Describe what problems with data consistency can arise when implementing distributed transactions across multiple services using PostgreSQL, and how you usually solve them.

Distributed transactions between services on PostgreSQL encounter the following data consistency issues:

• Lack of atomicity: Standard PostgreSQL transactions do not extend across multiple services, making it impossible to ensure atomicity of operations affecting multiple databases.
• Consistency problem: Data may become inconsistent if one service successfully completes a transaction while another does not.
• Locking and deadlock issues: Attempts to coordinate transactions between services can lead to locking.

Common approaches to address these issues include:

• Two-phase commit (2PC) — coordinating commits through a transaction manager, but this is complex and reduces performance.
• Eventual consistency and the SAGA pattern — breaking the transaction into a series of local transactions with compensating actions in case of errors.
• Using messaging and queues (e.g., Kafka) for asynchronous data synchronization.

In practice, the SAGA pattern is most often chosen because it is more resilient and scalable. In Java, this is implemented through orchestration or choreography of microservices with compensating methods.