It has been reported in multiple issues #24 and #26 that the deployment is broken. This PR attempts to fix that - we have tested with two options of user storage (persistent, key-value), and worker/distributor queues (dynamodb, sqs).
What is still missing:
[x] Tutorial explaining the meaning of configuration options.
[x] Make S3 buckets unique
[x] Reenable CI
[x] An SQS channel to transfer results back from Lambda functions to the user. Right now, we use TCP connections, which requires that the client application has a public IP and can accept connections.
It has been reported in multiple issues #24 and #26 that the deployment is broken. This PR attempts to fix that - we have tested with two options of user storage (
persistent
,key-value
), and worker/distributor queues (dynamodb
,sqs
).What is still missing: