Abstract: Disclosed are various embodiments for simulating hardware-based faults in a cloud provider network. A first computing device can send a command to an offload card installed on a second computing device to introduce a simulated hardware fault into the second computing device. Then the first computing device can determine whether the second computing device has successfully recovered from the simulated hardware fault. Alternatively, an entry in an access control list (ACL) of a network switch can be modified to block network traffic to a first network interface of a host machine that is connected to the network switch. Then, a command can be sent to a second network interface of the host machine to instruct the host machine to perform a hard reset. Then it can be determined whether the host machine has successfully booted subsequent to the hard reset.

Abstract: Availability zone and region recovery are described. For an availability zone (AZ), a recovery availability zone (rAZ) may be identified based on available computing capacity of the recovery availability zone and geographic proximity of the availability zone relative to the recovery availability zone. In an instance in which the availability zone is impacted in which at least one of hardware and software of the availability zone is not fully operational, a virtual private cloud (VPC) is generated that establishes a peered connection between the availability zone and the recovery availability zone. A service is executed in the recovery availability zone, permitting any other services executing in the availability zone to invoke the service and become partially or fully operational.

Abstract: Network services are deployed in a networked environment in association with a user account. Dependencies of a network service, such as other network services, may be identified based on an online analysis and an offline analysis of the network service. Further, anomalies associated with the dependencies may be identified in some situations. A call graph may include nodes corresponding to the network services and its dependencies, and may include an identifier corresponding to a part of the call path that has the anomaly. An inspection of the call graph allows software developers to readily recognize that their service depends on a potentially flawed software that may cause a service failure or outage.