Degraded deploy speeds?

zaydek
PRO

a year ago

Has anyone noticed deploys are going out pretty slow? I'm deploying vanilla Go servers, no docker or special NIX dependencies required, but I'm seeing deploy times up to several minutes. Sometimes they go by quickly, <1 minute, and sometimes it takes several minutes per deploy. If an admin sees this, here is a relevant build:

https://railway.app/project/b508b3ae-a488-4863-a24f-3990f5b88fb3/service/78a4c936-e42e-48be-9966-f5ba72a9b7db?id=54363a84-2429-4ab7-843d-150971d5699c

It's the container step that's taking much longer that I'm used to.

8 Replies

zaydek
PRO

a year ago

Logs if it helps

1260769737974091876


a year ago

yeah deploy times are not the best right now, the team is working on speeding them up.

cc @angelo


zaydek
PRO

a year ago

Tried redeploying and immediately saw this

1260770137158582300


zaydek
PRO

a year ago

zaydek@Zaydeks-MacBook-Pro orderediting-go % railway up --service=headers-experimental --detach
  Indexed                                                                                              
  Compressed [====================] 100%                                                               
  Failed                                                                                               Failed to upload code with status code 500 Internal Server Error
zaydek@Zaydeks-MacBook-Pro orderediting-go % railway up --service=headers-experimental --detach
  Indexed                                                                                              
  Compressed [====================] 100%                                                               
  Failed                                                                                               Failed to upload code with status code 500 Internal Server Error

zaydek
PRO

a year ago

The original deploy took about 7 minutes in deploy time alone

1260770320906846500


zaydek
PRO

a year ago

Thanks


zaydek
PRO

a year ago

@Brody If I use a Dockerfile will it bypass the deploy container step or not necessarily?


zaydek
PRO

a year ago

I'm seeing 1-10 minute deploys


a year ago

container still has to be deployed regardless dockerfile, nixpacks, source image, etc


zaydek
PRO

a year ago

Kk, thanks


a year ago

Yea, the publish step is the same regardless


a year ago

Hope to see this improve soon. Love railway, but these recent deployment times are really killin it for me.


rojwilco
HOBBY

a year ago

experiencing the same issue here


rojwilco
HOBBY

a year ago

stuck on step Deploy > Create Container for 15 minutes, and then it timed out with activity StartToClose timeout


a year ago

the team is actively working on improving this


rojwilco
HOBBY

a year ago

can we get an announcement or something in the incidents channel when the fix is deployed?


a year ago

its possible that the team will post some kind of post mortem


lucasgarcia98
HOBBY

a year ago

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died

container event container died


lucasgarcia98
HOBBY

a year ago

?????


lucasgarcia98
HOBBY

a year ago

my application only dont work


a year ago

please open your own thread as this seems unrelated to the topic of this thread.



a year ago

thats a private thread, this topic does not need to be a private thread


a year ago

Going to investigate- in this case, I don't mind.


a year ago

Point being is that the team, and esp. myself know how frustrating this whole experience is right now.


lucasgarcia98
HOBBY

a year ago

as we are pro, and because of what it says it was supposed to be faster, I opened it this way, as we have been stopped several teams because of this problem


a year ago

i understand, but please know that it says that message no matter the option selected


lucasgarcia98
HOBBY

a year ago

ok


a year ago

Okay @here, raised a post-facto incident, https://discord.com/channels/713503345364697088/846875565357006878/1261026504574369894 and we expect to have a fix to this in 4 hours. Early reports are looking like a host config issue, we plan to share more on what happened.


marc
HOBBY

a year ago

Angelo, if it's not too much trouble, could you re-share that incident on this forum or in a spot with more visibility?


a year ago

You can check the incidents here -

https://status.railway.app/


marc
HOBBY

a year ago

@Brody / @angelo As of 9:18 PDT, the Railway status incidents page shows All Systems Operational. However, I'm still observing degraded deploy speeds. I've seen this across a few different apps including MySQL. Here are a couple of screenshots of a stuck MySQL deployment. Are you observing these issues on your end as well?


a year ago

Both those screenshots show a single active deployment, in this case the deployment event stage listing as not started is likely just a bug.


marc
HOBBY

a year ago

@Brody I agree it's a bug. I want to add that another observation: when I push a new deployment, I need to manually remove the previous one. That suggests there's still an issue. Is the team planning to investigate this issue?


a year ago

We are working on that one as we speak, does that deployment have a volume per chance?


marc
HOBBY

a year ago

Thanks.

Yes, it does. I had a hunch it was related to volumes and wrote up what I'm seeing here: https://help.railway.app/questions/issues-with-volumes-68850731


zaydek
PRO

a year ago

Is this still an issue? I tried a deploy just now and the deploy phase took 4m.


a year ago

yes this is still an issue


lucasgarcia98
HOBBY

a year ago

wtf

1262481790900109300


lucasgarcia98
HOBBY

a year ago

1262481830972756000


lucasgarcia98
HOBBY

a year ago

1262481968986329000


a year ago

you would want to optimize your image down to below 1gb ideally


lucasgarcia98
HOBBY

a year ago

locally, my folder builded, has 851 mb, idk what's using more than 2GB.


a year ago

thats not the same as the built image, please work on writing a dockerfile that results in a smaller image