When you start a Package, the package builder places the start time and the parameters being passed in onto the screen. This information can help you to know how long it will be before something is finished, or if it is being run correctly during a package run.

Status information is information written out by each object to the log screen. With status information you will know:

1. The time that the object started.
2. The name of the object.
3. The number of record operations that object has completed.
4. Any errors or messages that object was supposed to write out to the screen.

The purpose of status information is to keep the person running the package updated about what is currently executing in the package. All other messages will typically happen between the start and stop messages for an object.

Errors do happen. Sometimes it's a design flaw, and sometimes it's an execution flaw. Design errors typically don't show up until you are looking at the output from the conversion.

When you do get an error on the log screen, it's going to be an execution flaw. Execution error can be SQL that does not run, variables that are in the object and not in the variables used list files, (at least they don't exist at that location), etc.

An error message will include information on the object and its state as well as the error message of the exception that actually caused the error.

Some error messages will include the SQL (if the object uses SQL) that was active in one or more connections at the time of the error.

Some error messages will contain messages written by the designer to give you more user-friendly feedback.

Some error messages will be a little cryptic. Let us know when about these messages and we will include more user-friendly error messages in a future release.

Detail information includes special messages like the ones reported by Validate objects.

When a package completes it will also (optionally) list all the variables used by that package.

Statistical information is also written to the Log screen. This includes the total number of records, total running time, and average time per record. This kind of metrics data is valuable for predicting performance on your server.

Performance measurement is going to be affected by the following things:

1. Machine speed.
2. Data types. (input coming from Flat files is going to be faster)
3. SQL. (there are fast and slow ways of doing anything--choose the fast way)
4. Target database. (the place the data is being sent can be a slow structure too)
5. Network speeds.
6. Hard drive speeds and configuration.

There are many considerations when making a package run fast. If your package is running slower than you want, use the information on each step to determine which steps to optimize.