Skip to content

ServiceRadar

Open navigation

Product announcement

November 20, 2026 Michael Freeman

From Browser to Wire: Real User Monitoring, Akamai Edge, and WAN Visibility in One Console

ServiceRadar v1.6.0 unites Akamai Boomerang Real User Monitoring, native CDN Server-Timing, W3C trace stitching, and embedded agent netperf so teams can tell in seconds whether slowness lives in the browser, at the Akamai edge, at the origin, or on a branch WAN link.

rum boomerang akamai netperf observability starrocks release

Every operations team knows the Monday-morning ticket: "The site feels slow." Within ten minutes, four different teams are looking at four different dashboards. Frontend checks Core Web Vitals and blames the CDN. The CDN team sees healthy cache ratios and points at the origin. Backend checks CPU graphs and sees nothing. Meanwhile, nobody has checked whether the branch office reporting the issue is saturating its WAN link.

In ServiceRadar v1.6.0, we are closing that gap. By bringing together Akamai Boomerang Real User Monitoring, native Akamai CDN timing telemetry, and embedded agent network performance testing (netperf), ServiceRadar gives you one continuous view from the user's browser all the way down to the wire.

Figure 1 · Browser to Wire

One continuous signal across browser, Akamai edge, origin, and branch WAN

Passive browser beacons, CDN Server-Timing headers, and private WAN probes converge into the same real-time warehouse.

BROWSER EDGE / CDN NETWORK PLATFORM Akamai Boomerang RUM LCP · INP · CLS · FCP · JS Errors Already on Akamai mPulse? 5-line dual-beacon tee · zero second agent Akamai CDN Server-Timing HIT / MISS · Edge RTT · Origin RTT · PoP W3C Trace Context Stitching Slow page load → backend OpenTelemetry trace Embedded Agent netperf & MTR Private WAN throughput · loss · jitter Branch Prefix Tags & TLS Posture Site cohort attribution · cert expiry checks ServiceRadar Web RUM + Network Performance JetStream-first ingest · StarRocks analytics · SRQL in:websites & in:rum · PII-scrubbed

Why we built on Akamai Boomerang instead of reinventing a browser agent

Measuring real browsers accurately is deceptively hard. Single-page apps, back-forward cache restores, interaction latency (INP), layout shifts (CLS), and compressed resource timing have years of edge cases behind them. Rather than shipping a fragile homegrown snippet, ServiceRadar bundles the battle-tested, open-source Akamai Boomerang library with a thin ServiceRadar trace plugin.

Privacy-first by default

Query strings, session tokens, and URL fragments are stripped at our Rust ingress gate before a beacon ever touches storage. Client IPs are anonymized after site and ASN attribution, and strict CSP sites run with zero unsafe-inline.

First-class Akamai CDN visibility

When your properties sit behind Akamai Delivery, ServiceRadar parses W3C Server-Timing headers directly into Edge RTT, Origin RTT, Cache HIT/MISS, and PoP — so you see instantly whether latency happened at the edge or behind it.

Coexists with Akamai mPulse

Already running Akamai mPulse? You do not need to load a second RUM script. A five-line subscription hook tees beacons to ServiceRadar alongside mPulse so your network and SRE teams share the exact same experience data.

Clean L7 Web Property inventory

Websites are modeled as first-class L7 Web Properties, never crammed into your L2/L3 device table. Rotating CDN Anycast VIPs stay out of device inventory while still linking down to Kubernetes gateways, origin hosts, and TLS posture.

Ending the "is it the site or the branch WAN?" debate

Traditional APM tools stop at the load balancer. Traditional network monitors stop at interface counters. Neither can tell you whether a spike in Largest Contentful Paint came from an uncached origin query or a saturated branch circuit.

In ServiceRadar v1.6.0, your edge agents now include an embedded, mTLS-authenticated netperf engine for private site-to-site throughput, jitter, and packet-loss testing over WireGuard or MPLS — without ever exposing a bandwidth port to the public internet. Combine that with Prefix Tags for your branch egress networks, and every browser beacon automatically tags which office it came from.

Figure 2 · Root-Cause Split

Three "slow site" tickets, three completely different causes

Side-by-side latency budgets separate healthy edge delivery, origin misses, and congested branch WAN links in seconds.

1. Akamai Edge HIT LCP 310 ms · Healthy Edge 14 ms · Render 296 ms 2. Origin Cache MISS LCP 1.4 s · Trace linked Origin 980 ms → OTel span 3. Branch WAN Congestion site:branch-01 · netperf loss WAN 1.2 s · 3.4% loss Branch WAN (netperf / MTR) Akamai Edge RTT Origin RTT Browser Render

One open platform instead of five consoles

Because every RUM beacon, CDN timing header, synthetic HTTP/TLS journey, and private netperf sample lands in the same JetStream-to-StarRocks pipeline, you query them with the same SRQL bar, view them in the same Blueprint dashboards, and alert on them with the same rules engine — without paying per-seat, per-beacon SaaS overages.

Try it live today

ServiceRadar's own marketing sites are already instrumented and streaming live beacons into our public demo cluster. Explore the Web RUM dashboard or read the setup docs to instrument your own properties in minutes.