📋 Technical documentation for CompactEdge AI: View Spec Sheet → ← Back to product page
Live
Edge Analytics
Real-time latency, cache performance, and bandwidth metrics across all 180 PoPs.
Requests / sec
284k
↑ 12% vs yesterday
Avg Edge Latency
5.2ms
↓ 0.4ms vs last hour
Cache Hit Rate
97.8%
↑ 0.3pp vs yesterday
Bandwidth (24h)
8.4TB
↑ 8% vs yesterday
Latency Percentiles — Last 60min
Live
● P50: 5.2ms ● P95: 18.4ms ● P99: 42.1ms
Cache Hit Rate — Last 24h
97.8% avg
● Cache Hits: 97.8% ● Misses: 2.2%
Top Routes by Traffic
RouteReq/secLatencyCache
/api/products48,2044.1ms99.1%
/static/assets/*42,8012.8ms99.8%
/api/users31,4007.2ms81.4%
/images/*28,9903.3ms98.4%
/api/search21,30211.8ms74.2%
Bandwidth by Region (24h)
North America 3.2 TB
Europe 2.4 TB
Asia Pacific 1.9 TB
South America 0.6 TB
Other 0.3 TB
Edge Node Topology
180 PoPs across 70+ countries — all operational. Hover nodes to view real-time stats.
Total Nodes
180
All operational
Healthy
177
98.3% healthy
Degraded
3
Auto-rerouting active
Avg Node Latency
5.8ms
↓ 0.2ms vs last hour
Healthy Degraded
All Nodes
LocationRegionStatusLatency (P50)Req/secCache HitUptime (30d)
New York, USNA-East● Healthy4.2ms48k99.4%99.99%
Los Angeles, USNA-West● Healthy3.8ms42k99.1%99.99%
London, UKEU-West● Healthy5.1ms38k98.8%100%
Frankfurt, DEEU-Central● Healthy4.7ms34k97.9%99.98%
SingaporeAPAC● Healthy6.2ms31k97.4%99.97%
Tokyo, JPAPAC● Healthy5.8ms28k97.8%99.99%
Stockholm, SEEU-North● Degraded22.4ms8k91.2%99.82%
São Paulo, BRSA● Healthy8.9ms14k96.2%99.95%
Sydney, AUAPAC● Healthy7.4ms12k97.1%99.96%
Mumbai, INAPAC● Healthy7.8ms16k96.8%99.94%
AI Optimization Engine
Configure AI routing rules, compression settings, and cache policies. Changes apply globally within 30 seconds.
AI Routing Efficiency
94.2%
Optimal path selected 94.2% of requests
↑ 2.1% this week (AI learning)
Compression Savings
72%
Average bandwidth reduction
↑ 5% after Brotli upgrade
Cache Efficiency Score
97.8%
Predictive pre-warming active
↑ 3.4% since pre-warm enabled
🧠 AI Routing Rules
Auto-Learning Active
Predictive Congestion Avoidance
ML model predicts network congestion 2–5 seconds ahead and pre-routes traffic away from congested paths before packets are dropped.
Impact: −1.8ms avg latency · Updated: 500ms cycles · Confidence: 97.2%
BGP Path Multi-Homing
Evaluates all available BGP paths and selects the one with lowest RTT + packet loss, regardless of standard routing tables.
Impact: −0.9ms avg latency · Active on: 180/180 nodes
Origin Shield Auto-Selection
AI dynamically selects the optimal origin shield PoP based on current network topology, reducing origin load by up to 90%.
Current shield: Frankfurt → reduces origin hits by 88%
Latency-Based Failover
Automatically fails over to secondary node if primary node latency exceeds 2× the rolling average. Failover completes in <100ms.
Threshold: 2× rolling avg · Last triggered: 4h ago (Stockholm)
📦 Compression Settings
Brotli Compression (Level 11)
Brotli is enabled at maximum compression level for all compressible content types (HTML, CSS, JS, JSON, SVG). Average 20% better than gzip.
Coverage: 68% of responses · Avg reduction: 78%
WebP/AVIF Image Conversion
JPEG and PNG images are automatically converted to WebP (or AVIF for supported browsers) at the edge. No origin changes required.
Coverage: 24% of responses · Avg reduction: 65% vs JPEG
JS/CSS Minification
Unminified JavaScript and CSS are minified at the edge using an LLVM-based optimizer. Works on cached content without touching your build pipeline.
Status: Disabled — enable only for non-minified origins
🗄 Cache Policies
Pre-warming Active
Predictive Cache Pre-warming
AI analyzes historical traffic patterns and pre-populates cache at relevant PoPs before traffic arrives. Eliminates cold-start latency for predicted content.
Pre-warm accuracy: 94.8% · Cache hit rate improvement: +3.4pp
Dynamic TTL Optimization
AI adjusts cache TTLs based on content change frequency. Static assets get extended TTLs; frequently-updated API responses get shorter TTLs automatically.
Avg TTL: 4.2h (up from 1h static baseline)
Stale-While-Revalidate
Serve stale cached content immediately while asynchronously fetching a fresh copy. Users never wait for cache revalidation.
Grace period: 60 seconds · Applies to: all cacheable content
Edge Provider Integration
Connect and manage edge computing providers. Deploy unified across Cloudflare, Deno, KubeEdge, OpenNebula, and EdgeCDN-X.
🚀 Deploy to Edge Providers
Unified Deploy
Alert Configuration
Configure threshold-based or AI-triggered alerts. Alerts deliver context — not just noise.
➕ New Alert Rule
Active Alert Rules
🔴
P99 Latency Critical
Fires when P99 latency exceeds 100ms for more than 60 seconds
Critical PagerDuty + Slack Last fired: 3h ago (Stockholm node)
🟡
Cache Hit Rate Drop
Fires when cache hit rate drops below 90% on any PoP for 5+ minutes
High Email + Slack Last fired: 2d ago
🟡
Node Health Degraded
Fires when any node's health score falls below 80/100
High Slack Active now — Stockholm node at 72/100
🔵
Bandwidth Spike Warning
Fires when bandwidth exceeds 150% of 7-day rolling average (potential traffic surge)
Medium Email Last fired: Never
🟢
Error Rate Threshold
Fires when 5xx error rate exceeds 0.5% of total requests over any 10-minute window
High PagerDuty + Email Last fired: 5d ago
CompactEdge AI
Edge Optimization Assistant
Hi! I'm your edge optimization AI. I can analyze your current routing, flag underperforming nodes, and suggest cache configuration improvements. What would you like to optimize?
P99 latency analysis
Cache optimization
Node performance
Routing analysis