API calls longer than expected after netlify outage

This is about: pensive-goldwasser-f25fb0

Since netlify outage I started to notice that my api calls are taking longer than expected, I moved the site to vercel or to a vps (aws - using nginx / static hosting) and the api calls are 2x faster.

is something wrong with my settings?

Not sure which API calls you’re talking about - I’m not seeing any Netlify Functions on your site. Is the API hosted on Netlify?

Exactly, I know how weird it sounds like, but when performing api call using netlify hosted service it takes longer than when using other hosted services. API likes ‘login’, if I curl, same request, from my pc it works 2.5 faster then thru the browser and it is happening with more devices & locations.

Since it is just frontend, the request starts from my pc, after loading the page.

I was just out of ideas about the root cause of it.

Also, you are correct, I’m not using anything other than hosting frontend with a very low traffic.

As you are using a third-party for API, and as you state above, this is not going through Netlify, but directly to them from your computer. You are not using Netlify functions for API, or proxying through Netlify to a backend. I am not sure how you think this is an issue with Netlify.

Completely unrelated…

I also see your site sistema.kitnetbar.com.br is proxying through Cloudflare. See [Support Guide] What problems could occur when using Cloudflare in front of Netlify?

I also see you have not disabled source maps (see GENERATE_SOURCEMAP.)

I also see the domains kitnetbar.com.br and www.kitnetbar.com.br are both inaccessible with an Origin is unreachable from Cloudflare. Is this intentional?

I appreciate your answers, and yes, everything is intentional BUT as I stated, this slowness behavior started after the outages Netlify had.

I opened this as a last option to understand what is happening. As I said, if I try from versel or a vps with nginx serving the static files it runs 2.5x faster

Example:

But honest I can’t blame the frontend for this behavior, so I was just looking for some enlightenments

Hi, @ragazzid. From the technical support side of things, I’m missing important information required to troubleshoot.

I use the analogy of me being a veterinarian and the issue being a sick horse. If I cannot see the sick horse, I cannot tell you why it is sick. In this case, the “sick horse” would be the slow API call. I need as much information about that API call as possible to troubleshoot.

Now, if the HTTP response is not being routed via Netlify, then you are chasing a red herring. Nothing at Netlify could impact a third-party API hosted elsewhere.

If, on the other hand, the HTTP response does route through Netlify, what we would need to identify the slow response is the x-nf-request-id HTTP response header for the slow response. We have a support guide about that header here:

If the header is not available we need the following details to troubleshoot:

  • the full URL requested
  • the client IP address making the HTTP request
  • the server IP address responding to the request
  • the date, time, and timezone of the request

With that information we can research the issue to see what the reason is for the slowness.

If there are questions about how to gather any of this information or why we require it, please let us know.

1 Like