Select or create a project to retain assessment snapshots in this browser.
Selected region comparison
Azure Region Pulse Assessment
Project No active project
Executive decision cockpit
Awaiting assessment
READY
Run an assessment to combine network evidence, resilience, cost, service fit and deployment readiness into a placement recommendation.
—Median RTT
—Confidence
—Reference monthly
—DR candidate
Decision status
Region eligibility pipeline
Shows how the candidate set narrows through scope, residency, mandatory architecture requirements, authenticated service availability and deployment readiness.
Awaiting assessment
Service availability becomes live evidence when an Azure subscription readiness check is performed with selected service requirements.
Customer decision priorities
Adjust the decision-support emphasis without changing the raw measurements. AVD always preserves strict lowest-median placement as the specialist recommendation.
Workload cost modeller
Build an indicative two-tier compute workload and add known monthly storage, backup and network costs. The model uses the same live regional Retail Prices data as the pricing comparison.
Reference model
Run an assessment firstThen calculate an indicative regional workload total.
Ancillary values are user-supplied budget assumptions and are not live Azure quotes. Compute uses the selected OS, currency, pricing model and hours from Live compute pricing.
Multi-site topology
Visualises saved source sites, their user weighting and the current weighted Azure recommendation.
No sites
Project intelligence
Track recommendation evolution, compare scenarios and share a compact read-only project snapshot.
Assessment evolution
Scenarios
Read-only links contain a compact project summary in the URL fragment. They do not include Azure tokens, IP/ISP identity, exact probe raw payloads or private access credentials.
Baseline monitoring mode
Use a saved baseline to answer “has anything materially changed?” after a repeat test.
No baseline
Save a baseline after a representative assessment, then use monitoring checks to detect winner changes, routing deterioration, confidence drops and material reference-pricing changes.
Azure placement & network assessment
Measure first. Place workloads smarter.
Azure region assessment using repeated HTTPS RTT samples, median, p95, timing variation, reliability and workload-aware scoring. By default, the tool runs a Quick test across European Azure regions using the General Azure workload profile. You can change the scope to another geography or Global at any time.
No additional data-residency restriction. Test scope controls which Azure regions are measured.
Architecture & service requirements
Mandatory requirements can exclude unsuitable regions. A region is shown as Eligible only against the requirements actually selected and verified for this assessment; outstanding checks are shown separately.
Adaptive Top-5 on
Mandatory placement requirements
Required Azure services · advisory verification
No service-specific requirements selected. Microsoft product availability remains a deployment-time verification step.
What does Eligible mean? The region meets the mandatory requirements selected for this assessment, such as residency, Availability Zones, a documented region pair, the reference VM listing and selected Azure services where they have been verified. Subscription deployment readiness is shown separately and is never implied when it has not been checked.
Source location
Not detected
Browser, manual place, or remote probe source.
Test ID
—
Reference for reports.
Completed
0
No test running.
Browser downlink
—
Connection API estimate only.
Previous winner
—
No local history yet.
Last test
—
Local browser time.
Ready to assess.
Primary region recommendation
Awaiting test
—
Run an assessment to generate a primary placement recommendation.
—Median RTT
—Availability zones
—Network quality
Suggested DR region
Not yet verified
—
The secondary recommendation considers region pairing, separation, resilience and measured network quality.
—Median RTT
—Region separation
—DR score
Primary and DR guidance will appear after the latency assessment. DR is a design recommendation, not a guarantee of service availability, capacity or compliance.
Practical interpretation
Run an assessment to identify materially equivalent regions.
Strict ranking is preserved; this layer prevents tiny RTT differences being overstated as architectural significance.
Recommendation confidence
—Awaiting evidence
Confidence combines separation from the runner-up, stability, reliability, sample depth, probe diversity and unresolved service requirements.
Geographic view
Actual country/coastline context with latency-coded Azure regions and a continuous interpolated RTT heat surface.
