Website Performance Testing Proxies: Measure Your Site From Real User Locations

Quick answer

To test how your own site performs for real visitors, you need vantage points that look like real visitors. Residential proxies give you consumer IPs in 195+ countries, so your CDN, DNS and geo-routing respond exactly as they would for a user in that city. Add ISP proxies when you need a fixed IP for a repeatable baseline. Residential is the recommended default.

Why website performance testing breaks without the right proxies

Synthetic monitoring tools run from a handful of cloud regions. Your CDN sees a datacenter ASN, picks an edge that serves cloud traffic well, and returns a time-to-first-byte that no real customer experiences. The same applies to DNS-based geo-routing: a resolver in a cloud region does not map to a consumer ISP in Jakarta or São Paulo.

Some stacks treat datacenter traffic differently. WAFs add JavaScript challenges or serve a lighter page to suspected automation, so a test from a cloud IP measures the challenge, not the site. Rate limiting can return HTTP 429 to a monitor firing every minute from one address.

The result is a green dashboard while a share of your users wait longer on an edge or A/B bucket you never measured.

Which Geonode proxy type fits

For measuring what real users see, route your checks through residential proxies. Geonode's pool has 2.5M IPs online across 195+ countries, with targeting by country, state, city and ASN. ASN targeting matters here: you can test from a specific consumer ISP and see which CDN edge its peering actually lands on. Rotating sessions give a different exit each run, which is what you want for a distribution of results; sticky sessions up to 24 hours hold one exit when you are debugging a single path.

For a stable baseline, ISP proxies give you static IPs from real ISPs with unlimited bandwidth per IP, from $1.25/IP. Run the same check from the same address every hour and any change in timing is on your side, not the network's. Datacenter proxies reproduce the cloud-region problem, so keep them for cheap availability pings only.

Residential — Best for: Real-user vantage points, CDN edge and geo-routing checks, per-ISP timing; IP source: Real consumer devices, 195+ countries, ASN targeting; Session: Rotating or sticky up to 24 h; Price from: $0.27/GB at volume
ISP (static residential) — Best for: Repeatable hourly baselines from a fixed consumer-grade IP; IP source: Static IPs from real ISPs; Session: Unlimited, static; Price from: $1.25/IP
Datacenter — Best for: Low-cost availability pings on non-protected endpoints; IP source: Rotating datacenter ranges; Session: Rotating; Price from: $0.14/GB

Recommendation: residential for the measurement itself, ISP for the control group.

How to set it up

Choose the markets that matter from your analytics, not from a world map. Pick the top countries by revenue plus any region where support tickets mention slowness.
Create one residential username per market with -country-xx, and add an ASN target (syntax in the docs) for the dominant consumer ISP. Leave the session suffix off so each run uses a fresh exit.
Fetch the full page, not just the HTML document: record DNS, connect, TLS, TTFB and total time, and log which CDN edge served the response from the response headers.
Add one ISP proxy per core market as a fixed control. Run it on the same schedule and alert when the residential median drifts away from the control.

# verify endpoint against docs.geonode.com before publishing
# Measure TTFB and total time from a residential exit in Germany
PROXY="http://USERNAME-country-de:PASSWORD@premium-residential.geonode.com:9000"

curl -x "$PROXY" -s -o /dev/null \
  -w "dns=%{time_namelookup} connect=%{time_connect} tls=%{time_appconnect} ttfb=%{time_starttransfer} total=%{time_total} code=%{http_code}\n" \
  -D - https://www.example.com/ | grep -iE "^(x-cache|cf-ray|x-served-by|via):"

Get the exact snippet for your language from the API Code Generator in the dashboard.

Website performance testing-specific settings that matter

Rotating for distributions, sticky for debugging

Use rotating sessions for scheduled checks so each sample comes from a different consumer exit and you get a realistic spread. When a specific route misbehaves, switch to a sticky session, hold the same IP, and run repeated requests to isolate whether the edge or the origin is slow.

Target the ISP, not just the country

A country median hides the ISP that peers badly with your CDN. Use ASN targeting to measure from the two or three largest consumer providers in each market separately. If one of them is consistently slower, the fix is a CDN or peering conversation, not a code change.

Keep concurrency low and schedules staggered

Performance checks should not become load tests. One request per market per interval is enough; stagger schedules so your own monitoring does not spike the origin. Residential proxies have no fixed thread limit, but the point is accuracy, not throughput.

Retry rarely, and record the failure

A failed check is data. Retry once after a short pause, then record the failure with status code, IP country and ASN. Repeated retries hide intermittent edge errors that your users are also seeing.

Log headers that identify the edge

Capture CDN headers that name the serving node and cache status alongside timings. Without them, a slow sample from Spain is a mystery; with them, it is a specific edge that missed cache.

What Geonode gives you for website performance testing

Residential exits in 195+ countries with country, state, city and ASN targeting
2.5M IPs online, so scheduled checks keep getting fresh consumer vantage points
ISP proxies with static IPs and unlimited bandwidth per IP for fixed baselines
Sticky sessions up to 24 hours for step-by-step path debugging
HTTP, HTTPS, SOCKS5 and UDP, and 24/7 support if a route needs investigating

Pricing for website performance testing

Performance checks use little bandwidth, so the residential entry plan of 10 GB for $7.92 per month covers a large monitoring matrix. Pay-as-you-go pricing drops to $0.27/GB at volume, a 3-day trial costs $5, and there is no contract. ISP proxies for fixed baselines start at $1.25/IP with unlimited bandwidth, and datacenter proxies for cheap availability pings start at $0.14/GB. See the pricing page for the full tiers.

FAQ

Why use website performance testing proxies instead of a synthetic monitoring service?

Monitoring services run from cloud regions, so your CDN and DNS respond as they would for a datacenter visitor. Residential proxies measure from consumer ISPs in the exact markets your customers use, including the edge selection and routing those ISPs actually get.

Can I test CDN edge selection for a specific ISP?

Yes. Geonode residential proxies support ASN targeting, so you can run the same check from the largest consumer providers in a market and compare which edge each one reaches. That separates a slow ISP peering from a slow origin.

Do I need ISP proxies as well as residential?

Not always. Residential gives you the realistic distribution. ISP proxies add a fixed address that never changes, which is useful when you want hourly baselines where any drift is attributable to your site rather than to a different exit each run.

How much bandwidth do website performance testing proxies use?

A full page fetch is usually a few megabytes including assets. Multiply by markets and checks per day to size your plan; most monitoring matrices fit well inside the 10 GB entry tier, and the $5 trial is enough to validate the setup.

Is this allowed on sites I do not own?

Use these proxies to test sites you own or manage with authorisation, and respect each platform's terms and rate limits. Geonode's acceptable use policy prohibits fraud, spam and credential abuse. Load testing third-party sites without permission is outside those boundaries.

Can I combine performance checks with uptime monitoring?

Yes. The same request that records timing also records the status code, so one scheduled job gives you both. Keep the ISP control running at a fixed interval and treat residential samples as the user-facing view of the same uptime data.

Start

Pick your top five markets, create a residential username for each, and run your first comparison against a fixed ISP baseline this week. Get started with the $5 trial, or Contact sales for a monitoring setup across many regions.