DX Application Performance Management

Hi,

I have some questions about "Historical metrics" :

I always thought that historical metrics related with the smartstor database.

But yesterday what i thought needs to be reconsidered -

After an issue (FS full) on our traces FS, i did this actions

1/ stop the impacted collector
2/ erase all under traces directory
3/ start the collector

After that I have noticed, that the number of historical metrics has been divided in two.

Should i consider that there is some historical metric in the smartstor database and in the event DB (traces)?

Hi Amor,

You are correct, they are different dbs.

***SmartStor database: SmartStor records all application performance data (Introscope metrics) at all times while it is in production, without the need for an external database. This enables you to analyze historical data, to identify root causes of application downtime, or perform capacity analysis-without the need for an external database.

***Transaction Events database: The Enterprise Manager stores transaction traces and other event data in the Transaction Event database.

Thanks,

Jose

you can also refer to the Sizing and Peformance guide for detailed information:

"

Enterprise Manager Internal Databases

The Enterprise Manager maintains three internal databases: SmartStor, Transaction Events database (traces.db), and the metrics baselining (heuristics) database (baselines.db).

Introscope features such as Transaction Tracing, Transaction Trace sampling, and metrics baselining (heuristics) make limited and short-term demands on the disk subsystem. For this reason, the Transaction Event database (traces.db) and the metrics baseline (heuristics) database (baselines.db) do not require dedicated disk resources and can be collocated on the same disk. However, for optimal performance, CA Technologies requires you to locate SmartStor on a separate dedicated disk or I/O subsystem.
Note: For information about the internal database file directories and how to backup and handle metadata, see the CA APM Configuration and Administration Guide.

"

Regards,

Jose

Just some added info that by default metrics become "historical" after 8 minutes and Knowledge Base Article TEC603652 has more details on this:

https://comm.support.ca.com/?legacyid=TEC603652

Very interesting deails here. Two questions for my understanding based on what is given in this thread.

1. Is it true that if you delete traces directory and restart EM, historical metrics count will go down?

2. Based on KB article, is it true that if we increase memoryCache.elements, this will reduce historical metrics? How big value can be specify?

--------------------

In 9.5 the memory cache can be configured to be larger by increasing the
value of this property and restarting the EM:

introscope.enterprisemanager.memoryCache.elements

--------------------

1. No, traces directory stores transaction traces and other events. It does not store metrics, Hence deleting it would not make historical metrics count go down.

2. No, the purpose of this property is as follows:

introscope.enterprisemanager.memoryCache.elements

This property controls the number of data points that the Enterprise Manager keeps in an internal compressed memory cache to speed historical data queries.

Each data point contains 15 seconds of metric data. The default of 32 data points contains 8 minutes of historical metric data.

The range is 32 (8 minutes) to 11520 (48 hours), and must be divisible by 32.

You can set the cache larger than 32 data points for triaging situations when there are spikes in Workstation usage. For example, a value of 240 data points allows the Enterprise Manager to provide every historical query for the last hour from the memory cache.

Thanks for your detailed reply puran02.  One more follow up question.

How do we reduce historical metrics from exploding? There is another thread here also. Historical metrics limit

Thanks.

To be honest, to reduce the number of historical metrics, the method would be to reduce the number of the metrics. You can do this by either of following approach.

- Adjust the smarstor aging:

introscope.enterprisemanager.smartstor.tier1.age,

introscope.enterprisemanager.smartstor.tier2.age, and

introscope.enterprisemanager.smartstor.tier3.age.

- Clamp the metrics reported by agent in EM

with introscope.enterprisemanager.agent.metrics.limit

- Reduce the number of the metrics reported from the agent by changing the PBD configuration.

Marking as answered.