Hover or tap an Azure region marker to expand the map and inspect that region. The heat surface uses inverse-distance interpolation between measured Azure endpoints and fades as confidence falls with distance from a measured endpoint. It is an illustrative placement surface, not a direct latency measurement for every point on the map. Lines show geographic relationship only — they do not represent the actual ISP/BGP or Microsoft backbone route.
Azure region ranking
AVD keeps absolute median-first ranking. Other profiles use workload-aware network scoring. Mandatory requirements are applied before placement, Quick tests can adaptively deepen the Top 5, and recommendation confidence distinguishes statistical rank from practical significance.
AVD profileDeep test
Rank
Region
Distance
Median
p95
Variation
Success
Stability
Score
Workload
Zones
Paired region
Samples
No test results yet.
Final recommendation matrix
Combines network quality, cost, resilience, service fit, residency and deployment-readiness evidence. It explains the decision but does not override strict AVD median-first ranking.
Awaiting assessment
Run an assessment to build the placement matrix.
AVD specialist readiness
Azure Virtual Desktop-specific interpretation of the regional proxy, recommendation confidence and post-deployment validation path.
AVD specialist
Live Azure retail pricing comparison
Compare a representative VM SKU across the fastest measured regions using Microsoft's Azure Retail Prices API.
PAYG · Savings Plan · Spot
Run an assessment first
Pricing is compared across the fastest measured regions.
Compute-only retail comparison. Savings Plan values use Microsoft's 2023-01-01-preview Retail Prices API where the meter is eligible. Spot is variable and not suitable for every workload. Non-USD prices are budget references.
Latency vs cost frontier
Highlights the Pareto-efficient regions: candidates for which no other measured region is both faster and cheaper under the selected pricing model.
Load live pricing to calculate the latency/cost frontier.
Current region comparison
Compare the recommended placement with an existing Azure deployment.
Choose the current Azure region to quantify potential latency and cost change.
Azure subscription deployment readiness
Optional authenticated check against Microsoft.Compute Resource SKUs for the selected reference VM size. This can identify subscription/location restrictions and zones; it is not a guarantee of real-time capacity.
Not checked
No readiness check yetRun an assessment, then optionally authenticate to Azure.
Obtain a management-plane token with an identity that can read Compute SKU metadata. The token is sent only to the same-origin readiness proxy and is cleared from the form after the request.
Multi-site weighted placement
Build an enterprise view from several completed browser or remote assessments. Run a site, add the current result with its user population, then repeat for other offices. The planner compares only measured data and weights each site's latency by affected users.
0 sites
No sites added yet.
For a fair weighted comparison, regions with measurements covering all site-user weight are ranked first. Synthetic remote sites represent the actual Globalping probes shown in each source assessment.
Local history
Up to 8 assessment summaries are stored only in this browser. This lets you compare changes across time without a backend.
No history yet.
Routing trend & change detection
Tracks recent median RTT for a selected region using local history or the active project's saved assessments.
Run repeat assessments to establish a routing baseline.
Probe diversity
Remote recommendations are stronger when independently confirmed across different networks and ASNs.
—
Browser mode does not use a distributed probe set.
\n
Client report
Generates a dedicated, client-facing A4 report window with executive summary, methodology, recommendation matrix, map, pricing, risks and an optional engineer appendix. Use the browser print dialog to save it directly as PDF.
Azure Region PulseClient Confidential
Platform health & hardening
Checks the same-origin proxies and browser performance path. Proxy endpoints are destination-pinned, rate-limited and response-size bounded in this build.
Pricing proxyNot checked
GeocoderNot checked
GlobalpingNot checked
Azure readiness proxyNot checked
Page security headersNot checked
Region catalogue cacheNot checked
Heatmap workerInitialising
Region catalogue caching and Web Worker heat interpolation reduce repeat network calls and main-thread rendering work.
Throughput validation
Latency probes use tiny files and cannot honestly measure bandwidth. For a real Azure throughput check,
use a sequential 100 MiB Azure Blob download. Parallel region tests would divide your connection and distort results.
Run this after the latency assessment against your top 1–3 shortlisted regions.
