Portugese Version: https://bmcsites.force.com/casemgmt/sc_KnowledgeArticle?sfdcid=000187038

Where can one find information on how to create a CSV file parser ETL in TrueSight Capacity Optimization (TSCO)?
 

There are two different types of CSV file parsers that can be created. One CSV parser imports data in a columnar format (so each row of data has multiple columns with different metrics) and the other in a vertical format (each row of data in the input file is for a single metric). 

Here is the TSCO 20.02 documentation regarding the two available CSV Parser ETLs (for other TSCO versions please consult the official product documentation):

• BMC Capacity Optimization integration with the CSV file parser
• BMC Capacity Optimization configuration with CSV columnar file parser
   

Outside of the TrueSight Capacity Optimization (TSCO) product documentation, there is also documentation that provides additional examples and recommendations related to importing data via a CSV parser.  

For example, this Hot Tip document:

• Hot Tip - Integrating CSV files in BCO

The TSCO Technical Support Knowledgebase also includes useful information related to implementing the TSCO CSV Parser ETL -- particularly in relation to common error messages and version-specific data import problems.

When first creating a CSV parser some recommendations include:

• Keep it simple at first. Once you have a simple data input file that processes correctly in your test environment it is a lot easier to add additional data for import than trying to debug an initial large data load.
• If you do have problems and are looking for help from Technical Support try to be really specific about what you are trying to accomplish. For a CSV parser, it is usually best to send a sample CSV file that isn't working and the ETL Run Configuration you currently have.

Note that for the CSV parser there can be a steep initial learning curve but once you've created one successfully creating others will be much easier.

Q: Can a TSCO Generic CSV Parser ETL import Business Driver (transaction count) data?

Yes, the CSV parser can import both System Data (metrics like CPU Utilization, I/O Rates, and so on) and Business Driver data (metrics that represent business drivers such as web page hits, number of users logged in, and so on).

When you are creating a CSV parser to import Business Driver data, for example, Logged In Users, you'll want to select the "WKLDAT - Workload Data" Dataset.  Then when you import the data you'll want to pick the right BCO Metric for the import.  To do that I'd go to the Administration -> Data Warehouse -> Datasets and then search for a good metric in the table.  For the number you are talking about the "USERS_CURRENT" (Described as 'The users currently in the system.') might be a good one to use.

Once you have the number of users logged into the system imported into TSCO what you are talking about creating is a Performance Load Analysis (PLA).  That is where TSCO  plots the utilization of a resource (CPU, I/O, memory) against a Business Driver (users in this case).  You can then also use that PLA to do a Correlation Analysis (to determine how well correlated the performance metric is to the measured number of users on the system.

If importing both System Data and Business Driver data it is necessary to create two separate CSV Parser ETLs -- one to bring in the system data and one to bring in the business drivers data.  In the Supported Datasets section of the documentation, it states, "This ETL supports vertical datasets: SYSDAT, WKLDAT. Note that only one dataset at a time is supported.".  So, you'd want to separate your input data into two separate CSV input files broken down by data type and then input the system data via one CSV Parser ETL and the business driver data via a 2nd separate CSV parser ETL.

Q: Can a single TSCO Generic CSV Parser ETL import several different System Metrics via a single CSV input file?

Yes.  In the BMC Capacity Optimization integration with the CSV file parser online help page in the "Input file format for system metrics" section at the bottom take a look at the first sample input file.  That input file brings the CPU_UTIL, NET_IN_BYTE_RATE, and BYIF_IN_BYTE_RATE metrics into BCO via a single ETL.  TSCO can import as many different system metrics as you'd like via a single CSV input file and separately can you have as many different business driver metrics as you'd like imported via a single CSV input file.  But they need to be separate files imported by two separate CSV Parser ETLs.

Q: What are the standard field separators to use for the CSV parser ETL?

Q: What is the best Timestamp format to use in a CSV input file?

Q: Can the pipe (|) character be used as a field separator?

Sample steps to create an Open ETL template

• Select "Generic - CSV columnar file parser" as the "ETL Module"
• Select "Based on an Open ETL template" in the Module Selection.
• In the 'Select Open ETL template' field select the ETL template you just created.
• Under Entity Catalog select 'Shared Entity Catalog'.
• In the 'Sharing with Entity Catalog' field select the Entity Catalog where the machines this ETL should augment with additional metrics are defined.
• Under 'Object Relationships' select "leave all new entities in 'Newly Discovered'" (since this ETL isn't defining relationships it is just enhancing existing entities with a new metric.
• Under File, Location specifies the location of your input file.
• Click the 'Advanced' button and under the 'Loader configuration' section set "Skip entity creation" to 'true' if you only want this ETL to add this configuration metric to entities that already exist (you don't want it to create a new entity that only carries this one metric)

Custom BIRT Reports

The following KA has the links to the BCO How-To Guides: How to create custom reports in BCO which is a high-quality document related to the creation of custom BIRT report templates:

  000029447: Where can I find documentation related to the creation of custom ETLs and BIRT Advanced Report templates in the TSCO Integration Studio?

When first creating BIRT reports some recommendations are:

• Check to see that you can't accomplish what you want using a report created from a TSCO Analysis. It is much easier to create a report from an Analysis -- but the ease of creation does limit your flexibility.
• Check to see that there isn't an existing out-of-the-box TSCO BIRT report template that can accomplish what you want to do with your custom template (or one that is close to what you want). You can use that TSCO out-of-the-box template as the basis for your custom template (you download it from the TSCO  console and then you can upload it to the TSCO Integration Studio).
• If you are embedding custom SQL in a BIRT report check that the functionality can't be better implemented using the TSCO DataMart Views.  This is a way to create materialized views that can then be referenced by a BIRT report as the data source
• If you need to engage Technical Support it is always best to be very specific about what you are trying to accomplish and how far you've gotten (what points are you stuck on). A screenshot or example of what you are trying to display in TSCO is a good way to give Technical Support a good idea of what you are trying to do.