Abstract: At a rule processing unit of an evolving, self-organized machine learning-based resource management service, a rule of a first rule set is applied to a value of a first collected metric, resulting in the initiation of a first corrective action. A set of metadata indicating the metric value and the corrective action is transmitted to a repository, and is used as part of an input data set for a machine learning model trained to generate rule modification recommendations. In response to determining that the corrective actions did not meet a success criterion, an escalation message is transmitted to another rule processing unit.

Abstract: Non-blocking processing of federated transactions may be implemented for distributed data partitions. A transaction may be received that specifies keys at data nodes to lock in order to perform the transaction. Lock requests are generated and sent to the data nodes which identify sibling keys to be locked at other data nodes for the transaction. In response to receiving the lock requests, data nodes may send to lock queues indicating other lock requests for the keys at the data node. An evaluation of the lock queues based, at least in part, on an ordering of the lock requests in the lock queues may be performed to identify a particular transaction to commit. Once identified, a request to commit the identified transaction may be sent to the particular data nodes indicated by the sibling keys in a lock request for the identified transaction.

Abstract: A system for transmitting data over a network may include data processed according to a data-redundancy encoding technique such as erasure coding to be transmitted via an unreliable, connectionless transmission protocol, for example user datagram protocol (“UDP”). A transmitting manager may receive a request from an application to transmit data. The transmission manager may select encoding parameters and encode data to be transferred using a data-redundancy encoding technique. The transmission manager may initiate transmission of the encoded data and payload header data may via a network, for example via a network communication stream or protocol stack. A reconstruction manager associated with the receiving node may reconstruct the original data using the received erasure coding data and payload header data. In some embodiments, the receiving node may transmit telemetry data to the erasure coding algorithm.

Abstract: Non-blocking processing of federated transactions may be implemented for distributed data partitions. A transaction may be received that specifies keys at data nodes to lock in order to perform the transaction. Lock requests are generated and sent to the data nodes which identify sibling keys to be locked at other data nodes for the transaction. In response to receiving the lock requests, data nodes may send to lock queues indicating other lock requests for the keys at the data node. An evaluation of the lock queues based, at least in part, on an ordering of the lock requests in the lock queues may be performed to identify a particular transaction to commit. Once identified, a request to commit the identified transaction may be sent to the particular data nodes indicated by the sibling keys in a lock request for the identified transaction.

Abstract: Non-blocking processing of federated transactions may be implemented for distributed data partitions. A transaction may be received that specifies keys at data nodes to lock in order to perform the transaction. Lock requests are generated and sent to the data nodes which identify sibling keys to be locked at other data nodes for the transaction. In response to receiving the lock requests, data nodes may send to lock queues indicating other lock requests for the keys at the data node. An evaluation of the lock queues based, at least in part, on an ordering of the lock requests in the lock queues may be performed to identify a particular transaction to commit. Once identified, a request to commit the identified transaction may be sent to the particular data nodes indicated by the sibling keys in a lock request for the identified transaction.

Abstract: Techniques to generate a distributed data structure using monotonic clock times are described herein. A set of data events are received at a plurality of replica locations and, based on delivery times, creation times, and timeout times, the data events are stored within a convergent replicated data structure. The data events are then ordered within the convergent replicated data structure based on the corresponding timeout times being less than a time value obtained from a locally accessible clock associated with each replica locations. The distributed data structure is then generated from the ordered data events by recursively selecting data events from the replica locations.