error database alert
This monitors the number of messages in the error database.
The "error count" alert looks at the error count statistic for the thread as stored in shared memory. This is the count that shows up in the GUI under "failed" when you look at a thread’s status.
The "error database" alert looks at the number of messages in the error database that arrived from the given thread.
How these numbers differ:
- You can change one number
without changing the other. For example, you can reset monitoring statistics, and
therefore the "error count" alert, without affecting the error database when using
hcimsiutil.
You can also change the number of messages in the error database, therefore changing the "error database" alert. This does not affect monitoring statistics by resending messages or reinitializing the database.
- Messages in the error database
do not necessarily come from a protocol thread. If you attempt to create an "error
database" alert, then you can select translation threads as the source for the
alert. For example, you can create a raw route and use the
xlateError
procedure to generate errors. If you set your alert to trigger on the protocol threads, then the alert never fires. If you set the source to be the translation thread, then the alert fires.
Therefore, select which alert to use depending on what to monitor:
- Use the error database alert to monitor error database depth.
- Use the error count alert to measure how many messages fail going through a protocol thread.
The source value is a set of thread names. The dialog box for configuring the source is the same as other thread alerts, such as thread status and protocol status.
Click Source to open the Select Alert Source dialog box.
This dialog box provides a choice for applying the alert to all threads or specific threads.
Selecting the
option implies that all error messages in the error database are selected. In this case, all existing thread names are shown in the source field.