Cold starts with DNS Cloudflare and runtipi traefik
Hello,
I've successfully finished runtipi guide on DNS cloudflare (https://runtipi.io/docs/guides/dns-challenge-cloudflare) so I will have a ssl https cerificates (with my cloudflare domain), edited settings config, traefik config, cloudflare records. But I have issues with cold starts of the epxosed app in runtipi. It happens always when the websites are opened after like 2 or more minutes. Something is happening in runtipi during this time. The loading takes very long to load the webpage. It does not happen in http localhost.
I am sending my configs and logs in the attachment. Thank you for your help.
14 Replies
so the config looks fine. in the logs i do not see anything (its on debug and there is lots of output not in that example - so i assume you showed all interesting things)
my next guess is its a DNS issue. it takes some time to resolve the name for each newly called app i suppose. easy test for that would be to have a host entry for one such an app and call it then to see if its faster.
Yes, I've tested it with the host entry and it loaded immediately everytime. But the issue is not only for newly called app, but also for already previously called apps. Which logs should I send? I didn't find any new logs appearing in runtipi during starting the app with the cold start.
Is there a possibility to solve this, please?
tested it with the host entry and it loaded immediately everytimeconfirms a dns issue for me tbh. so runtipi itself is not the issue here imo. best would be to lay out your setup in regards of using a tunnel , what dns service you use etc.
I use no tunnel, just DNS cloudflare from the runtipi guide. No proxy, just DNS there. That's all I think. Maybe something is set wrong?
i sssume you tested this at home? in your pc can you swap the probably domestic dns server for the cloudflare one 1.1.1.1 ?
Yes, I've changed it in DNS settings, but the issue persists.
and the hostentry you did where all went fast, points to the same ip as the cloudflare dns server is refering to?
so both time the external ip of your setup?
Actually, when I switched to 1.1.1.1, the issue disappeared. Although, again at first apps had to be cold started again. But now it works as it should. Thank you for your help. Seems like the 1.1.1.1 DNS needs to be set directly on router device. But this comes with another issue. If it will be added to the router, it means, that whenever the 1.1.1.1 will not work, the router will not provide the internet, right? Or wm I wrong?
Previously, I had the DNS router to be pointed to the adguard app in runtipi and whenever the adguard was updated, the whole internet at home was disconnected.
Previously, I had the DNS router to be pointed to the adguard app in runtipithat is what i meant by:
best would be to lay out your setup in regards of using a tunnel , what dns service you use etc.it probably has something to do with your adguard usage / setup. i cant help you here, since i dont use it.
No, I have no longer adguard deployed. Now I only changed DNS to 1.1.1.1
It was just an example π
ok. 1.1.1.1 is one of the most well known DNS server on the internet. whatever you had before is probably not more reliable.
also you can have more then 1 dns server specified. you could have 1.1.1.1 and 1.0.0.1 (both cloudflare) and then maybe a dns server of your provider configured
I see. Thank you for your help. I will try to add these to the router and see what happens.
So it didn't help. I set it on my laptop and a mobile phone, but still there are cold starts after some short time for most deployed apps.
So I have no idea what is the issue and how to fix it.
I just know that when I open the https app from runtipi, it makes cold start and these logs are displayed in runtipi proxy:
2025-09-13T19:46:34Z DBG github.com/traefik/traefik/v3/pkg/server/service/loadbalancer/wrr/wrr.go:175 > Service selected by WRR: http://172.18.0.8:1337
2025-09-13T19:46:34Z DBG github.com/traefik/traefik/v3/pkg/server/service/loadbalancer/wrr/wrr.go:175 > Service selected by WRR: http://172.18.0.8:1337
2025-09-13T19:46:34Z DBG github.com/traefik/traefik/v3/pkg/proxy/httputil/proxy.go:121 > 499 Client Closed Request error="context canceled"
2025-09-13T19:46:34Z DBG github.com/traefik/traefik/v3/pkg/server/service/loadbalancer/wrr/wrr.go:175 > Service selected by WRR: http://172.18.0.8:1337
2025-09-13T19:46:34Z DBG github.com/traefik/traefik/v3/pkg/server/service/loadbalancer/wrr/wrr.go:175 > Service selected by WRR: http://172.18.0.8:1337
2025-09-13T19:46:34Z DBG github.com/traefik/traefik/v3/pkg/server/service/loadbalancer/wrr/wrr.go:175 > Service selected by WRR: http://172.18.0.8:1337
2025-09-13T19:46:34Z DBG github.com/traefik/traefik/v3/pkg/proxy/httputil/proxy.go:121 > 499 Client Closed Request error="context canceled"
2025-09-13T19:46:34Z DBG github.com/traefik/traefik/v3/pkg/proxy/httputil/proxy.go:121 > 499 Client Closed Request error="context canceled"
yes those are all DBG messages not errors. so there is no helpful infos here.
in the end, given the infos you gave, its very hard to troubleshoot anything here tbh.
my guess is stll that that this is not tipi related , but something else in your homelab / network.
my suggestion would be to undo all your user configs and see if anything changed / is back for better.
from there on make adjustments to get to your prefered setup and test on all the steps.
also, since you are talking about "cold starts" - which is not the proper term to use here most likely - i would even suggest to use chatgpt for troubleshooting : you can prompt your setup there and provide logs - but be very careful regarding suggested changes - only if you have a complete backup of your data and are able to restore them, you should follow them - maybe better post those suggestions here first.
I see. Thanks. I would give you more logs for sure if you give me instructions on how to do that. I tried to restart runitpi/dockers/different domains from other providers, but it didn't help.
So I've deployed the same two apps with traefik in portainer, without runtipi, and the issue remains. So there is an issue with cloudflare imo. Not runtipi. Probably I cannot use them as DNS.
I've solved the issue. It was CF DNS. I changed it to Dynu DNS and it works flawlessly!