Hotfix: More accurate performace profiling

Moved the hardware inspection code to its own thread. This will lead to more accurate results and better performance. Added a section to the documentation on cleanup.

Hotfix: Connection pooling

We discovered that the connection pool was configured to drop connection when the number of threads exceeded the number of connections in the pool, by swagger. We also discovered that swagger doesn't use a thread pool, which is very unsafe. Both of these are now implemented more elegantly.

Hotfix: TrakerrClient defaults

Two of the new fields set at the TrakerrClient level weren't getting populated in events properly. This is now fixed. Log accepted parameters which should be set at the client level, that was removed this patch. Also improved speed of sending an event by about a second.

Feature Release: REST API V1.3 release

This adds a bunch of new paramaters, mostly based around describing hardware. Some of these are automatically gathered (cpu usage, memory usage), many of them need to be given to us. This also updated the log function along side the v1.3 API update allowing users to send a simple event with a lot of flexiblity on extra data they send. A notable one that isn't related to hardware is the CorrelationID which you can use to correlate events. Check out the generated/docs/AppEvent.md doc for more info.

The logger was also slightly improved. The API change brought along the tags feature, and the logger now gets tagged by name when sending a name. Originally, the logger would only substitute it's name for eventtype when the event type wasn't defined (like when you are logging an event without a stacktrace). For events with stacktraces, the logger name wasn't readily available anywhere. Now the handler will tag each event with the logger's name, allowing you search all events by a logger, if you suspect that module is reporting something odd. Of course you can combine tags with other search terms to further make it easy to locate critical issues.

This is of course along side bugfixes and major doc changes. To see the new stuff, be sure to check out the README.md and the generated/docs/AppEvent.md!

Hotfix: Import Fixes

When the import chain on the library was being worked on, a few ambiguous names slipped through the cracks. This led to some confusing names between the package.module and the module relative itself. Because trakerr.trakerr existed, when initialized it looked for TrakerrHandler in the trakerr module and not in the trakerr package to find the TrakerrHandler class. Thing is, python 2 fell back to general relative imports, which python 3 doesn't support. This is now fixed!

Welcome to the fold python 3.x users!

Enchancement: Handler code.

Our handler code to hook into the logger needed to be tested and edge cases ironed out, along with some usability, doc, and efficiency updates. All of those are now fixed! Happy Logging!

Hotfix: LogLevel issues, and poor examples.

When I brought Trakerr to it's 2.0.0 release, it's examples hadn't updated with it. That's fixed, along with some less maintainable internal code that I designed poorly when I first wrote out the code for loglevels. The code should need to be maintained less as I abstract the allowed values for loglevel. Also, doc stuff.

Feature Release: Relative StackTrace Output

The StackTrace now output paths relative to the current working directory. This should make grouping smarter on the website. Other improvements is now a realization that the setup.py file wasn't being maintained properly, and various other comments and document updates.

Hotfix: create_app_event

Hotfix for creating an app event with a stacktrace without a try-except for the first time. Changed default values for create_app_event. See issue #10 and #11 for more details.

Trakerr Python initial release

Send your events to Trakerr easily from python with the official release of the Trakerr API for Python.