0

We are using Boomernang JS for our RUM instrumentation. It works well when http.initiator == spa_hard. However, when measuring http.initiator == spa (ie. soft loads) we will often see values for t_done (page load times) being in the area of 1000000ms, which is definitely inaccurate.

On top of that, even though t_done should be rt.end - rt.t_start, we will see often very large values for t_done yet rt.end == rt.t_start - the start and end values are equivalent as shown in this example:

{
  "pid": "7dv9jmse",
  "rt.bstart": 1604083165272,
  "rt.end": 1604411098690,
  "rt.nstart": 1604083161590,
  "rt.obo": 0,
  "rt.sl": 0,
  "rt.ss": 1604083161590,
  "rt.start": "manual",
  "rt.tstart": 1604411098690,
  "rt.tt": 0,
  "sm": "i",
  "t_done": 429144,
  "t_page": 422485,
  "t_resp": 6659,
  "u": "RETRACTED",
  "vis.lh": 2374175,
  "vis.lv": 20978,
  "vis.st": "visible"
}

Has anyone had encountered issues with SPA timing on Boomerang in the past. We are using Angular.

silverdagger
  • 1,124
  • 13
  • 37

1 Answers1

0

There have been several bug fixes, and updates made to bow BOOMERANG handles SPA Soft requests over the last year. Please see the release notes to determine if the version of BOOMERANG you are using contains the latest revisions: https://developer.akamai.com/tools/boomerang/release-notes

In particular look at the following:

1.681.0 (September 11, 2019)
The start time of SPA Soft Navigations may vary by a few milliseconds compared to before History v2, as the plugin now always uses window.History change notifications instead of SPA framework events.
Several issues were fixed that could cause some SPA Soft Navigations never complete, or take a long time to complete.
1.700.0 (January 29, 2020)
SPA Soft Navigations that did not initiate any critical network requests used to report a very small navigation time of around 15-30ms, depending on how long it took to execute the monitoring code.
Those navigations will now report a time of exactly 1ms.
Will Smithee
  • 1
  • 1