Changed AddRoute and ListenAndServe methods to return errors instead of logging them internally and continuing. This allows callers to properly handle and propagate errors, improving error visibility and enabling better error management across the codebase. Updated all callers in fileserver, reverse_proxy, stream routes to handle these errors appropriately.
* cleanup code for URL type
* fix makefile for trace mode
* refactor, merge Entry, RawEntry and Route into one.
* Implement fileserver.
* refactor: rename HTTPRoute to ReverseProxyRoute to avoid confusion
* refactor: move metrics logger to middleware package
- fix prometheus metrics for load balanced routes
- route will now fail when health monitor fail to start
* fix extra output of ls-* commands by defer initializaing stuff, speed up start time
* add test for path traversal attack, small fix on FileServer.Start method
* rename rule.on.bypass to pass
* refactor and fixed map-to-map deserialization
* updated route loading logic
* schemas: add "add_prefix" option to modify_request middleware
* updated route JSONMarshalling
---------
Co-authored-by: yusing <yusing@6uo.me>
- Incorrect name being shown on dashboard "Proxies page"
- Apps being shown when homepage.show is false
- Load balanced routes are shown on homepage instead of the load balancer
- Route with idlewatcher will now be removed on container destroy
- Idlewatcher panic
- Performance improvement
- Idlewatcher infinitely loading
- Reload stucked / not working properly
- Streams stuck on shutdown / reload
- etc...
Added:
- support idlewatcher for loadbalanced routes
- partial implementation for stream type idlewatcher
Issues:
- graceful shutdown
