Todays software is broken. Nobody can really see what is going on inside their own code. At least not by deafult. The default is darkness. Traysee aims to lift this darkness and change the default. All code should be traceable and visible. Logging only goes so far. Text comprehension is slow. A new approach is needed. Visualizing code execution as a gantt chart or flame chart is a way to go.
There are no silver bullets. Knowing what to trace and where is up to you. Traysee goes over and beoynd to make instrumentation as easy as possible but you need to put in the tracing code. Tracing instructions are ordinary HTTP requests. That makes it possible to start tracing from virtually anywhere right now.
HTTP request are slow and they will affect performance of your code. It is a trade-off. The good thing is that you don't need any special tracing library. The plan is to make asynchronous HTTP request without waiting for the response. That should bring tracing overhead down to microsecond level.
Step 1 - Name your dashboard
GET /api/p - create an event
namename of traced event (string)
starttime of event start - epoch time in seconds with any precision(number)
endtime of event end - epoch time in seconds with any precision(number)
usernameyour unique username (string)
GET /api/trace/obj/activity - view activity dashboard
namefilter events by name (string)
lastfilter only last N seconds (integer)
usernamefilter events by username (string)
polling_offturn off automatic UI refresh (true)