Take the 2-minute tour ×
Database Administrators Stack Exchange is a question and answer site for database professionals who wish to improve their database skills and learn from others in the community. It's 100% free, no registration required.

I am a replication rookie. We have a very simple transactional replication set up in SQL Server. I am looking at the Replication Monitor on the subscriber and one of my subscriptions show Poor Current Average Performance. My understanding is this is derived from the latency threshold but when I look at the detail of the subscription, Performance is Excellent and Latency is 00:00:00. For what it's worth, Synchronization Status is fine.

Why is this current average performance poor? Over what time period is this "average"? This subscription has had no noticeable issues and has been active for some time.

(I am asking for my own knowledge as well as ensuring there is not something broken/sub-optimal that should be modified)

share|improve this question
1  
Agree w Mr. Denny re: excellent, poor, etc. My favorite repl measure (alert) is distributor latency. If the distributor can't talk to subscribers, latency climbs until it catches up. Set an alert for it (low threshold), and monitor what is normal for your setup, then adjust so you only get woke up for real problems. –  Eric Higgins Jun 5 '12 at 17:31

2 Answers 2

up vote 4 down vote accepted

Personally I ignore the excellent, poor, etc. As long as the number of records pending isn't very high, and the amount of time to catch up isn't very high I consider everything is fine. That front screen monitor is pretty misleading most of the time.

share|improve this answer

You can customize what the threshold is in ReplMonitor for the different levels, but you will go mad trying to watch ReplMonitor all day long or configure alerts for all your subscriptions. Like Denny mentioned, distribution latency is really the key. I wrote an email proc to send alerts for all subscriptions having latency issues: http://sqlwebpedia.com/content/sql-replication-undelivered-command-count and one for catching errors (ReplMonitor is green when retrying a failed event) http://sqlwebpedia.com/content/sql-replication-error-summary. I rarely use ReplMonitor now except for running TracerTokens for real-time issue investigation. Also check my site for SQL Saturday handouts and presentation decks. I also did one for SQL Pass Dev VC which was recorded.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.