Likewise, as we look forward to Notes and Domino 7, we're looking at bigger and bigger projects. Fortunately, Release 7 is going to provide us with even more and better troubleshooting tools.

In this article, we'll take a first look at what you can expect.

Domino Domain Monitoring
If you've read some of our previous R7 articles, you might have heard us mention DDM (Domino Domain Monitoring). DDM is a codebase that's been added into all server tasks and provides self-monitoring hooks into the server task operation.

For the record, since DDM is code inside the R7 server tasks, it can only monitor R7 tasks. It's not code that'll go out and look at server tasks running on other, non-R7 servers. There are some excellent third party products out there that will do remote server monitor. One such is GSX's Server Monitor (at http://www.gsx.net), shown in Figure A.

FIGURE A

If you want to monitor non-DDM enabled servers, Server Monitor from GSX should help. Click picture for a larger image.

But let's get back to what's built into DDM and some of the troubling troubleshooting problems we've seen in the past. If you've done any amount of Notes and Domino server development or maintenance, you know how very much of a hassle it can be to identify problematic agents.

It's been even more of a problem in a Web-serving environment. Here, you've got many agents running concurrently, with an entirely unpredictable access pattern coming in from all the people visiting your Web pages on either your intranet or over the Internet, or both.

When you've got all these hits coming in from any number of entities, accessing any number of unpredictable pages and triggering any number of unpredictable operations, how can you tell which agent is gobbling up too much CPU or RAM? Even if, after herculean efforts, you're able to figure out which agent or agents are the resource hogs, what chunk of code would benefit most from optimization?

It's a non-trivial problem that the new DDM system may be able to help.

To do this, DDM implements what Lotus calls "probes," sort of like the test probes you'd see on electronic test instrumentation. By looking at the performance of individual application agents, you can get a better idea of what's going on.