same here

Team is aware and looking into it

I noticed this in my mysql db's logs:

> 2026-02-11T14:55:06.361425Z 0 [System] [MY-013172] [Server] Received SHUTDOWN from user . Shutting down mysqld (Version: 9.4.0).

Is this mainly related to internal networking, otherwise I will start rerouting via public network?

It seems related to processes being terminated, team is looking into it, but no harm in trying.

Can you share your service URL? The one that appears in the URL bar (the one that starts with railway.com…) when you have the service in focus.

Seeing this too with my Caddy server. Keeps getting shutdown

This MongoDB indeed seems to have been shutdown: https://railway.com/project/e843ddc7-071d-430c-be13-52ab4fe102c0/service/d130e10f-dd87-45a2-aa5e-cc3ce0cd8798?groupId=6e585a96-f33a-482d-817b-417a7dd8eb42&environmentId=40e37c2c-937f-4965-89a2-76389262b957&id=7d81b9d3-3be6-40bd-9517-db468ec6a018#deploy

Same here, services are refusing to connect

I've tried redeployment, restart, different instances, nothing fixes it

true

Same here. Services crashing when accessing across multiple sites

if I restart my mysql db it seems to work for a minute or two, then it's shut down again

I'm getting massive response times suddenly, nothing seems to connect to my services either

same

all my Rabbitmq instance getting SIGTERM, shuting down

this is frustrating

same issue.

Even railway is having problems, like railway status

Restart did seem to help (for now)

And now dead again

Same Issue, my production MySQL is down. Affecting hundres of customers

Same for my services - restart helped briefly and then died again.

Same here, we restarted got brief recovery then fully down again

Same here

🚨｜incidents has been updated, we just have to be patient while the team fixes it

Same!

Any updatesM

?

None yet, and for updates be sure to look into the incident: https://status.railway.com/cmli5y9xt056zsdts5ngslbmp.

I have an uptime kuma app with sqlite db for monitoring and also this is going down. so it looks like it is related to volume mounting.

Don't believe so - hitting issues with services without volumes here.

Same here

For reference, the issue seems related to processes being killed randomly, not related to volumes.

Any ETA for this?

None at the moment, sorry.

we're having this issue too

Hope this gets solved quickly… My customers are beginning to notice 🙁

I keep having my services go down almost every month

Same here, customers are already complaining

Each time i stick with railway cuz i like the service but i can't fucking trust it if it keep going down every month

We're experiencing similar issues with our internal networking when connecting to both MongoDB and Postgres. All instances were running without any problems until approximately two hours ago, when we started encountering connection timeouts.

the status page is also incorrect, doesn't discuss the networking outage

Same here!

It's not related to internal networking.

the DB is not even reachable using public URL

Is your postgres server actually running though? In all of my services it doesn't seem to be networking related, it's that the services are receiving a shutdown command after a few minutes of being up.

now the services is just crashed randomly
for some reason, it become crashed, just crashed
without any logs

This is not acceptable. All our services are hard down right now.

The team will post updates on here: https://status.railway.com/cmli5y9xt056zsdts5ngslbmp. Really sorry for that.

It truly is, again really sorry for that

Is not just deployments tho, production went down completely

Yeah, the status update is deceptive. All our services are hard down, period.

This is insane, my app is completely offline

ya, the production is went down completely.

fair. i would say that builds failing doesn't fully describe the issue. some existing services went down as well

which is why i saw the status page and assumed it didn't impact us

I think everything's down, not just deployments

all our production services are down :/

Yeah this is not just deployments this is a major service outage effecting my customers

It is major outage, never been this bad before

even logs is not readable

yep, had this issue too

Lucky you this aint our first time ….

I think maybe post an incident in the incidents thread

can we at least get some updates? 🙁

my customers are getting anxious

they did

Can agree here, told the team if they can update it

There's an incident called: https://status.railway.com/cmli5y9xt056zsdts5ngslbmp. Updates will be posted there.

Literal seconds after I posted.

A new update has been posted, they're still looking into that.

since June last year, after migrate to the METAL stuff,
it start on and off having the outrage
average 1 month per incident

Same issue here

From the issues I've been observing, it seems related to the new influx of users to the platform. Been awhile since we got a major outage that affected running workloads (not builds).

yeah this one is different, it affects any services, even the already-running ones.

please fix it 🙁

A bit more than 1 month ago, all my services went offline, node js mongo etc etc

100% not just "deployments"

Internal routing is down as notionally my app is up but can't route to the DB. Nothing to do with deployments

Team is trying their hardest to fix it and I can agree that it's unacceptable, they'll be posting a post-mortem on this too.

I believe the services have been restored. Please 👍🏻 or 👎🏻 this message if it works for you or not

Things happen, as long as it gets fixed and there is an understanding on it and how to avoid it again, all is good - at least from my pov

It's not related to internal networking, the processes themselves seem to be killed randomly.

Appreciate the Railway team responding quickly!

What a joke. Responding quickly is not enough. It's one outage a month at this point. Not at all usable.

Bunch of apps fully down because of this outage.

