zlacker

[parent] [thread] 3 comments
1. hobo_i+(OP)[view] [source] 2026-02-03 16:21:15
Seeing how it hasn't survived the HN hug of death... Not sure how you've built it but consider putting it behind a CDN or something and caching the responses, esp since you're trying to pull live data
replies(1): >>tinuvi+1C
2. tinuvi+1C[view] [source] 2026-02-03 18:46:51
>>hobo_i+(OP)
It was never down - it was being patched and redeployed based on feedback here.

Webapp is light enough to handle 10000 concurrent hits.

replies(1): >>decept+NF1
◧◩
3. decept+NF1[view] [source] [discussion] 2026-02-04 00:13:02
>>tinuvi+1C
I can't tell much about your infrastructure, but memcached would probably increase that by several orders of magnitude. The NGINX module is pretty simple: https://nginx.org/en/docs/http/ngx_http_memcached_module.htm...

Cloudflare with Cache-Control headers is even simpler if you're okay with adding Cloudflare as a dependency.

From an ASN lookup, it appears you're hosting on Oracle Cloud, so Cloudflare would also give you free data egress: https://blogs.oracle.com/cloud-infrastructure/why-cdn-client...

Their Always-On feature would also help if Oracle has an outage.

I like the general idea, very lightweight and more likely to remain accessible when an emergency is overloading the mobile networks.

replies(1): >>tinuvi+XA2
◧◩◪
4. tinuvi+XA2[view] [source] [discussion] 2026-02-04 08:14:34
>>decept+NF1
Thank you! I will implement this - the server code is in plain C with NGINX on top so I presume adding memached would be straightforward.
[go to top]