Page tree

The tables in this topic describe the latency metrics in the exported data.

Host time

Host time is the time server took to process the object or page (measured in milliseconds).

Host latency metrics

Export headers

Description

x-host-time-sum

The sum of all host times for the time period (In other words, the total time spent by the infrastructure processing these requests)

x-host-time-min

The lowest observed host processing time

x-host-time-max

The highest observed host processing time

x-host-time-avg

The average host processing time

x-host-time-80

The 80th percentile host latency. 20% of all requests took this long or longer for the host to process.

x-host-time-85

The 85th percentile host latency. 15% of all requests took this long or longer for the host to process.

x-host-time-90

The 90th percentile host latency. 10% of all requests took this long or longer for the host to process.

x-host-time-95

The 95th percentile host latency. 5% of all requests took this long or longer for the host to process.

x-host-time-99

The 99th percentile host latency. 1% of all requests took this long or longer for the host to process.

x-host-time-norm

The normalized host latency. This is the average host latency, with the fastest and slowest 5% of observed hits excluded.

x-host-time-mode-low-end

The lower bound of the frequency bucket containing the mode (most typical) of host time

x-host-time-mode-high-end

The upper bound of the frequency bucket containing the mode (most
typical) of host time

x-host-time-mode-count

The number of observed hits in the mode bucket

x-host-time-tolerating-count

The number of requests that were unsatisfied and only tolerating the host latency (In other words, it took longer than the host latency PCL)

x-host-time-frustrated-count

The number of requests that were frustrated by the host latency (In other words, it took longer than the host latency PCL)

Network time

Network time is the time network took to transmit the object or page (measured in milliseconds).

Network latency metrics

Export headers

Description

x-nw-time-sum

The sum of all network times for the time period (In other words, the total time spent by the network transmitting these requests)

x-nw-time-min

The lowest observed network time

x-nw-time-max

The highest observed network time

x-nw-time-avg

The average network processing time x-nw-time-80 The 80th percentile network latency. 20% of all requests took this long or longer for the network to transmit.

x-nw-time-85

The 85th percentile network latency. 15% of all requests took this long or longer for the network to transmit.

x-nw-time-90

The 90th percentile network latency. 10% of all requests took this long or longer for the network to transmit.

x-nw-time-95

The 95th percentile network latency. 5% of all requests took this long or longer for the network to transmit.

x-nw-time-99

The 99th percentile network latency. 1% of all requests took this long or longer for the network to transmit.

x-nw-time-norm

The normalized network latency. This is the average network latency, with the fastest and slowest 5% of observed hits excluded.

x-nw-time-mode-low-end

The lower bound of the frequency bucket containing the mode (most typical) of network time

x-nw-time-mode-high-end

The upper bound of the frequency bucket containing the mode (most typical) of network time

x-nw-time-mode-count

The number of observed hits in the mode bucket

x-nw-time-tolerating-count

The number of requests that were unsatisfied and only tolerating the network latency (In other words, it took longer than the network latency PCL)

x-nw-time-frustrated-count

The number of requests that were frustrated by the network latency (In other words, it took longer than the network latency PCL)

SSL time

SSL time is the time the web browser and the web server took to negotiate SSL encryption for the object or page (measured in milliseconds).

SSL latency metrics

Export headers

Description

x-ssl-time-sum

The sum of all SSL times for the time period (In other words, the total time spent by the infrastructure negotiating encryption for these requests)

x-ssl-time-min

The lowest observed SSL processing time

x-ssl-time-max

The highest observed SSL processing time

x-ssl-time-avg

The average SSL processing time x-ssl-time-80 The 80th percentile SSL latency. 20% of all requests took this long or longer for the SSL to negotiate SSL encryption.

x-ssl-time-85

The 85th percentile SSL latency. 15% of all requests took this long or longer for the ssl to negotiate SSL encryption.

x-ssl-time-90

The 90th percentile SSL latency. 10% of all requests took this long or longer for the ssl to negotiate SSL encryption.

x-ssl-time-95

The 95th percentile SSL latency. 5% of all requests took this long or longer for the ssl to negotiate SSL encryption.

x-ssl-time-99

The 99th percentile SSL latency. 1% of all requests took this long or longer for the ssl to negotiate SSL encryption.