Ah! correct, my db was showing online but clicking into it shows deployment is offline - my node service then can't route to it for obvious reasons!

Can you please update the status. "We're investigating issues impacting deployments" is very incorrect. This is a complete outage. It's deceptive to write it this way when it's clear to everyone the impact is far wider than deployments.

O problema aparentemente é apenas no Build com Metal, desabilitei o uso e meus sistemas voltaram a funcionar normalmente, apenas ao desabilitar o Use Metal Build.

been thinking is it worth it for migrate to the METAL thing. before this is just fine. lolz

I've passed this down to the team to update it, they're looking into it!

I hope their postmortem will be hyper detailed

same here

it has become a major outage

Railway's postmortems are normally pretty good! Certainly more detailed than many other providers cough AWS cough.

December 16 outage was exactly the same @ThallesComH 'minor outage' - when everything was down. This is systematic dishonesty.

if it helps anybody, my servers are on Legacy, not on Metal! And they're no longer dying every 5 mins.

What GPT had spotted in the logs was that 270GB of physical storage were assigned to the service, so I do believe that as all our projects grow, once in a while Railway will be running out of storage

Looks like my services may be back 🤞

u still in legacy? nice.. good for u

I remember Railway having an issue with elevated networking latency, but nothing as critical as this, but I can understand your frustation if that affected your workload, sorry for that.

I had to move away from Metal because the connection with Db was lagging a lot

Agree this is the worst one in the past 1 year

Previously, it was just build failures etc

damn, i still remember that,
the connection is suddenly go for 3s after migrate,
super lagging, damn scary

I have an email from December 16 saying the following, and it definitely wasn't just build failures, perhaps because it was European time but everything was down

"Hello,

We recently experienced a major outage that impacted your Railway workloads. We're very sorry for the disruption this caused.

Our goal at Railway is to provide a best-in-class experience, and this incident fell short of that standard. To help make things right, we've applied a credit based on your average bill from the last three months (20) to your workspace, Projects. You can view your available credits here.

For a detailed breakdown of what happened and the steps we're taking to prevent this in the future, you can read the full post-mortem here."

allow to download database backups, we need an anternative to get back to work

u can go back to legacy after the going METAL

The option is there

I'm trying it

For reference, we run our entire company on Railway and we weren't affected by that issue (to the point of our services going down). I apologize if I gave a different impression, it was not my intention to be dishonest.

really, it dont have that. sad

can this be changed at least? certainly not caused by the application, it's the fault in hosting provider. 🙁

Every workload on Railway runs on metal, there's no "non-metal" option.

this would provide better visibility for the end users

That's a great feedback, will check with the team!

Idt we even received credits even though all our workspaces were down then

u really no mercy on that.. haha

when there's a major outage like this, it would be great to at least be notified by railway so I can at least push messaging to my client notifying paying users about our service outage

my services are back now, i'm not using metal, but my apps werent connecting to my database, now they are able

yes please, a link pointing to the status page would be great. that will at least mitigate the damage for end users a bit, at least they know that the fault is happening on the hosting provider rather than the developers.

You can subscribe to incidents on and, in case you use Discord, you can follow the #🚨｜incidents channel.

We've got clients calling and are pretty annoyed that it's the second time in the past months. It's looking like we might have to change provider in good faith.

It took 10 mins for incidents to be updated with the issue

Even subscribing to it wouldn't do much, clients will probably call before that

I can understand your frustration. Our company also runs on Railway, and we're just as upset as you are about this situation. However, we also recognize that outages are inevitable in any service and we love Railway way too much to leave.

We genuinely don't care if some stuff on your timeline takes longer to get added, we just don't want to have a major outage each month

yes, I did that once I found out about the incident. but I only found out about the incident 30min after it started because I happened to open my site to test something

i love Railway very much
just dont outrage every month
i need to sleep

Most issues for us seems resolved now after restarts

i actually dont care whether u have update or not
just dont outrage

I can guarantee that the team will prioritize stability over features at any time. For reference, the build issues we were having was their top one priority.

Do you have a plan for resolving the problem? Will the data be accessible? Answers are needed; several critical systems are currently down. I need to know if the data is secure.

The benefit to Railway as a developer is amazing, to the client, they just don't want their website to go down and we provide client services so our opinions are less important. It makes it difficult to justify to client why to continue using an unstable service just because I personally like it.

One of my service still not working.

Is this on the Legacy services or Metal?\

Metal, I first restarted database services and then any depending service

a simple restart fixed it

i still cant see any logs

Can agree here, but I don't think we would have gotten our product to the point it is now without Railway. Every time we don't have to worry about infrastructure is time we get to ship. I can guarantee you that Railway is actively working on improving stability.

This worked for me now, too!

RESOLUTION turn this OFF.
DONE. 🫰

it work for a while (15 mins)
now 502 again
all service down again

I dont think the build environment has anything to do with it

it seems to be working again, now I have to deal with a lot of angry customers :|

I don't think that's guaranteed, since some folks (including myself) saw improvement turning it on. In general, I'd use the Metal build environment

database is still down

Have you restarted it?

