FB TraceManager 3 - Documentation

Hotspots Analysis

Hotspots Analysis

Previous topic Next topic  

Hotspots Analysis

Previous topic Next topic  

Hotspots Analysis, part of the Enterprise Edition, is a new feature in V3.0. It offers a near-real-time, on-the-flow aggregated view on parsed trace data. Have a look on the following screenshot.

 

fbtm_hotspots_output.zoom25

 

What you see here in the so-called Statement Hotspot is an aggregated view on executed statements per database with ~ 60 attributes/measures. The most simple measures are:

 

Execution count (Exec Count)
Total Execution time (Exec Total [ms])
Preparation count (Prep Count)
Various I/O statistics
and much more ...

 

To answer if you are re-using a prepared statement with the same SQL is very easy with this sort of data. Prep Count towards Exec Count give you a very quick answer.

 

There is also a Procedure/Function Hotspot and a Trigger Hotspot, using the procedure/function/trigger name as part of the aggregation key beside the database name.

 

New in V3.5: V3.5 adds support in Hotspots Analysis for stored functions and package names of Firebird 3. The package name is now part of the aggregation key to have a distinct view on equally named PSQL modules coming from different or no package.

New in V3.6: V3.6 got support for an optional and configurable auto refresh possibility. Defaults for the trace session can be set in the Register Project wizard.

 

Hint 1: This sort of on-the-fly aggregated data, without first persisting into the FBTM3.fdb database, enables you to quickly get an high-level overview on the current database activity.

Hint 2: If you are interested in hotspot data only and not in raw/parsed trace data, make use of the new Disabled option in the Raw/Parsed visualization area. This results in a light-weight trace session background worker thread, without putting anything into the visualization processing queue.

Hint 3: To get the full "hotspot experience", you have to make sure to include various trace configuration options for the project in the Register Project wizard, otherwise you won't see useful data for e.g. Prep Count etc. The new project trace configuration templates make that easy. There is a pre-installed template called Hotspots Analysis.

 

 

What's next: At this point, you are finished with the Quick Start Guide. The next section Trace Data Management discusses various ways to work with persisted trace data.

 

See also

Main Screen

Register Server

Register Database

Register Project

Start Trace Session

Trace Data Visualization