You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've just become aware that awslimitchecker is beginning to fail on accounts in the Standard regions (not China or GovCloud) with the following unhandled exception:
ParamValidationError("Parameter validation failed:
Invalid type for parameter checkId, value: None, type: <class 'NoneType'>, valid types: <class 'str'>")
This seems to be related to the recent AWS announcement of new vCPU-based service limits for EC2 On-Demand Instances; at or around the time that accounts transition to the new vCPU limits (between October 24th and November 7th 2019, based on your account ID according to the table under the "How will the transition to vCPU limits happen?" heading on the EC2 FAQs page) the "Service Limits" Trusted Advisor check in the "performance" category disappears. awslimitchecker has been relying on this check for over four years (since June 21, 2015), and I cannot find any notice or public communication from AWS about this change. This also was not communicated to my employer, despite having Enterprise support.
I'm currently working on a release that will fix this problem, as well as a number of other bugs and support the new vCPU-based limits (#432). This release will also include support for Service Quotas (#413), which should drastically reduce or eliminate the need to manually set limit overrides for most services. Unfortunately, as I maintain awslimitchecker in my personal time, the release is not ready yet.
Install awslimitchecker from the develop branch of the git repository where the next release is staged. This will fix the issue, but introduces major changes (see the CHANGELOG for 8.0.0 for details) and also carries the caveats explained below. This can be installed from git via pip using a command such as pip install git+https://github.com/jantman/awslimitchecker.git@develop#egg=awslimitchecker
What's Holding Up The Release?
Previous versions of awslimitchecker rely largely on Trusted Advisor for finding the current service limits of services that don't expose them via API, which includes EC2. Most users who have limits increased above the defaults were relying on this data to get their current limit values, or else they would have to manually tell awslimitchecker what their limits were via limit overrides.
If the current develop branch were released as-is, with Trusted Advisor broken, many users would incorrectly see usage thresholds crossed, as limits would revert to their global default in the absence of Trusted Advisor information.
The fix for this is #413, switching from Trusted Advisor to the new Service Quotas service. That is a non-trivial change, and is going to take some time to implement, but is necessary for a reasonable user experience.
(This issue is a duplicate of #441 but was created for increased clarity.)
The text was updated successfully, but these errors were encountered:
I've just become aware that awslimitchecker is beginning to fail on accounts in the Standard regions (not China or GovCloud) with the following unhandled exception:
This seems to be related to the recent AWS announcement of new vCPU-based service limits for EC2 On-Demand Instances; at or around the time that accounts transition to the new vCPU limits (between October 24th and November 7th 2019, based on your account ID according to the table under the "How will the transition to vCPU limits happen?" heading on the EC2 FAQs page) the "Service Limits" Trusted Advisor check in the "performance" category disappears. awslimitchecker has been relying on this check for over four years (since June 21, 2015), and I cannot find any notice or public communication from AWS about this change. This also was not communicated to my employer, despite having Enterprise support.
I'm currently working on a release that will fix this problem, as well as a number of other bugs and support the new vCPU-based limits (#432). This release will also include support for Service Quotas (#413), which should drastically reduce or eliminate the need to manually set limit overrides for most services. Unfortunately, as I maintain awslimitchecker in my personal time, the release is not ready yet.
Temporary Workaround
I can offer two possible temporary workarounds:
pip install git+https://github.com/jantman/awslimitchecker.git@develop#egg=awslimitchecker
What's Holding Up The Release?
Previous versions of awslimitchecker rely largely on Trusted Advisor for finding the current service limits of services that don't expose them via API, which includes EC2. Most users who have limits increased above the defaults were relying on this data to get their current limit values, or else they would have to manually tell awslimitchecker what their limits were via limit overrides.
If the current
develop
branch were released as-is, with Trusted Advisor broken, many users would incorrectly see usage thresholds crossed, as limits would revert to their global default in the absence of Trusted Advisor information.The fix for this is #413, switching from Trusted Advisor to the new Service Quotas service. That is a non-trivial change, and is going to take some time to implement, but is necessary for a reasonable user experience.
(This issue is a duplicate of #441 but was created for increased clarity.)
The text was updated successfully, but these errors were encountered: