Presto

Network Timeout in Hive Metastore Communication

warning
Connection ManagementUpdated Feb 23, 2026

Presto queries fail sporadically with SocketTimeoutException when communicating with Hive metastore, often under high query concurrency or when metastore is under-resourced relative to query load.

How to detect:

Monitor for increasing presto_execution_external_failures_one_minute_rate with java.net.SocketTimeoutException stack traces in logs. Check hive_ms.log on coordinator for errors. Look for correlation between failure rate and presto_execution_running_queries or presto_execution_started_queries_one_minute spikes.

Recommended action:

Increase coordinator node size to provide more memory for embedded metastore. Raise hive.metastore-timeout, hive.s3.connect-timeout, and hive.s3.socket-timeout values (e.g., to 3m). Increase metastore heap memory allocation. Consider implementing connection pooling or metastore caching to reduce load.