x-ssl-time-norm

The normalized SSL latency. This is the average SSL latency, with the fastest and slowest 5% of observed hits excluded.

x-ssl-time-mode-low-end

The lower bound of the frequency bucket containing the mode (most typical) of SSL time

x-ssl-time-mode-high-end

The upper bound of the frequency bucket containing the mode (most typical) of SSL time

x-ssl-time-mode-count

The number of observed hits in the mode bucket

x-ssl-time-tolerating-count

The number of requests that were unsatisfied and only tolerating the SSL negotiation delay (In other words, it took longer than the SSL latency PCL)

x-ssl-time-frustrated-count

The number of requests that were frustrated by the SSL negotiation delay (In other words, it took longer than the SSL latency PCL)

End-to-end time

End-to-end (E2E) time is the end-to-end latency of the entire object or page, including HTTP redirects (measured in milliseconds). 

E2E latency metrics

Export headers

Description

x-e2e-time-sum

The sum of all end-to-end times for the time period (in other words, the total time spent by the network and infrastructure delivering these requests)

x-e2e-time-min

The lowest observed end-to-end delivery time

x-e2e-time-max

The highest observed end-to-end delivery time

x-e2e-time-avg

The average end-to-end delivery time

x-e2e-time-80

The 80th percentile end-to-end latency. 20% of all requests took this long or longer for the end-to-end to deliver.

x-e2e-time-85

The 85th percentile end-to-end latency. 15% of all requests took this long or longer for the end-to-end to deliver.

x-e2e-time-90

The 90th percentile end-to-end latency. 10% of all requests took this long or longer for the end-to-end to deliver.

x-e2e-time-95

The 95th percentile end-to-end latency. 5% of all requests took this long or longer for the end-to-end to deliver.

x-e2e-time-99

The 99th percentile end-to-end latency. 1% of all requests took this long or longer for the end-to-end to deliver.

x-e2e-time-norm

The normalized end-to-end latency. This is the average end-to-end latency, with the fastest and slowest 5% of observed hits excluded.

x-e2e-time-mode-low-end

The lower bound of the frequency bucket containing the mode (most typical) of end-to-end time

x-e2e-time-mode-high-end

The upper bound of the frequency bucket containing the mode (most typical) of end-to-end time

x-e2e-time-mode-count

The number of observed hits in the mode bucket

x-e2e-time-tolerating-count

The number of requests that were unsatisfied and only tolerating the end-to-end latency (In other words, it took longer than the end-to-end latency PCL)

x-e2e-time-frustrated-count

The number of requests that were frustrated by the end-to-end latency (In other words, it took longer than the end-to-end latency PCL)

Page-render time

Page-render time is the time to render all the content on a page, when all or some of the content comes a source other than your origin server (measured in milliseconds).

Page-render time metrics

Export headers

Description

x-page-render-time-sum

The sum of all page-render times for the time period

x-page-render-time-min

The lowest observed page-render time

x-page-render-time-max

The highest observed page-render time

x-page-render-time-avg

The average observed page-render time

x-page-render-time-80

The 80th percentile page-render latency. 20% of all requests took this long or longer for the page to render.

x-page-render-time-85

The 85th percentile page-render latency. 15% of all requests took this long or longer for the page to render.

x-page-render-time-90

The 90th percentile page-render latency. 10% of all requests took this long or longer for the page to render.

x-page-render-time-95

The 95th percentile page-render latency. 5% of all requests took this long or longer for the page to render.

x-page-render-time-99

The 99th percentile page-render latency. 1% of all requests took this long or longer for the page to render.

x-page-render-time-norm

The normalized page-render latency. This is the average page-render latency, with the fastest and slowest 5% of observed hits excluded.

x-page-render-time-mode-lowend

The lower boundary of the frequency bucket containing the mode (most typical) of page-render time

x-page-render-time-modehigh-end

The upper boundary of the frequency bucket containing the mode (most typical) of page-render time

x-page-render-time-modecount

The number of observed hits in the mode bucket

x-page-render-time-toleratingcount

The number of requests that were unsatisfied and only tolerating the page-render latency

x-page-render-time-frustratedcount

The number of requests that were frustrated by the page-render latency

Related topic

Watchpoint summary export API