add initial thoughts on /canary endpoint changes

TODO

@@ -0,0 +1,16 @@
+TODO:  Implement /canary endpoints
+
+GET /canary/:configName -- get "recent" runs for this config.  Augment with params to allow pagination later.
+    Future to allow selection of date range, and pagination.
+
+GET /canary/:configName/:id -- get results for this specific run (id == pipeline id returned from POST /canary/:config
+   This should have a basic structure to indicate if it's running or not, completed, etc.
+   { status: completed, result: <extracted from pipeline or storage result object> }
+
+GET /canary/:configName/:id/data -- get the data (the joined metric set) for this canary run
+   This is used to validate the data from other scripts, not generally by the UI due to size.
+
+GET /canary/:configName/:id/data/:metricName -- get just the data for the experiment+control for this one metric
+   The UI can use this to display graphs for a specific metric.  Format is to extract just the given metric by name.
+   Tags should also be handled, how?  Can we allow a GET with a body here?  Should there be an UUID that is in the
+   results output that can be used here?