I don't see any purge mechanism, though I've only been collecting data since May 21. So maybe by TC16, I will have a clearer answer as to whether it actually purges old data.
you could purge yourself by stopping tabmon temporarily and then connecting to the PG via a SQL tool and running a command such as:
delete * from countersamples
where timestamp <= 'yyyy-mm-dd'
My workbook this morning has data from 19Nov2015 to 20July2016. The packaged workbook is 200,891KB in size.
A couple of observations:
- Absolutely no information is displaying on the Network tab.
- Given the latest push-back from our IT dept. for additional hard drive space I think logging total disk consumption over time would be a better metric for me. Then I can show how free space is gradually disappearing as time goes on due to the increase in the Desktop licenses and thus reporting & extract activities.
- Logs show "Unable to query MBean." for multiple entries and QueryCache hits/misses counters can't be found according to the log. I am using "iwebclientappservice" and "service.jmx_enabled true" is set.
- Not being a dyed-in-the-wool server guy much of the TabMon report doesn't makes sense to me and I'm not really finding as useful as it's supposed to be (lack of documentation surrounding the metrics is a huge factor though). Maybe because much of the server infrastructure environment is so stable I don't have to worry about it. I dunno.
1 of 1 people found this helpful
For #1, I have no thoughts here how to get the network info as we're currently not capturing this metric.
For #2, it's kinda ashame as disk space is so inexpensive nowadays. In other words, we don't really automatically actively monitor this as our servers have so much disk capacity that it would take us a decade to get even get close to max. And the general sense here is that time is better spent on other places.
For #3, we have JMX enabled=true (tabadmin set service.jmx_enabled true) and are currently capturing cache and the java info via the defined metrics within the config. But we are on 9.2, and I recall a post from Mr. Russell The new (and fiendishly hidden) Tableau Server monitoring improvements in 9.3 - Tableau Love stating that the metric names have changed. I'm not sure if they ever closed the loop on this?
For #4, we only use tabmon to capture cache and java metrics that perfmon could not capture. Therefore, it has some utility for us, but is fully automated and gives us a good view into the metrics. I have a post somewhere else if you want me to find it about the metrics that we are capturing in relation to Java and cache.
Thanks for the #3 link, Jeff. TabMon is now sort of useless to me
Loop is not closed yet. That namespace is still "bad" for 9.3. Sorry.