19913d85-f662-4518-a9a1-7adc1793dfd4.png){kind=small}
Channels
dagster-de
dagster-cube
gigs-freelance
dagster-wandb
dagster-machine-learning
dagster-bigquery
tools
dagster-dask
dagster-snowflake
dagster-releases
data-platform-design
events
dagster-serverless
dagster-noteable
introductions
dagster-airflow
dagster-cloud
random
dagster-dbt
faq-read-me-before-posting
dagster-airbyte
dagster-spatial
dagster-support
announcements
github-discussions
豆瓣酱帮
dagster-feedback
dagstereo
dagster-ecs
dagster-kubernetes
jobs
dagster-showcase
Title
w
Will Gunadi
01/04/2022, 7:52 PMMy Dagit instance intermittently hangs up and seems to freeze the networking of the AWS EC2 server. Has anyone else notice this?
d
daniel
01/04/2022, 7:56 PMHi Will - what version of dagster/dagit are you using?
c
Chris Retford
01/04/2022, 7:57 PMalso what instance type? the free-tier tends to have terrible networking
w
Will Gunadi
01/04/2022, 8:01 PMdagit, version 0.13.3
@Chris Retford what do you mean by free tier? Isn't dagster open source?
@daniel dagit, version 0.13.3
c
Chris Retford
01/04/2022, 8:02 PMfree tier of AWS
w
Will Gunadi
01/04/2022, 8:02 PMah, this one is a t3-medium
not free
c
Chris Retford
01/04/2022, 8:03 PMyeah. still t-class tho
I would try with m or c class
see if you see the same behavior
w
Will Gunadi
01/04/2022, 8:03 PMlemme verify
@Chris Retford can you refer me to info where the network issues with t3 is mentioned? Other than bandwidth, i don't see anything that would indicate the m or c class EC2 networking to be superior
c
Chris Retford
01/04/2022, 8:08 PMjust experience. but if you look here: https://aws.amazon.com/ec2/instance-types/ you will see lots of asterisks around network bandwidth and cpu
w
Will Gunadi
01/04/2022, 8:08 PMThank you! I'll chew on those.
@daniel is there a way to specify more than one "worker threads" like what you can do with uwsgi with Dagit?
I kinda suspect that the lock up is caused by single thread trying to process more requests than it can handle.
d
daniel
01/04/2022, 9:40 PMWe're actually working on revamping the dagit server to use starlette instead of flask which would give you some more options here in terms of specifying workers, that's coming soon but isn't quite ready yet.
w
Will Gunadi
01/04/2022, 9:43 PM@daniel flask is fine if we can use something like uwsgi. I used those configurations for production server and they are quite robust.
d
daniel
01/04/2022, 9:49 PMI don't think we have hooks for that in the flask implementation currently, but I expect we will in the new starlette version
w
Will Gunadi
01/04/2022, 9:56 PMI just read on Starlette, looks like optimized for websocket stuff. Should be good for Dagit. Hopefully it can scale (worker-thread-wise)
k
Kyle Downey
01/21/2022, 6:24 PMHi @daniel, will this change also include an upgrade to a modern GraphQL version? We just hit an issue today where we could not use an Ariadne version later than 0.11.0 -- from April 2020 -- in combination with Dagster because you guys are still on graphql-core 2.x and they are on 3.x.
I think I have a way around it (fingers crossed it looks like Tartiflette does not conflict) but Graphene, Ariadne, Strawberry all conflict with Dagster due to this issue.
a
alex
01/21/2022, 6:27 PMyea thats the plan
k
Kyle Downey
01/22/2022, 12:50 AMThanks for confirming. Meanwhile Tartiflette seems to work in combination with Dagster.