What you might be seeing is typical HTTP/2 behaviour
Under HTTP/1.1 the browser prioritises the requests - each TCP connection can only be used to one request/response at a time, so once all six connections are in the use the remaining requests queue until there’s unavailable connection
Under HTTP/2 only one connection is used to the main origin, and the responses are multiplexed over it - the browser can send multiple requests at a time (with a priority attached), and then the server fulfils then. This means some requests will queue on the server while higher priority requests are served and this is probably what you’re seeing in the long request bars.
If you want to compare the difference test in Chrome, and then create a comparison run with the --disable-http2 command line flag
HTTP/2 works best when the content comes over as few as connections as possible so self-hosting jquery etc should bring benefits