Currently, there is a plethora of low-power wide-area IoT networking solutions available, each targeting a specific niche of use-cases and deployment scenarios. Existing studies on reliability evaluations of IoT solutions rely on the assumption that a single technology is deployed in the service area, or different IoT technologies operate over dedicated spectrum bands. Here, we investigate the reliability performance of IoT communications in coexisting scenarios, where multiple competing radio-access technologies share spectrum resources. Our focus is on solutions exploiting grant-free communications, which are gaining traction due to their potential to lower the energy consumption, and have been adopted in recent IoT technologies like SigFox and LoRa. We first derive an analytical model of the interference, comprising both inter- and intra-technology interference sources. We then leverage the Poisson Cluster Process for modeling distribution of devices in the service area, and derive expressions for the communication reliability, energy consumption, and battery lifetime of IoT devices. Exploiting these expressions, we study the energy-reliability trade-offs and investigate strategies to maintain or improve communication reliability, while minimizing energy consumption in coexisting scenarios by proper adjustment of communications parameters at the device side and provisioning resources at the network side. We verify the analytical results via numerical evaluations, confirming their accuracy and performing optimization in some example networking setups.