The index of Flutter dashboards is available at: https://flutter-dashboard.appspot.com/
There are two, one for benchmarks derived from running flutter/flutter tests, and one derived from flutter/engine tests.
Each data set (results from a benchmark value) consists of date/value pairs associated with a set of parameters. These parameters include things like the branch that was being tested, the benchmark test name (“test”), and the specific value (“sub_result”). There are other parameters but they aren't important (see below).
You can navigate the X axis of the graph using WASD.
The parameters for a data set are as follows:
sub_result: the specific data stream provided by the test. Tests can provide multiple data points, in a JSON map; each key in that map becomes a sub_result in the Skia Perf system.
branch: the branch that was being tested.
config: always “default”.
originId: always “devicelab” except for data that was migrated from the old database, which are labeled “legacy-flutter”.
test: the name of the test that ran to collect the data.
unit: in theory, the units in which the data was collected (so each sub_result should only ever be associated with one unit). In practice this parameter is often incorrect.