Microsoft Technology Licensing, LLC

Vamshidhar Radha Kommineni

Yi Meng

Gerald Francis Maffeo

William Hunter Hudson

Jose Bernabeu-Auban

Corneliu I Lupu

Michael H Krause

Stephan Doll

Paul Matthew Pietrek

Richard E Wilbur

Christopher Marsh

Frederick Justus Smith

Pavel Anatolievich Dournov

Anil A Ingle

Matthew Kerner

David Gregory Grant

Kenneth Kai-Baun Ma

Adam Liechty

Matthew M Kerner

An application programming interface (API) that leverages operating system instrumentation to provide a chain of threads and processes may alleviate some debugging complications. Specifically, the chain may start with the first thread in the process that experienced the original failure and end with the last thread upon which the first thread directly or indirectly depends. The API may aid debugging efforts by classifying all threads related or dependent upon an original failed thread into specific categories of failures, requesting further information from the originating OS concerning specific failed threads, and using that information to debug the failed application or process more thoroughly.

Failed process reporting

Methods, systems, and computer-readable media for monitoring information passed from instances of role(s) of a service application installed on a distributed computing platform and for indexing and analyzing the information within a data store are provided. Monitoring involves a monitoring agent that is integrated within an operating system of computing devices that make up the distributed computing platform. In operation, the monitoring agent retrieves information from the role instances and converts the information into parameters that are assembled to documents. Generally, the parameters are assembled to a uniform-structured format in accordance with a schema imposed by the distributed computing platform. Accordingly, each of the documents across the platform have a common format that promotes processing and analysis without a need for reconfiguration. Accordingly, the health of the service application, although dispersed throughout the distributed computing environment, may be observed and evaluated in a consistent manner.

Monitoring agent programs in a distributed computing platform

Methods, systems, and computer-readable media for deploying an update to nodes propagated throughout a data center are provided. Launching new upgrade to hosting environment residing on the nodes typically invokes a mechanism (e.g., fabric controller) to form a group of nodes that are independent of one another with respect to upgrade domains, which are assigned to tenants (e.g., program components of service applications running within the data center) presently hosted by the nodes. The constraints of the update domains are articulated by service level agreements established for the service applications, respectively. Forming the group involves identifying independent nodes for membership, where no two members of the group host analogous tenants (belonging to a common service application) that are assigned to distinct update domains. However, it is acceptable to join to the group those nodes hosting analogous tenants that are each assigned to the same update domain.

Updating nodes considering service model constraints

Instrumentation to find the thread or process responsible for an application failure

A method of analyzing problem data from a computer application is disclosed. The method evaluates a memory dump, identifying call stacks within the memory dump that are related to application failures, creates a hash of the identified call stack and adds the hash to a database. The database may then be evaluated to look for trends in the error data such as whether the same call stack is causing problems.

Analyzing problem signatures

In embodiments of the invention, selected aspects of a process' call stacks are hashed, and the hash is used to capture the execution state of the process in a concise form and with minimal impact on the performance of the process and with no modification to the process code. The hash allows vendors and developers to classify failures in order to associate them with their known root causes to support rapid customer feedback and to maximize efficiency of correction. Further, the hash provides a metric usable to prioritize diagnosis of failures. In an embodiment of the invention, the identities of modules on the call stack are hashed in combination with some but not all offset information to minimize the affect of patches and minor changes to the code, and improve the ability to discriminate different execution paths.

Using a call stack hash to record the state of a process

Growing community of inventors

Seattle, WA, United States of America

Vamshidhar Radha Kommineni