Polyfill.io

Usage and performance

The public instance of the polyfill service is hosted by the Financial Times, with the generous support of Fastly, who provide CDN distribution. View Fastly network map

Traffic volume

This shows the number of requests we have served per day, over the last 180 days, measured by Fastly:

DateRequests
27 Jul 2016 00:0040
28 Jul 2016 00:0077
29 Jul 2016 00:00167
30 Jul 2016 00:0074
31 Jul 2016 00:0071
1 Aug 2016 00:00184
2 Aug 2016 00:00110
3 Aug 2016 00:00118
4 Aug 2016 00:00109
5 Aug 2016 00:0095
6 Aug 2016 00:0056
7 Aug 2016 00:00106
8 Aug 2016 00:00184
9 Aug 2016 00:00371
10 Aug 2016 00:00245
11 Aug 2016 00:00159
12 Aug 2016 00:00122
13 Aug 2016 00:0059
14 Aug 2016 00:0068
15 Aug 2016 00:00197
16 Aug 2016 00:00147
17 Aug 2016 00:0091
18 Aug 2016 00:00106
19 Aug 2016 00:001190
20 Aug 2016 00:0091
21 Aug 2016 00:0081
22 Aug 2016 00:00132
23 Aug 2016 00:0056
24 Aug 2016 00:0072
25 Aug 2016 00:0059
26 Aug 2016 00:00125
27 Aug 2016 00:0080
28 Aug 2016 00:0070
29 Aug 2016 00:0037
30 Aug 2016 00:00347
31 Aug 2016 00:0051
1 Sep 2016 00:0044
2 Sep 2016 00:00735
3 Sep 2016 00:0065
4 Sep 2016 00:0092
5 Sep 2016 00:00128
6 Sep 2016 00:0080
7 Sep 2016 00:00171
8 Sep 2016 00:00175
9 Sep 2016 00:0069
10 Sep 2016 00:0068
11 Sep 2016 00:0051
12 Sep 2016 00:0067
13 Sep 2016 00:0054
14 Sep 2016 00:00128
15 Sep 2016 00:00106
16 Sep 2016 00:00100
17 Sep 2016 00:0049
18 Sep 2016 00:0054
19 Sep 2016 00:0077
20 Sep 2016 00:0082
21 Sep 2016 00:0091
22 Sep 2016 00:0068
23 Sep 2016 00:00104
24 Sep 2016 00:0037
25 Sep 2016 00:0085
26 Sep 2016 00:00127
27 Sep 2016 00:0081
28 Sep 2016 00:0066
29 Sep 2016 00:00179
30 Sep 2016 00:0086
1 Oct 2016 00:0043
2 Oct 2016 00:0071
3 Oct 2016 00:00226
4 Oct 2016 00:00160
5 Oct 2016 00:00110
6 Oct 2016 00:001579
7 Oct 2016 00:0048
8 Oct 2016 00:0064
9 Oct 2016 00:0067
10 Oct 2016 00:00149
11 Oct 2016 00:00223
12 Oct 2016 00:0078
13 Oct 2016 00:0081
14 Oct 2016 00:00118
15 Oct 2016 00:0055
16 Oct 2016 00:0070
17 Oct 2016 00:0063
18 Oct 2016 00:00103
19 Oct 2016 00:00157
20 Oct 2016 00:00236
21 Oct 2016 00:0047
22 Oct 2016 00:0079
23 Oct 2016 00:00204
24 Oct 2016 00:0091
25 Oct 2016 00:0064
26 Oct 2016 00:0065
27 Oct 2016 00:0076
28 Oct 2016 00:0097
29 Oct 2016 00:0077
30 Oct 2016 00:0049
31 Oct 2016 00:0034
1 Nov 2016 00:0096
2 Nov 2016 00:0065
3 Nov 2016 00:0080
4 Nov 2016 00:00100
5 Nov 2016 00:0030
6 Nov 2016 00:0040
7 Nov 2016 00:0078
8 Nov 2016 00:0072
9 Nov 2016 00:0034
10 Nov 2016 00:0054889
11 Nov 2016 00:00679929
12 Nov 2016 00:00825820
13 Nov 2016 00:00814720
14 Nov 2016 00:001021753
15 Nov 2016 00:001102019
16 Nov 2016 00:001099720
17 Nov 2016 00:001123739
18 Nov 2016 00:0067517
19 Nov 2016 00:0035108
20 Nov 2016 00:0031363
21 Nov 2016 00:0028867
22 Nov 2016 00:0040263
23 Nov 2016 00:0032979
24 Nov 2016 00:0024253
25 Nov 2016 00:0022254
26 Nov 2016 00:0022030
27 Nov 2016 00:0018731
28 Nov 2016 00:00182565
29 Nov 2016 00:001140876
30 Nov 2016 00:001186124
1 Dec 2016 00:001122908
2 Dec 2016 00:001395852
3 Dec 2016 00:002909070
4 Dec 2016 00:002038248
5 Dec 2016 00:002301520
6 Dec 2016 00:001214877
7 Dec 2016 00:001095901
8 Dec 2016 00:001064559
9 Dec 2016 00:001046807
10 Dec 2016 00:00904317
11 Dec 2016 00:00940400
12 Dec 2016 00:001213397
13 Dec 2016 00:001743987
14 Dec 2016 00:001664367
15 Dec 2016 00:001175534
16 Dec 2016 00:00999240
17 Dec 2016 00:00832129
18 Dec 2016 00:00835799
19 Dec 2016 00:001053727
20 Dec 2016 00:001189654
21 Dec 2016 00:001077093
22 Dec 2016 00:001022188
23 Dec 2016 00:00872246
24 Dec 2016 00:00726560
25 Dec 2016 00:00620632
26 Dec 2016 00:00834653
27 Dec 2016 00:00989326
28 Dec 2016 00:001089302
29 Dec 2016 00:001028335
30 Dec 2016 00:00979850
31 Dec 2016 00:00818599
1 Jan 2017 00:00785550
2 Jan 2017 00:001001862
3 Jan 2017 00:001194243
4 Jan 2017 00:001162339
5 Jan 2017 00:001159689
6 Jan 2017 00:001239661
7 Jan 2017 00:00911146
8 Jan 2017 00:00922117
9 Jan 2017 00:001143238
10 Jan 2017 00:001171298
11 Jan 2017 00:001116548
12 Jan 2017 00:001075348
13 Jan 2017 00:001011575
14 Jan 2017 00:00864698
15 Jan 2017 00:00880394
16 Jan 2017 00:001129440
17 Jan 2017 00:001157111
18 Jan 2017 00:001139520
19 Jan 2017 00:001121942
20 Jan 2017 00:001052632
21 Jan 2017 00:00913862
22 Jan 2017 00:00928253

