A spike to investigate options before requiring to build elastic/apm-server#3912.
This spike is to do some analysis of the data we already capture from the user agent string, and how we can use this to determine the device type.
An ES issue covers this, but this spike is to look at how we can distinguish between:
- Phone (or Smart Phone - TBC)
- Tablet
- Desktop/Laptop
- Bot
- Other (everything else)
With a future enhancement to be able to categorise the Elastic (Synthetics) Agent separately as well (e.g. for overlaying Synthetic performance data with that from RUM).
This spike should consider if we need to configure a device type field in RUM data (and how it might get there, such as via an ingest pipeline), or being able to do this just at query time, by searching the user agent string.
It might be useful to start by querying the user agent strings we currently get, and how this can be modelled into different device types at query time, to see how possible and performant this might be.
A spike to investigate options before requiring to build elastic/apm-server#3912.
This spike is to do some analysis of the data we already capture from the user agent string, and how we can use this to determine the device type.
An ES issue covers this, but this spike is to look at how we can distinguish between:
With a future enhancement to be able to categorise the Elastic (Synthetics) Agent separately as well (e.g. for overlaying Synthetic performance data with that from RUM).
This spike should consider if we need to configure a device type field in RUM data (and how it might get there, such as via an ingest pipeline), or being able to do this just at query time, by searching the user agent string.
It might be useful to start by querying the user agent strings we currently get, and how this can be modelled into different device types at query time, to see how possible and performant this might be.