I did some rough Performance comparison between the Connect edition and a Forge version of our macros.
Many Forge apps will have to at least one network trip to the backend to display anything useful. So I was looking into how long it takes to that: As soon as the iframe is loaded, how long does it take until there is a initial request done.
What I expected:
- Frame and the static resource is loaded
- The Forge Brige/GraphQL is invoked and the race is off to fetch the data
- The Forge Bridge call is completed and the macro finally can display something useful.
However, I get a ~1second pause until the Forbe Bridge finally gets going:
- Frame and the static resource is loaded
- ~1 seconds of ‘nothing-ness’
- The Forge Bridge/GraphQL is invoked and the race is off to fetch the data
- The Forge Bridge call is completed and the macro finally can display something useful.
That means the Forge version of the Macro already looses 1 second before it even can do a request to the backed.
My questions:
- Did anyone else notice this behavior?
- If so, is this a performance bug in Forge or something we do wrong?
Example timings.
- The iframe resource requested at 15.88s, taking 18ms.
- I navigated to the page, so there is navigating time for this late start =)
- The forge_ui_invokeExtesion requested at 16.87, nearly 1 second later.
