The need for advanced autonomous operations in spacecraft systems recently has significantly increased, not only in deep space missions as e.g. the ESA projects MarsExpress or Rosetta where autonomy is inherent, but also for Earth Observation and other missions, since there is increasing interest to reduce operational costs.In order to implement S/C autonomy, a hierarchical FDIR architecture needs to be in place to support the overall system autonomy during nominal operations ("nominal autonomy") and during non-nominal operations ("failure case autonomy"). Nominal autonomy is usually supported by means of a mission timeline or mission planning driven concept, where different implementations can be conceived, depending on the respective mission goals. More sophisticated methods are required, if failure cases have to be considered. In this case the autonomy and FDIR implementation depends on the specific mission requirements and can vary from high levels of autonomy with automatic reconfigurations of S/C systems or subsystems to lower levels of autonomy, in which the spacecraft is switched to a safe state. In addition to the traditional FDIR complexity for deep space missions due to contact limitations and special manoeuvres as e.g. orbit insertions, also for Earth observation there is an increasing tendency to apply more complex autonomy and FDIR concepts in order to reduce the operational costs on ground. This paper presents concepts of hierarchical FDIR based on the experiences from ongoing ESA projects and addresses the required resources to cope with various autonomy requirements. The hierarchical FDIR concept is characterised by the definition of specifically acting instances and clearly defined interfaces between those instances as the individual FDIR levels. Usually the highest FDIR level is in charge of the execution of vital functions to ensure the S/C integrity, while lower level hierarchies operate on system or subsystem level and are usually software driven. A very important aspect in this context is the increased level of technological know-how regarding the FDIR implementation. The presented project examples show, that the overall system FDIR is usually implemented by means of hardware and software resources with a high level of "crosstalk" between both. As a consequence, each of the spacecraft subsystems needs to implement parts of the overall FDIR architecture.