Intermittent SSM Connectivity Issue on Snap Start Lambda Restore Phase

Hello Everyone, We have a snapstart enabled lambda where in Restore phase using Crac hooks, we fetch the SSM using AWS SSM client GetParameter. The afterRestore hook was successfully fetching the SSM then the handler processes the request. We have performed a Resiliency testing to checking the availability by disconnecting the subnets (AZs) and adding them back to the lambda. After this activity, we are facing with an Intermittent SSM Connectivity Issue on Snap Start Lambda Restore Phase. In the Restore Phase, the lambda goes to fetch the ssm and exception is thrown. This is observed in an irregular interval and some executions are successful as well. Interestingly the successful execution which happened occasionally was from all the three AZs.

This was an working code base and lambda function which is QA tested and running successfully for months.