We continue to work to identify the root cause of the event resulting in connectivity issues for some Network Load Balancers within the US-EAST-1 Region. Some Network Load Balancers in USE1-AZ2, USE1-AZ4, and USE1-AZ5 are experiencing connection issues for TLS (SSL) requests. Non-TLS (SSL) requests are not affected by this issue. We are also experiencing some delays in the provisioning of new Network Load Balancers and registration of new targets behind existing Network Load Balancers. We have identified the underlying subsystems responsible for this issue and continue to work towards root cause and resolution.
We have identified root cause and are now working to test a fix for the issue affecting connectivity for Networking Load Balancers in the US-EAST-1 Region. Once we have confirmed that it resolves the issue, we will update all affected Network Load Balancers. The issue continues to affect TLS (SSL) connections for Network Load Balancers within the USE1-AZ2, USE1-AZ4, and USE1-AZ5 Availability Zones. The issue does not affect other Availability Zones in the Region and is not expected to affect any additional Network Load Balancers that are currently operating normally. Affected customers can create new Network Load Balancers (with or without TLS) in Availability Zones that are not affected by this issue for immediate mitigation.
We have confirmed that the fix to resolve the issue is working as expected and are currently deploying it to the affected Network Load Balancers. We will be deploying one Availability Zone at a time, in the following order: USE1-AZ4, USE1-AZ2, and USE1-AZ5. We expect to see recovery as we progress through the affected Availability Zones.
We have made steady progress and are close to completing the rollout of the fix in the USE1-AZ4 Availability Zone. Most of the affected Networking Load Balancers in that Availability Zone are now operating normally. We’ll continue at this pace through the remaining Network Load Balancers in USE1-AZ4, followed by the Network Load Balancers in USE1-AZ2 and USE1-AZ5 Availability Zones. Some customers have reported impact for PrivateLink endpoints, which is related to this issue and will see recovery as we continue to deploy the fix.
We continue to make progress in deploying the fix to Network Load Balancers in the USE1-AZ4 Availability Zone. While the vast majority of the affected Network Load Balancers in that Availability Zone have now recovered, we are making slower progress than expected on the final cell within the Availability Zone. We are working to resolve that issue before moving onto the remaining Availability Zones. We’ll continue to keep you updated on our progress.