver=4.10.0 AND issue=A02. - A/B-gate the WebRTC SDK upgrade for India iOS 4G cohort. Current bad-rate trending +18% WoW. Either roll back this segment to 4.9.2 or split-test against the current version with a 50/50 ramp.
- Expand the push-token refresh to more regions. Current ramp is limited; the +0.6pp setup-success improvement suggests broader rollout would help. Suggested next: Egypt, India, Philippines.
- Set up auto-alert on the Pakistan corridor. The +23% jump went unnoticed for 24h; an alert at +10% would have caught it before customer complaints arrived.
Predicted trajectory · next 7 days
If no intervention, the bad call rate is forecast to reach 12.8% by next Sunday, driven primarily by Pakistan and India regressions. If the recommended actions are taken, the forecast drops to 10.4%.
⛔ Without intervention
12.8%
95% CI: 11.9% – 13.7%
✓ With actions above
10.4%
95% CI: 9.8% – 11.1%
⚠ AI-generated analysis · cross-check with raw data in Explorer before acting on recommendations