Monitoring events
Use these guidelines:
- Select an optimal combination for best load balancing.
- It is pointless to monitor a configuration whose application database contains no event handlers or fires no events. However, note that Infor or another vendor from whom you purchased an add-on product might have included event handlers or code that fires events in their distribution, so you should check for these when you consider which configurations to monitor.
- You can create new configurations solely for monitoring by the Event Service.
Each configuration in the monitoring list represents an independent software timer on an independent operating system thread whose time slices can be scheduled (by the operating system) to run on a different CPU than the others, where available.
Example 1
Hardware: One superfast/multiprocessor server.
Goal: You want to handle all events on one server.
Implementation: Pick a set of configurations, each of which points to a different application database, and which together span all application databases. Monitor that set.
This table provides the details:
Configuration | App database | Monitor? |
---|---|---|
ILL_Prod | ILL_App | Yes |
ILL_Test | ILL_App | No |
CAL_Prod | CAL_App | Yes |
OH_Prod | OH_App | Yes |
Example 2
Data: Six sites (ILL, CAL, OH, COL, CHINA, and FRANCE) and two entities (USA and CRP). The entities will not have events in this example, because they are only used for G/L roll-up.
Hardware: Three servers, one of which (Util1) is comfortably loaded with other functions such as replication, TaskMan, or IDO communications.
Goal: Spread the load from the six site application databases over the two free servers (Util2 and Util3).
Implementation: Monitor the largest, heavy-traffic site alone or with another lesser site. Monitor the rest together.
This table provides the details:
Server | Configuration | App database | Monitor? |
---|---|---|---|
Util2 | ILL | ILL_App | Yes |
Util2 | CAL | CAL_App | Yes |
Util3 | OH | OH_App | Yes |
Util3 | COL | COL_App | Yes |
Util3 | CHINA | CHINA_App | Yes |
Util3 | FRANCE | FRANCE_App | Yes |
Example 3
Data: One heavily used site (ILL_App) with many events firing per second, and/or a need for immediate response to queued and adjourned events.
Hardware: One superfast/quadprocessor server (Util1), plus other slower servers.
Goal: Select items from the super-site’s event queues on as many processors as possible.
Implementation: Monitor the super-site multiple times from the superfast server (once per CPU) and once from the other servers. Monitor the other sites from the other servers, evenly.
This table provides the details:
Server | Configuration | App database | Monitor? |
---|---|---|---|
Util1 | ILL_1 | ILL_App | Yes |
Util1 | ILL_2 | ILL_App | Yes |
Util1 | ILL_3 | ILL_App | Yes |
Util1 | ILL_Prod | ILL_App | Yes |
Util2 | CAL_Prod | CAL_App | Yes |
Util2 | OH_Prod | OH_App | Yes |
Util2 | ILL_Prod | ILL_App | Yes |
Util3 | ILL_Prod | ILL_App | Yes |
Util3 | COL_Prod | COL_App | Yes |
Util3 | CHINA_Prod | CHINA_App | Yes |
Util3 | FRANCE_Prod | FRANCE_App | Yes |