Hey, this was almost certainly on Railway's side there was an incident on their Edge Network causing elevated latency, with partial outages in US West (California) and US East (Virginia):

https://status.railway.com/incident/A97BS3FC

Since it's been resolved for a while now, trigger a fresh redeploy and see if things go back to normal. If it's still slow after that, it's likely unrelated to the incident worth checking your recent changes or opening a ticket with Railway support.

No, It is the South Asia region.

It is unrelated to that outage.

Ah try this:

Go to your service -> Deployments -> find the last deployment that was fast -> Redeploy it. If:

• Fast again -> the problem is in your new build, not Railway
• Still slow -> platform/infra side, open a ticket

Yes, it remains just as slow. As I mentioned earlier, I made zero changes to my code. The problem is clearly with your service.

Can i have the public url of your project ?

https://www.furima-alert.app

FWIW I tested your app from Europe: TTFB is ~240ms after connect, which is healthy for a route through Europe to a Southeast Asia origin. So the origin itself responds fine on this route the slowness is either specific to your network path, or to dynamic/API routes. Can you tell where exactly you see it: page loads in the browser, API responses, or server-side calls to a DB/external service?

When it's working properly, the speed is normal and fast. What can you possibly tell from just a single test? It's taking an unbelievable amount of time.

It seems it's sometimes taking over 2 seconds to load just a simple static page. This is absolutely ridiculous.

Details of the situation:

・I deployed the app around 3:00 AM today (JST), and the performance has been consistently poor ever since.

・The build took noticeably longer than usual, which already seemed abnormal.

・As I have stated multiple times, I have not made any changes to the code or versions.

・There is a clear, drastic drop in overall performance compared to before the deployment.

・Delays are occurring even on simple static pages, and there is massive latency on the API as well.

・Furthermore, regarding the API (App Router), since it only handles very lightweight processes, the response time used to be consistently under 1 second. However, it now sometimes takes more than 10 seconds.

・The API includes database operations, but I have confirmed there are zero delays on the DB side.

Based on all this, it is highly obvious that there is an issue on Railway's end. Please investigate this properly.

I can't pinpoint exactly where the bottleneck is, but it seems like the delay is happening at the routing level.

I am seeing similar delays in the past hour-ish, from US

As shown in the 1-week log in the attached image.png, there is a sudden drop in memory right after the deployment. Actually, there are two sudden drops in memory in the image (the most recent one being the one I am referring to), and both of these occurred immediately after a deployment. Notably, the older one (about a week ago) was also accompanied by a drastic performance drop immediately after that deployment, just like this time. Could this be related to the issue?

As shown in the 1-week log in the attached image.png, there is a sudden drop in memory right after the deployment. Actually, there are two sudden drops in memory in the image (the most recent one being the one I am referring to), and both of these occurred immediately after a deployment. Notably, the older one (about a week ago) was also accompanied by a drastic performance drop immediately after that deployment, just like this time. Could this be related to the issue?

As shown in the 1-week log in the attached image.png, there is a sudden drop in memory right after the deployment. Actually, there are two sudden drops in memory in the image (the most recent one being the one I am referring to), and both of these occurred immediately after a deployment. Notably, the older one (about a week ago) was also accompanied by a drastic performance drop immediately after that deployment, just like this time. Could this be related to the issue?

While I’m not certain if this memory anomaly is the root cause of the issue I'm reporting, I would appreciate it if you could take action to address it as soon as possible.