nighthawk
nighthawk copied to clipboard
L7 (HTTP/HTTPS/HTTP2/HTTP3) performance characterization tool
Consider a configured threshold as follows: ```bash --termination-predicate benchmark.http_2xx:99 ``` The way this is implemented, when the client stops, one would expect to see `benchmark.http_2xx` at `100` when the client...
This is a follow-up to comments on #338 which has added `echo_request_headers` to the test server filter `ResponseOptions`. I'd like to add more response options that I think might be...
If the test server could be asked to add response headers with high-res timing information about when it received a request, and when it started replying (or just the delta...
As an opt-in feature, we could add a ringbuffer per worker for holding more information regarding the requests that saw the top x highest latency replies. This information should be...
Issue If a customer wants to connect to multiple clusters via the same load balancer instance, it may be difficult to do so with the #current request source implementation. A...
How do i configure response payload in test server so that any request to test server would return status code, header along with given response payload Will appreciate for any...
In #265 we hard-code that statistics postfixed with "_size" are routed to `StreamingStatistic` instead of to `HdrHistogram`. It would be nicer if we could have options for that so that...
If a consumer needs to saturate a very large backend, perhaps one running on multiple machines, that consumer may want to start multiple nighthawks in order to adequately saturate it....
In addition to allowing the following dynamic configurations in RequestSpecifiers, it would be great if the RequestSource could override uri, so that it can explicitly dictate load balancer behavior, which...
For more complete control, have an alternative option for RequestSource where the remote source specifies the whole requests, either in an existing protobuf definition or as serialized json/yaml. This should...