Response time

Mean global response time, ms, from Pingdom:

DateResponse time (ms)
16 Jan 2017 21:0099
16 Jan 2017 22:0089
16 Jan 2017 23:0095
17 Jan 2017 00:0089
17 Jan 2017 01:0088
17 Jan 2017 02:0093
17 Jan 2017 03:0088
17 Jan 2017 04:0088
17 Jan 2017 05:0095
17 Jan 2017 06:0096
17 Jan 2017 07:0083
17 Jan 2017 08:0088
17 Jan 2017 09:0098
17 Jan 2017 10:00106
17 Jan 2017 11:0092
17 Jan 2017 12:0094
17 Jan 2017 13:0092
17 Jan 2017 14:0091
17 Jan 2017 15:0093
17 Jan 2017 16:00111
17 Jan 2017 17:0096
17 Jan 2017 18:00106
17 Jan 2017 19:0091
17 Jan 2017 20:0087
17 Jan 2017 21:00101
17 Jan 2017 22:0095
17 Jan 2017 23:0088
18 Jan 2017 00:0093
18 Jan 2017 01:0091
18 Jan 2017 02:0085
18 Jan 2017 03:0094
18 Jan 2017 04:0097
18 Jan 2017 05:00140
18 Jan 2017 06:00109
18 Jan 2017 07:00115
18 Jan 2017 08:00118
18 Jan 2017 09:00104
18 Jan 2017 10:00141
18 Jan 2017 11:0099
18 Jan 2017 12:00212
18 Jan 2017 13:00158
18 Jan 2017 14:0090
18 Jan 2017 15:0087
18 Jan 2017 16:00112
18 Jan 2017 17:0098
18 Jan 2017 18:00104
18 Jan 2017 19:00104
18 Jan 2017 20:0092
18 Jan 2017 21:00106
18 Jan 2017 22:00124
18 Jan 2017 23:00105
19 Jan 2017 00:00117
19 Jan 2017 01:00103
19 Jan 2017 02:00104
19 Jan 2017 03:0097
19 Jan 2017 04:0098
19 Jan 2017 05:0095
19 Jan 2017 06:00113
19 Jan 2017 07:00120
19 Jan 2017 08:00116
19 Jan 2017 09:0093
19 Jan 2017 10:00115
19 Jan 2017 11:00140
19 Jan 2017 12:00115
19 Jan 2017 13:00114
19 Jan 2017 14:00108
19 Jan 2017 15:00115
19 Jan 2017 16:00119
19 Jan 2017 17:00141
19 Jan 2017 18:00103
19 Jan 2017 19:00122
19 Jan 2017 20:00110
19 Jan 2017 21:00107
19 Jan 2017 22:00103
19 Jan 2017 23:0094
20 Jan 2017 00:00111
20 Jan 2017 01:00120
20 Jan 2017 02:0089
20 Jan 2017 03:0099
20 Jan 2017 04:00108
20 Jan 2017 05:00103
20 Jan 2017 06:00132
20 Jan 2017 07:0098
20 Jan 2017 08:00108
20 Jan 2017 09:00101
20 Jan 2017 10:00109
20 Jan 2017 11:00114
20 Jan 2017 12:00115
20 Jan 2017 13:00135
20 Jan 2017 14:00112
20 Jan 2017 15:00120
20 Jan 2017 16:00136
20 Jan 2017 17:00117
20 Jan 2017 18:00118
20 Jan 2017 19:00110
20 Jan 2017 20:00106
20 Jan 2017 21:00305
20 Jan 2017 22:00104
20 Jan 2017 23:00104
21 Jan 2017 00:0099
21 Jan 2017 01:00108
21 Jan 2017 02:00108
21 Jan 2017 03:00102
21 Jan 2017 04:0093
21 Jan 2017 05:00104
21 Jan 2017 06:00102
21 Jan 2017 07:00106
21 Jan 2017 08:00117
21 Jan 2017 09:00109
21 Jan 2017 10:00110
21 Jan 2017 11:00105
21 Jan 2017 12:00100
21 Jan 2017 13:00112
21 Jan 2017 14:00102
21 Jan 2017 15:00104
21 Jan 2017 16:00129
21 Jan 2017 17:00135
