-
Notifications
You must be signed in to change notification settings - Fork 2.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
AutoPublishAlias: Support updating version when things other than CodeUri Changes #413
Comments
Appears to be a bug in how we process |
Just wanted to bump this, it's been an issue for us as well. Our issue is when changing memory allocations. |
This is similar to #479 for APIs only deploying when the Swagger changes. We do have a proposal for APIs to always deploy even when we can't detect a change #660. However, I believe that Lambda won't allow you to create a new version of a Function if there have been no changes at all, making this solution not viable in this case. We could detect changes to the entire |
Any updates on this? Is there a solution to make a new lambda version created by environment variable change be tied to the alias? |
Also an issue for us; we use an environment variable to change log level and should to be able to update this without rebuilding the code. |
I don't understand the use for |
Any updates on this one? |
Adding |
Relaying from #2418:
|
Closing this, the fix will be released over the coming weeks |
Description:
Per the documentation, AutoPublishAlias only supports publishing another function version when the CodeUri changes. While having AutoPublishAlias in its current state is still beneficial, I have found this has caused much confusion and even bugs that took quite some time to track down. Specifically, when things such as environment variables are changed on the stack.
The "gotcha" that occurs is because an environment variable might be passed as a stack parameter (for instance). When doing a stack update and updating the parameter, everything appears to execute cleanly. The stack change set picks up that the function needs to be updated, etc and the update is marked as successful. This causes the false belief that your update got applied successfully, only on close inspection of the lambda function do you discover that the lambda function was updated, but a new version was not tied to the alias.
I am curious if there is plans to remove this limitation? While the documentation does spell this out clearly, unless someone updating the stack knows that function(s) in the stack are using the
AutoPublishAlias
feature and knows about this caveat - they they might think an update was successful with no other action needed.The text was updated successfully, but these errors were encountered: