Hydrogen is increasingly being used as an alternative renewable energy carrier because of its potential to reduce carbon emissions in transportation and heavy industry. Nevertheless, this transition necessitates establishing an infrastructure for storage, transporting, and distributing hydrogen. This infrastructure is typically equipped with pressure relief devices (PRD) to protect systems from uncontrolled pressure increases. Without PRDs, a substantial pressure increase has the potential to rupture equipment and lead to a hydrogen release, which could lead to fires, explosions, and significant damage. However, recent incidents have shown that these PRDs can also be the root cause of leaks and releases. Therefore, there is a need to understand the conditions when PRDs increase the risk versus when these devices effectively mitigate the risk. This paper presents the first comprehensive analysis of past hydrogen incidents involving PRDs and considers when these components succeed and when they fail. We then analyze a diverse, representative subset of events involving PRDs using a root cause analysis (RCA) methodology known as Tripod Beta. Further, we introduce a taxonomy breakdown of the root causes of these incidents. Finally, we provide conclusions related to the observed root causes. The most frequently observed failure modes leading to PRD incidents are 1.) spurious operations of PRDs and 2.) high input to the PRD from components upstream of the PRD, which then activate PRDs. When comparing PRD-mitigated events to PRD-initiated events, we found that PRDs initiated more events than they mitigated; however, it is yet to be seen if this is an artifact of poor (success) data reporting or a true prevalence of PRD activations which undermines their value as a protective feature. The use of Tripod Beta shows that using systematic approaches to incident investigation and analysis provides more causal insight into risk mitigation than simply documenting the occurrence of a single failure. To understand the risk tradeoff that PRD offers, we need more rigorous assessments for the risk & reliability of these components. The results of this study motivate research, design and operational changes, and upcoming codes and standards developments to ensure the continuous, safe, and reliable operation of hydrogen systems.