stack launch in AWS sub-account fails #26
Comments
|
see if the following post can help: https://forum.serverless.com/t/unreservedconcurrentexecution-below-its-minimum-value-of-100/10323 also read up on the Lambda concurrency model here: |
urbien
changed the title
|
@martinheidegger please close if resolved |
|
Using my original account and creating a sub-account: for some reason the sub-accounts are consistently limited to 50 concurrent lambda executions. Having requested a change for one account fixes the issue for a specific account but doesn't relieve the issues for other accounts. However with a new account that is different: In the new account, sub-accounts automatically received the default concurrency of 1000 concurrent executions which is sufficient. In short, the issue persists on my original account and likely for some other users. My theory is that there is a setting internally at AWS that determines the value for new sub-accounts which is not exposed to the user (at least I couldn't find it). My original account is pretty old, newer accounts probably don't 🤷♂️, so we could mark it as wont-fix?! One thing that could be done is to look into the AWS limitations and refuse to deploy to an account that doesn't fulfill the limitations?! |
|
I don't understand how "50 concurrent entries" can be limiting for your testing. You don't have 50 sub-accounts created, do you? |
|
I mistook the word. What I meant was "50 concurrent lambda executions" (I updated the previous comment) |
|
you should not have been able to hit the limit of 50 concurrent Lambda executions with this project, how did you manage to do that? |
|
It seems that the CloudFormation setup process runs these executions at startup. My guess is that it is part of the warmup process combined with the tests to make sure that the lambda functions are properly deployed. |
|
It is very hard to imagine 50 lambdas running simultaneously. Even if we
run 50 requests at stack launch, which I doubt very very much, they would
never be simultaneous. Each lambda ends in a couple of hundred
milliseconds. --
Gene Vayngrib, CEO & Co-founder Tradle, #1 in EU and Japan, awards in US,
UK, SE Asia
KYC, client life cycle and leasing on blockchain, AI, 3rd gen cloud
Signal & WhatsApp +1 201 925 8899
Typed on mobile, please forgive my spelling errors
|
|
I can not explain the initial error otherwise. As mentioned before: when I asked the AWS support team to increase the limit from 50 to 1000 it works consistently fine. The error occurs while deployment. I am open for other suggestions how this could happen. |
|
Myclouds in all the regions show peaks of 90+ concurrent executions and an average when nothing is happening of 40+. |
Today I tried launching MyCloud using the Deployment product an organizational Account on a fresh AWS organization. It automatically rolled-back due to following error:
Preceding log statements
The text was updated successfully, but these errors were encountered: