You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When loading a dataset (warm cache), the duration until the first data bucket is downloaded is ~2.5 seconds. I think, there is a chance to optimize quite a few things:
make loading of olvy script completely async ("defer" is used but this still influences the DOMContentLoaded event)
load web worker modules lazily (e.g., they are not even needed in the dashboard)
reduce size of web workers (by reducing imports)
avoid impact of standard requests (organizationIsEmpty, features, getActiveUser, ToS)
the requests themselves could be faster (e.g., organizationIsEmpty waits 40 ms for the server so that a simple false is returned)
these requests could also be baked into the served HTML to avoid roundtrips (not sure about the downsides though)
the initial getMaximumZoomForAllResolutionsFromStore JS function takes 600ms. hopefully this can be sped up
GETing the user organizations can be made lazily (either for ToS or avatar in navbar)
Not all optimizations might reduce the perceived page loading speed because requests and JS code can happen in parallel. still, less work => better.
The text was updated successfully, but these errors were encountered:
I added some small request speedups in #8290 but the big fish in the network tab remains olvy (compare #8207).
Maybe some requests could also be parallelized (as they spend most of their frontend time waiting?) e.g. getActiveUser, buildInfo, toSAcceptance? Maybe some already are, I did not look into this further.
When loading a dataset (warm cache), the duration until the first data bucket is downloaded is ~2.5 seconds. I think, there is a chance to optimize quite a few things:
getMaximumZoomForAllResolutionsFromStore
JS function takes 600ms. hopefully this can be sped upNot all optimizations might reduce the perceived page loading speed because requests and JS code can happen in parallel. still, less work => better.
The text was updated successfully, but these errors were encountered: