Home
...
FAQs
How do I compare gateway latency with backend latency?

How do I compare gateway latency with backend latency?

Find out whether the response time of your calls was spent in the gateway or in the backend using the duration fields in Sensedia Analytics.

NOTE
Three fields record the time of each call: sensedia.duration_millis (total time), sensedia.gateway_duration_millis (time spent by the gateway) and sensedia.target_duration_millis (time spent by the backend). Comparing the last two shows where the time went.

The duration fields

Field	What it records
`sensedia.duration_millis`	Total time of the call
`sensedia.gateway_duration_millis`	Time spent by the gateway
`sensedia.target_duration_millis`	Time spent by the backend

Values are in milliseconds and the three fields are available in both the Traces and the Calls indices. See the field mapping for the names used in previous versions.

Checking call by call

Step 1

In General trace, set the time range you want to investigate.

Step 2

In the field list on the sidebar, add the fields sensedia.gateway_duration_millis and sensedia.target_duration_millis to the search result.

Step 3

Compare both values on each row. The higher one shows which stage consumed the time of the call.

Following the trend over time

To see the trend instead of going call by call, create a visualization combining:

a Date histogram bucket on the date field of the index;
an Average metric on sensedia.gateway_duration_millis;
a second Average metric on sensedia.target_duration_millis.

Both curves then appear on the same chart and the gap between them becomes visible.

TIP
Prefer the Percentiles metric over Average when you want to keep isolated spikes from distorting the reading. See aggregations.

How happy are you with this page?

How does data rotation work in Analytics?

FAQs