Fargate timeout problem

Question

Hi. I've got an unexpected error, I guess it's about timeout. Task takes time quite long. generally less than 1minute. There is no problem when it finish before 1 minute. But sometimes task takes more than 1 minute. Error occurs when it takes more than 1minute. Please see below.

[ec2-user@ip-000-00-0-00 ~]$ curl --location --request POST 'userid.ap-northeast-2.elb.amazonaws.com:port/service' -d "video.mp4" -o output.json -v
Note: Unnecessary use of -X or --request, POST is already inferred.
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0*   Trying 172.31.74.35:5000...
* Connected to loadbalancer.ap-northeast-2.elb.amazonaws.com (000.00.00.00) port 0000 (#0)
> POST /service HTTP/1.1
> Host: userid.ap-northeast-2.elb.amazonaws.com:port
> User-Agent: curl/7.79.1
> Accept: */*
> Content-Length: 37
> Content-Type: application/x-www-form-urlencoded
>
} [37 bytes data]
100    37    0     0    0    37      0      0 --:--:--  0:00:59 --:--:--     0* Mark bundle as not supporting multiuse
< HTTP/1.1 500 Internal Server Error
< Date: Tue, 19 Apr 2022 01:45:23 GMT
< Content-Type: application/octet-stream
< Content-Length: 0
< Connection: keep-alive
< Server: Python/3.7 aiohttp/3.7.4.post0
<
100    37    0     0    0    37      0      0 --:--:--  0:01:00 --:--:--     0
* Connection #0 to host loadbalancer.ap-northeast-2.elb.amazonaws.com left intactm left intact

With curl verbose option, I get 500 Internal Server Error at 0:00:59. How can I finish my task which takes more than 1minutes? I've tried

increasing Health check grace period for ECS Service
increasing idle timeout of Load Balancer
increasing timeout and interval of Target group
curl options (like keep-alive, max-time) My EC2 Instance
type : t2.micro
Amazon Linux My Service
Service type : REPLICA
Launch type : FARGATE My task in service
Network mode : awsvpc
Compatibilities : EC2, FARGATE
Requries compatibilities : FARGATE
EFS mounted
Docker Appreciate,

Answer

I know this is an old question, but I was having a similar problem and thought I would provide my experience in case it would help anyone else (for the record, I think the answer of "Make your application faster" is not helpful at all). In my case, we are hosting a website with Fargate and fronting it with CloudFront origins. Long running APIs would timeout and return a 504 to the client due to the CloudFront request timeout setting on the origin. By default, the max you can set this timeout is 60 seconds, but you can do a service limit request and have it bumped up to 180 seconds. That is their current hard limit and they won't go higher than that. Once I did that and updated my CloudFront origin, my 504 errors went away.

Answer

Request / Response taking longer than a minute is already in very bad standing.

Make your application faster or redesign it to be batched / async?

Fargate timeout problem

相關內容