Description of problem: Currently, Pacemaker supports the "failure-timeout" resource meta-attribute, which will automatically clear a resource's failure history once it has no new failures in that much time. However, Pacemaker now tracks failure counts per operation, and is planned to offer more flexible failure handling, so it will be helpful to be able to set such timeouts individually per operation.

Actual results: Pacemaker supports setting failure-timeout as a resource meta-attribute.

Expected results: Pacemaker additionally supports setting the option as an operation property.

Additional info: The option might be renamed as part of this change.