Source location / privacy / assessment notes
Use a place as the source location
Quick uses up to 3 probes; Deep uses up to 5. The same selected probe set is reused across Azure regions on a best-effort basis.
This runs real HTTP measurements from Globalping probes near the selected place. The exact probes, networks and distance from the requested place are shown after selection.
Probe agreement will appear after a remote assessment.
or enter exact coordinates
Privacy: browser geolocation is used only in this page to calculate geographic distance. It is not uploaded by the tool. Place-name searches send only the text you enter to the same-origin geocode proxy, which queries OpenStreetMap Nominatim and returns matching coordinates. If you choose remote probes, the selected city/country context is sent by the same-origin Globalping proxy to Globalping so it can select live measurement probes. Browser-mode latency still originates from this browser/network; remote-mode latency originates from the actual Globalping probes shown in the result.
Browser latency requests go to public Azure Storage probe endpoints. The RTT probes use opaque no-CORS HEAD requests so the measurement does not depend on the website's Origin being whitelisted by each Azure Storage account. Each current AzureSpeed storage endpoint is validated first with a HEAD request and a 2-second timeout. Timed-out probes are dropped immediately, and a responsive probe is dropped if it later fails twice consecutively. IP/ISP information is fetched only when you press the network button.
History uses browser localStorage. Export, copy, share-link and print actions happen only when you request them.
AVD: Microsoft recommends client-to-host-pool-region RTT below 150 ms, and AVD Insights describes interaction below 100 ms as generally ideal.
This tool measures HTTPS RTT to Azure Storage in the region, so it is a regional proxy rather than an end-to-end AVD session-host test.
RDP Shortpath/UDP and actual session-host performance still require validation in the deployed AVD environment. For AVD ranking, the lowest healthy median RTT wins. p95 and stability are used only when calculated medians are exactly equal; the composite network-quality score is informational and cannot make a slower-median AVD region win. Stability uses median-centred robust variation so an isolated slow browser/network sample does not collapse the stability score.
Remote probes: remote mode uses Globalping's live distributed network. The requested place and the actual measurement source are deliberately shown separately because a probe may be in a nearby city or on a different ISP. Quick requests up to three probes and Deep up to five, preferring eyeball-network probes when selected. The first successful measurement fixes a probe set which is then reselected for the remaining Azure regions on a best-effort basis. Each probe counts as a Globalping test and API limits apply; set GLOBALPING_TOKEN on the supplied proxy to use an authenticated Globalping account and higher limits.
Service availability and residency: region placement must also consider service availability, cost, compliance, data residency and resilience. The residency selector is a conservative placement filter, not a compliance certification; verify service-specific residency and regulatory requirements before deployment.
The workload selector changes scoring and guidance. Mandatory requirements in Architecture & service requirements can exclude regions using measured metadata; named Azure services remain explicitly marked for Microsoft product-by-region verification because product/model availability changes independently of latency.
Build: 2026-08-16 global v10.0 · application workspace / decision intelligence / project scenarios / monitoring / hardened runtime ·
Metadata snapshot verified against Microsoft Azure public region documentation on 14 Aug 2026 (availability zones, pair, geography and programmatic name).
AVD latency guidance: Microsoft Learn AVD prerequisites / Insights. Latency probe infrastructure: the current azurespeed.com Azure Storage convention (s8<regionId>.blob.core.windows.net/public/latency-test.json), using the same regional storage-account family as the reference latency test.
Remote synthetic measurement: Globalping distributed probe network via the same-origin Globalping proxy supplied with this build. Azure retail pricing: Microsoft unauthenticated 2023-01-01-preview Retail Prices API via the same-origin hardened pricing proxy, including eligible Savings Plan and Spot rates. Heat interpolation runs in a Web Worker where browser policy allows, with synchronous fallback. Same-origin PHP proxies are destination-pinned, rate-limited and response-size bounded. Throughput validation recommendation: sequential 100 MiB Azure Blob download methodology used by AzureSpeed.com.