Excellent advice Jesper!
On Sunday, 3 May 2020 10:34:12 UTC+1, Jesper Louis Andersen wrote:
>
> The general rule is to run a corollary of Dijkstra: experiments can only
> show the presence of problems in a system, not their absence :)
>
> That is, seek to validate your assumptions rather than try
The general rule is to run a corollary of Dijkstra: experiments can only
show the presence of problems in a system, not their absence :)
That is, seek to validate your assumptions rather than trying to guess what
is wrong with the system.
The obvious things to check against are HTTP/1.1 vs HTTP/2
Thanks Amnon & Hi Jesper,
You're correct. Most of the time is actually spent in *ioutil.ReadAll()* .
Trying to debug network issues related to GCP, GKE & the functionality of
ISTIO sidecar reverse proxy, but seems okay so far couldn't find anything.
But, I've performed an interesting experiment.
On Tue, Apr 28, 2020 at 6:48 PM Sarath Prabath Redlapalli Jaya <
harrysarath2...@gmail.com> wrote:
> We've instrument the above code block and reading request body bytes is
> taking very long i.e., upto 5 secs sometimes reaching upto 20 seconds at the
> throughput of ~4-5K RPM
>
>
> The Request
req := &mystruct{}
json.NewDecoder(r.Body).Decode(req)
We've instrument the above code block and reading request body bytes is
taking very long i.e., upto 5 secs sometimes reaching upto 20 seconds at
the throughput of ~4-5K RPM
The Request Body Size Metrics are as follows
Average: 73190 Byt