21 Jan 2017 18:00108
21 Jan 2017 19:00104
21 Jan 2017 20:00130
21 Jan 2017 21:00113
21 Jan 2017 22:00114
21 Jan 2017 23:00105
22 Jan 2017 00:00131
22 Jan 2017 01:00128
22 Jan 2017 02:0097
22 Jan 2017 03:00102
22 Jan 2017 04:00112
22 Jan 2017 05:00124
22 Jan 2017 06:00106
22 Jan 2017 07:00120
22 Jan 2017 08:00111
22 Jan 2017 09:00105
22 Jan 2017 10:00114
22 Jan 2017 11:00109
22 Jan 2017 12:00111
22 Jan 2017 13:00105
22 Jan 2017 14:00119
22 Jan 2017 15:00112
22 Jan 2017 16:00107
22 Jan 2017 17:0097
22 Jan 2017 18:0098
22 Jan 2017 19:00108
22 Jan 2017 20:00122
22 Jan 2017 21:00131
22 Jan 2017 22:0098
22 Jan 2017 23:00118
23 Jan 2017 00:00113
23 Jan 2017 01:00129
23 Jan 2017 02:00129
23 Jan 2017 03:00140
23 Jan 2017 04:00133
23 Jan 2017 05:00141
23 Jan 2017 06:00122
23 Jan 2017 07:00111
23 Jan 2017 08:00120
23 Jan 2017 09:00125
23 Jan 2017 10:00126
23 Jan 2017 11:00118
23 Jan 2017 12:00127
23 Jan 2017 13:00118
23 Jan 2017 14:00117
23 Jan 2017 15:00148
23 Jan 2017 16:00145
23 Jan 2017 17:00155
23 Jan 2017 18:00180
23 Jan 2017 19:00180
23 Jan 2017 20:00177
23 Jan 2017 21:00174

Caching

Hit ratio, last 7 days, from Fastly:

Cache resultRequests
Hits68254535
Misses1003162

Availability

Total downtime, according to Pingdom:

Last 30 daysNone
Last 3 months4 minutes
Last 12 monthsan hour

Caching responses that vary by User-Agent is very hard to do with good cache performance. We use a custom Fastly VCL configuration that separates the UA normalisation from the polyfill bundle.

Network performance in detail

95th percentile and median resource timing metrics, broken down by Fastly edge point of presence, measured using the resource timing API from live polyfill service requests that opted in to anonymously report performance data. Last 30 days, showing only POPs that have served more than 10000 RUM-enabled requests.

PoP Sample Per connection phase (95th percentile and median) Overall
IAD 14,116
603ms
(73ms)
DFW 12,006
656ms
(81ms)

Only non-zero samples are counted, which in practice means there are far fewer DNS lookup and TCP connect datapoints, because if reusing a keep-alive connection, both will be zero. Timings for individual connection phases do not sum to the overall RTT because, for example, the request that exhibits the 95th percentile DNS time, will most likely not also be the 95th percentile on all other connection metrics.

Typically Polyfill.io is loaded at a point where the browser is very busy. To speed up loading, consider using preconnect or preload.