Benchmark report

Is esm.sh slow?

Module response times from 6 cities

Fast globally

esm.sh responds quickly from all 6 test locations. Tokyo was fastest at 21 ms; Mumbai took 47 ms.

Fastest: 21 ms; Tokyo
Slowest: 47 ms; Mumbai
Successful requests: 100%; 6 locations

Response time around the world

Tokyo
21 ms
San Francisco
27 ms
Singapore
30 ms
Amsterdam
39 ms
Montreal
45 ms
Mumbai
47 ms

Why is Mumbai sometimes slow?

On eight of the 14 days Mumbai had the module in 19 to 55 ms. On the other six (15, 17, 20, 21, 22 and 25 September) its daily median was 223 to 253 ms, about ten times longer. The controlled run of 27 September caught both in one run: the median request took 23 ms and the slowest 5% about 233 ms. From outside we cannot see why; the pattern looks like a cache that sometimes answers in Mumbai and sometimes has to fetch the file from further away.

Click the location to see each stage.

ConnectTLSServer waitDownload

Mumbai

23 ms

Finding the server: 5 ms
Reaching the server: 3 ms
Setting up security: most of the time: 10 ms
Waiting for the server: 10 ms
Receiving the response: 0 ms
Total: 23 ms

Finding the server is measured once per location, so it sits outside these bars. Open a city to see it.

Bars show the typical request, so the parts add up to its total. Why

Compare all locations

Click a city to see its stage-by-stage breakdown.

ConnectTLSServer waitDownload

Amsterdam

38 ms

Finding the server: 9 ms
Reaching the server: 9 ms
Setting up security: 12 ms
Waiting for the server: most of the time: 16 ms
Receiving the response: 1 ms
Total: 38 ms

San Francisco

21 ms

Finding the server: 13 ms
Reaching the server: 4 ms
Setting up security: 7 ms
Waiting for the server: most of the time: 10 ms
Receiving the response: 0 ms
Total: 21 ms

Montreal

45 ms

Finding the server: 4 ms
Reaching the server: 11 ms
Setting up security: most of the time: 18 ms
Waiting for the server: 16 ms
Receiving the response: 0 ms
Total: 45 ms

Singapore

28 ms

Finding the server: 3 ms
Reaching the server: 2 ms
Setting up security: 11 ms
Waiting for the server: most of the time: 15 ms
Receiving the response: 0 ms
Total: 28 ms

Tokyo

20 ms

Finding the server: 5 ms
Reaching the server: 2 ms
Setting up security: 5 ms
Waiting for the server: most of the time: 12 ms
Receiving the response: 1 ms
Total: 20 ms

Mumbai

23 ms

Finding the server: 5 ms
Reaching the server: 3 ms
Setting up security: most of the time: 10 ms
Waiting for the server: 10 ms
Receiving the response: 0 ms
Total: 23 ms

The phases come from the controlled run of 27 September 2026. Typical and slower times pool the daily requests. 6 of 6 test locations included: Amsterdam, San Francisco, Montreal, Singapore, Tokyo, Mumbai.

Finding the server is measured once per location, so it sits outside these bars. Open a city to see it.

Bars show the typical request, so the parts add up to its total. Why

esm.sh still feels slow?

esm.sh has no status page: It runs behind Cloudflare, so a Cloudflare incident can show up here. Cloudflare's status page is the closest thing to one.
This is one resolve step: We fetch the preact 10.19.3 entry module, an 81-byte file that points to the build. The build itself, a full bundle, a package esm.sh has not built before, and npm install are not measured.
Montreal is answered from Newark: In the controlled run of 27 September Montreal's requests were answered by Cloudflare in Newark, while every other city had a nearby site. Montreal's daily median went from 17 to 23 ms to 44 to 54 ms on 20 September, which suggests its route changed that day.

Response time over the last 18 days

Daily median of 5 requests per location, 9 September 2026 to 27 September 2026.

A dot marks a day well above that location’s usual level.

