Nils Goroll
Nils Goroll
taking note of an issue which was almost forgotten: #3546 is just a special case of a broader issue: We should either keep object references until the end of the...
I am opening this RFC because the topic has been brought up by @rezan in general on IRC and in this PR in particular https://github.com/varnish/varnish-modules/pull/180 I am making an effort...
#3529 makes me want to come back to an issue which has bothered me for some time and now that the VDP API groundwork has been laid, a gzip vdp...
This tickets comes from a very true comment by @bsdphk https://github.com/varnishcache/varnish-cache/issues/3094#issuecomment-548274037 > Wasn't there something about CLI wandering the backends getting in the way of traffic when creation of dynamic...
This just popped into my mind working on a somehow related issue: Do we want to be agnostic against VCL changes to `Transfer-Encoding` and `Content-Length` and do the right thing...
@ernestojpg writes in #3312 > Don't you think that Varnish should provide some kind of counter/gauge to expose the current number of stablished connections to a specific backend? It is...
beresp 304 header merging makes impossible to distinguish between headers set by varnish and backend
Reported by @martin-uplex in https://varnish-cache.org/lists/pipermail/varnish-dev/2019-October/009469.html Basically, the way we merge headers for 304 responses makes it impossible to distinguish headers originating from the backend from ones which we set in...
I wonder why we do not issue [413 Payload Too Large](https://httpwg.org/specs/rfc7231.html#status.413) for _too many_ or _too long_ headers and [414 URI Too Long](https://httpwg.org/specs/rfc7231.html#status.414) for exactly that, but rather only send...
I noticed that, after a `kubectl delete pod` of one pod from a service configured for kube-keepalived-vip, client connections (https) would time out for >>10 minutes even after the pod...
for validating untrusted inputs, we need to limit the recursion depth in order to avoid running into stack overflows. We also introduce a new extensible API interface: `json_walk_args()` takes additional...