On Mon, Jul 18, 2016 at 10:46 AM, Bakul Shah <ba...@bitblocks.com> wrote:
> A number of nodes and machines and networks are involved in > your test so you need to somehow further isolate the problem. > ...... On Mon, 18 Jul 2016 06:03:26 PDT Rayland <guianul...@gmail.com> wrote: > > I have an HTTP server in AWS on an instance with 4 cores and 8g of memory. > > The server is not processing anything, just replying with 500. Can you be specific about your HTTP server and how host names are resolved. If your AWS instance sender/receiver has a hostname that is resolved slowly you may have an answer. In some cases adding host names in /etc/hosts and resolving from "file" before "DNS" can speed things. I have seen X-server benchmarks suffer badly because "localhost" had to get resolved in another building. A local caching name server has value and is one of the standard bind9 configurations. Is a local caching DNS server a service option for an AWS instance? Latency in seconds may involve lost packets and retry timers. Since you have 4 cores... try connect inside a single box Also try connecting by IP address. Check that the server will respond to the sender's IP address and need not lookup up a name. There are security issues in not looking up host names but it helps to know if the external service is the issue. Wireshark on a single core can capture enough packets to tell you if your HTTP server is looking up DNS services and how quickly it gets answers. Firewalls -- are there network firewall services on the local network. AWS internal security same local network may have more impact than external connections. 10,000 packets a second can look like a flood ping or DOS and trigger firewall filters. -- T o m M i t c h e l l -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