Show the numbers

Day	Amsterdam	San Francisco	Montreal	Singapore	Tokyo	Mumbai
9 September 2026	46 ms	26 ms	20 ms	30 ms	32 ms	238 ms ▲
10 September 2026	38 ms	26 ms	21 ms	27 ms	33 ms	39 ms
11 September 2026	41 ms	32 ms	24 ms	49 ms	20 ms	26 ms
12 September 2026	44 ms	23 ms	20 ms	32 ms	25 ms	227 ms ▲
13 September 2026	41 ms	34 ms	23 ms	42 ms	31 ms	35 ms
14 September 2026	38 ms	28 ms	19 ms	31 ms	24 ms	25 ms
15 September 2026	39 ms	24 ms	21 ms	33 ms	26 ms	253 ms ▲
16 September 2026	37 ms	29 ms	23 ms	28 ms	20 ms	55 ms
17 September 2026	43 ms	24 ms	21 ms	46 ms	27 ms	223 ms ▲
18 September 2026	39 ms	30 ms	17 ms	30 ms	21 ms	26 ms
19 September 2026	39 ms	29 ms	19 ms	28 ms	21 ms	19 ms
20 September 2026	37 ms	27 ms	44 ms	27 ms	23 ms	233 ms ▲
21 September 2026	42 ms	26 ms	50 ms	29 ms	21 ms	226 ms ▲
22 September 2026	37 ms	27 ms	54 ms ▲	34 ms	19 ms	233 ms ▲
23 September 2026	42 ms	32 ms	49 ms	30 ms	22 ms	28 ms
24 September 2026	43 ms	27 ms	51 ms	29 ms	20 ms	21 ms
25 September 2026	39 ms	26 ms	47 ms	25 ms	21 ms	235 ms ▲
26 September 2026	35 ms	26 ms	53 ms ▲	33 ms	19 ms	24 ms
27 September 2026	41 ms	26 ms	48 ms	28 ms	18 ms	21 ms

About this measurement

GET esm.sh/preact@10.19.3 · 70 requests per location over 14 daily runs · 14 September 2026 to 27 September 2026

What we tested · Static file. The preact 10.19.3 entry module from esm.sh: an 81-byte file that re-exports the build, immutable behind Cloudflare. One resolve step of an npm import.

How LatencyRadar measures response time →

Technical details

Doesn’t measure: The module it points to, a full bundle, esm.sh's build step for a package it has not seen, or npm install.

One esm.sh surface: the entry module for preact 10.19.3, an 81-byte file that re-exports the build, marked immutable and served behind Cloudflare. A browser that imports a package from esm.sh asks for this file first and then follows it to the build, so this measures that first step of the import, not the whole thing.

Not measured: the build the entry module points to, full bundles, esm.sh building a package it has not seen before, or npm install.

Between 14 and 27 September 2026 the typical time was 21 ms in Tokyo, 27 ms in San Francisco, 30 ms in Singapore, 39 ms in Amsterdam, 45 ms in Montreal and 47 ms in Mumbai. Mumbai's typical hides two states, 19 to 55 ms on eight days and 223 to 253 ms on six, which is why its slower time is 259 ms, more than five times its typical. San Francisco also has a long tail for a file this small, 101 ms, from a handful of slow requests; its daily median never left 24 to 32 ms.

Test location	Typical response time	Slower response time	Requests
Amsterdam	39 ms	47 ms	20
San Francisco	27 ms	101 ms	20
Montreal	45 ms	55 ms	20
Singapore	30 ms	56 ms	20
Tokyo	21 ms	33 ms	20
Mumbai	47 ms	259 ms	20

Typical: half of the requests finished within this time (technical: p50). Slower: 95% of requests finished within this time (technical: p95). Statistics

Request: GET esm.sh/preact@10.19.3
Measured from: Amsterdam · San Francisco · Montreal · Singapore · Tokyo · Mumbai
Requests: Daily measurements from 14 September 2026 to 27 September 2026. 6 of 6 test locations included: Amsterdam, San Francisco, Montreal, Singapore, Tokyo, Mumbai. · 14 September 2026 to 27 September 2026
Timings taken: DNS, connect, TLS, waiting for the server, download
Report coverage: 6 of 6 test locations included: Amsterdam, San Francisco, Montreal, Singapore, Tokyo, Mumbai.
Daily requests in each location: Amsterdam: 70 · San Francisco: 70 · Montreal: 70 · Singapore: 70 · Tokyo: 70 · Mumbai: 70
Typical response time: Median of the eligible daily requests in each included location.
Slower response time: The 95th percentile of the same eligible daily requests in each included location.

esm.sh

https://esm.sh/preact@10.19.3

Independent measurement by LatencyRadar. Not affiliated with esm.sh. · How we measure (v1.2)

How fast does your site load around the world?

Run a free speed test from multiple cities and find out where your users are waiting. No setup, no account required.

Test my site

No account required · Takes about 30 seconds.

More benchmarks

All benchmarks →