Phoenix Row-Level Write Lock Configuration for Concurrent Requests

Phoenix Row-Level Write Lock Configuration for Concurrent Requests

Goal: Determine which Phoenix configuration parameters control row-level write lock granularity for concurrent INSERT and UPDATE operations on non-partitioned tables.

Constraints: Phoenix 5.0 includes optimistic concurrency control, yet write requests targeting the same row or region can still serialize and increase latency. Behavior differs across Phoenix versions, and network partitions may mimic lock contention. The effect of phoenix.client.operation.timeout on actual lock wait time remains unclear for non-partitioned schemas.

Questions: Which settings directly narrow row-level write lock scope without changing the table schema? Can region placement or salting reduce hot-spot contention while preserving the current layout? What log indicators separate lock wait latency from general region server overload?