Hi.
Hive is a multi-fact application require a number of monitoring tools to ensure that it is performing as it should. You need to ensure among other things: 1. Hive server is up and running through os level check via cron say every minute 24x7. --. netstat -plten|grep port 2. The connection to hive metastore are active though cron similar to above 3. Check Hive errorlog hive.log for any errors/issues 4. check the metastore errorlog. We run ours on Oracle DB and as usual monitor for errors in the alert.log 5. With Hive .14 version and the locking/concurrency handled by the metastore, check for blocking locks and stale connections in Oracle DB 6. You can now check for locks and Open/Aborted transactions within Hive. These need to be tracked on a periodic basis for any issues. For example, does it make sense to have an Open transaction running for two daya! 7. As Hive utilises Yarn and MapReduce any issues with the aforementioned applications will impact Hive functionality. You can see hive tasks being submitted by not running for resource issues. I am sure there are other checks that need to be done as well. HTH, Mich From: Ashok Kumar [mailto:ashok34...@yahoo.com] Sent: 13 November 2015 17:49 To: User <user@hive.apache.org> Subject: Best practices for monitoring hive Hi, I would like to know best practices to monitor the health and performance of Hive and hive server, trouble shooting and catching errors etc. to be clear we do not use any bespoke monitoring tool and keen on developing our own in house tools to be integrated into general monitoring tools to be picked up by operations. greetings and thanks