Spike: Use retry middleware with reverse proxy #64
Conversation
| log.Printf("Proxying request to host %v: %v\n", host, id) | ||
| newReverseProxy(host).ServeHTTP(w, proxyRequest) | ||
| proxy := newReverseProxy(host) | ||
| NewRetryMiddleware(h.retriesOnErr, proxy).ServeHTTP(w, proxyRequest) |
There was a problem hiding this comment.
This is how the middleware is integrated with the reverse proxy
|
|
||
| // NewDiscardableResponseWriter creates a new instance of the response writer delegator. | ||
| // It takes a http.ResponseWriter and a function to determine by the status code if content should be written or discarded (for retry). | ||
| func NewDiscardableResponseWriter(writer http.ResponseWriter, isDiscardable func(status int) bool) XResponseWriter { |
There was a problem hiding this comment.
I think wrapping the http.ResponseWriter comes with thorns. You can lose functionality if you don't implement all of the interfaces that ResponseWriter can be: https://go.dev/play/p/3B0yYC8kTvf
This is one the reasons I think that the middleware pattern in Go typically leads to more code than necessary. In the case of Lingo where we have a single primary handler (the proxy) I think that the middleware approach leads to more a more verbose/complicated solution than leaning on helper functions.
There was a problem hiding this comment.
Bringing benefits being customizable discussion here: #63
alpe
changed the title
|
Closing this now in favour of #61 |
Same middleware as in #51 but set up for the reverse proxy