[Dump] DumpCollector can be slow with large amount of data

[barryvdh]

The DumpCollector works by continuously loading the Dump entries. There are 50 dump entries loaded on each request by default, which are refreshed every 2 seconds.
For the other collectors, there is a ?take=1 parameter which is compared to the last entry. Perhaps that should be done with Dump also?

And besides that, when dumping large object, Dump can become very heavy. Try generating 50 dumps of app() and loading the page (not very real-life, but demonstrates the problem -> generating 10+MB of HTML and refreshing that every 2 seconds.

Would the Dump screen perhaps make more sense as a normal preview page?

[barryvdh]

Or, if you want to create a 'streaming dump server' kind-of-thing, maybe create a regular index/preview page like the rest, but add a 'streaming' page (the current Dump screen) which will only load the last entry and add new entries on the fly, instead of loading alle of them from the API directly.