Digital health technology utilizing wearables, IoT and mobile devices has been successfully applied in the monitoring of numerous diseases and conditions. However, intervention, in response to monitored data, is yet to benefit from technological support and continues to follow a traditional point-of-care delivery model by providers and health professionals. Mental health is an example of a critical health area in dire need for technology solutions to enable timely, effective and scalable interventions. This is especially the case with an increasing prevalence of mental health conditions and a declining capacity of the healthcare professional workforce. Numerous studies reveal the potential for peer support groups as an effective, scalable, cost-effective, first-line of response in mental health interventions. Peer support helps participants, at low and moderate risk, better understand their diseases or conditions and empowers them to take control of their own health. Peer support interactions also seems to inform health professionals with insights and intricate knowledge, making it effectively a learning health system. This paper proposes a software architecture to better enable "peer-sourcing". We present related work and show how the proposed architecture might draw similarity to and differences from crowd-sourcing architectures. We also present a study in which we interacted with service users (mental health patients) and mental healthcare professionals to better understand and elicit the key requirements for the software architecture.