2026-02-11T15:52:23.923886Z 1 [ERROR] [MY-012574] [InnoDB] Unable to lock ./ibdata1 error: 11
2026-02-11T15:52:24.081102Z 1 [ERROR] [MY-012574] [InnoDB] Unable to lock ./ibdata1 error: 11

Pretty sure that disabling metal build isn't doing much here, as your service received a redeploy which seems to be temporarily fixing the issues for some users. Please keep metal build enabled, it's much faster.

i restart , but the mysql doesnt start

I don't doubt that the team is working on fixing issues on Railway but the fact is that major outages are occurring every month

Restarting seems to have fixed it

ty

Railway can of course fail, like every service. But at least they hear out and actively fix our concerns. AWS would just put all of us on voicemail…

I can assure you they're aware. We have an internal chat with the team and receive updates on any issues they're actively fixing. One of the examples was the GitHub authentication problem where they told what was happening and the steps they're taking to fix it.

The main difference is AWS doesn't go down all the time

this is true actually

And when it does it's so bad that our clients are more accepting of it not being our fault

That's what I did in my database and applications, and all 6 had Metal enabled for testing. After disabling it, they all started working again right now. I'm telling you what I did and it worked. What you think based on guesswork without testing is just guesswork.

This can be re-enabled again after the problem is solved, and the problem is with Metal; it's on the resolution page!

Every workload runs on Railway Metal, which is Railway's own server infrastructure. Service builds were previously running on GCP, and now they're experimenting with running these builds on their own servers as well. Whether your server is built on Metal or GCP should not affect the outage.

Unfortunately we've to wait until the incident is fixed.

Excuse me, I don't understand. I don't have any Metal services; they're all legacy, and my MySQL database wont starts. Is anyone else having problems with the database?

still down in London FYI

restart

Have you run railway services on GCP? Metal's much more stable.

"Legacy" services are still running on Metal iirc, it's just the runtime that's legacy

What I did worked, and I shared it. I'm fine, do whatever you want. For me, it's already resolved. I don't care if it's just guesswork.

still down in Singapore, FYI

Yes, but only the applications on metal were affected. I disabled them until they fix it; it's a temporary solution. It doesn't matter what's best right now, what matters is what's working.

Hey, just got confirmation from the team that redeploying the affected services should fix the issue for some users (this explains why @Gandalf service is working, which is not related to Metal Builds). Just doing a redeployment on each service for example your API and database should fix the issue for some.

Not a great solution, I can agree, but the team is still working on a fix for everyone else

I did a simple "restart" of the already deployed SQL service and everything came back for me

Yep, restart also works

Still down for me

I never used Metal for my database and it's down as well 🙁

I did redebploy. two of my three services are fine. But one is still down

I redeployed however I am still running into 502's with my caddy proxy to my services

Hey, can you try redeploying each of your services as mentioned above? That should fix for some users.

This actually helped my database to go online

I cant redeploy the database, it would erase it… can only restart it I presume.

You don't have a volume attached to your db?

I did restart it

I am afraid to restart my database! what would happen with my data

If you've a volume attached to your service, that shouldn't happen. If you don't know what a volume is, it's this small "card" displayed below your service.

Hey, can you try redeploying each of your services as mentioned above? That should fix for some users.

for me the data wasn't corrupted or removed. I'm still checking the integrity, but based on what I see - data seems to be okay

this? @ThallesComH

Yep, you can redeploy it no problem.

Hey, can you try redeploying each of your services as mentioned above? That should fix for some users.

Ok let's go

My postgresql service can't be reached

Whoever has a database and the data in a single deployment should be able to SSH in to it right? Though no one should be doing that ever.

Redeployed didn't work but seems services are coming back after

Hey, can you try redeploying each of your services as mentioned above? That should fix for some users.

Maybe, but I don't see why you would need to SSH now though.

Just had three client websites crash and the automatic restart policies didn't work at all.

Just an idea but people who use Railway are people who are not comfortable with infrastructure or understand it completely. Railway creates this abstraction and makes it visually easy to understand. But this doesn't mean you can have completely no knowledge of infrastructure like Vercel, It's still involved.

https://svelte.dev/tutorial/svelte/welcome-to-svelte

Svelte has this really nice interactive tutorial… Imagine if Railway had this.

Oh they worked but all the restarts failed

For people still having issues after restart/redeploying, please open a #✋｜help thread and I'll escalate it to the team.

Confirming that redeploying my Postgres and app services brought them back online. Thanks @ThallesComH 👍

I don't think https://discord.com/channels/713503345364697088/1471157274314539098/1471171555592638698 is the resolution at all.

I'm back online

Can you please confirm we can safely redeploy a Postgres DB without affecting it?

Railway has an SSH option by using the CLI, just type railway ssh.

Can you share a screenshot of your service on the canvas?

Huh?

Is it fixed? Because i can't redeploy

Do not want to share the name of my services publicly

I just did and it's fine - the service mounts a persistent volume that survives the redeploy.

You can blur it out but shouldn't be a problem.

or temporarily renaming it to another name

redeploying the web app part

You can redeploy it, no problem.

Sorry the DB is mysql

sure? No risk?

Yep