Response time on pingdom is equivalent to the base page download time on WebPagetest. It’s only exposed in the raw data/xml interface though. The time to first byte is close but you also need to include the time to transfer the base page (so the time to first byte measurement is only a part of the response time).
As far as a good response time, “it depends” It mostly depends on where your servers are located relative to their test agents. If they are in the same region then under 100ms is a reasonably good time.
The google webmaster tools time is the same as the pingdom response time measurement (though it’s the average of all of the pages it crawled on your site, not just the pages that are being tested by pingdom). 891ms is pretty high (mine is running ~90ms in webmaster tools).
Neither of them has much relation to the time it takes to load for real users (which is just about always significantly higher because of other resources on the page). The google labs performance information is the end-user performance.
For real data for users you can add the server processing time to your access logs which will tell you how long each request takes, use the google analytics support for end-user performance or use something like new relic which will give you both in a pretty dashboard.