Main Report Request Page


There's certainly a lot on this page, isn't there? :) Well, let's go through everything one by one...


"Select Stored Report To View"
This dropdown list lets you select a stored report that you want to view. When something is selected in here, every other option on the page is ignored. Yes, every other option. Once this is selected, the report requested will be generated dynamically (that is, a "stored report" only stores the options you want to use to generate the report; it doesn't save the graphs or anything like that). You can't "edit" an existing report on this page; that can only be done in the administration section. However, you can create new reports here (but it won't overwrite reports that already exist should you give a new one the same name as an existing one). If selecting a report here doesn't automatically submit the form, you can hit the "Run It!" button.


"Run It!"
This is the form submission button. There's another one at the bottom of the page; they both do the same thing, which is to submit the form with whatever options are selected on it.


"Report Data Configuration"
This is where you configure what you want to see on the report. The interface selection can take two forms; if you have Javascript enabled in your browser, and the "Fancy Javascript hostname/interface selection method" option is enabled in the administration section, you'll see the fancy Javascript hostname/interface selection boxes. If not, you'll see one selection box with every interface listed in it. If you only monitor 10 or less interfaces, it's usually best to leave the fancy method disabled; it's easier to select from a 10-item list than it is to have to use three select boxes. But if you monitor hundreds of them (in which case, congratulations on getting that 10-terabyte hard drive), believe me -- the fancy method is better. Experiment with it and determine for yourself which method works better for you.

With the non-fancy method, you just go down the list and select each interface you want the report to cover. Since it's a multiple select box, some browsers behave differently; on Windows, for example, you usually have to hold down the "Control" key on your keyboard and click on each one you want; otherwise each click you make will deselect everything else on the list. In Netscape on Linux, you just click on each one you want (much more sane behavior, in my opinion). I don't know what Macs do. After you have every interface you want selected, continue with the other options.

With the fancy method, selection is a 2-step process. First, you select the hostname you want in the first box; that will populate the interfaces box with a list of all interfaces on that host for which stats exist in the database. Then, you go down the interfaces list, clicking on each one you want. Clicking adds the interface you clicked on to the third select box. Just click on each one you want, changing the "Hostname" box as required, until the third box is filled with the list of interfaces you want reports on. Then, continue with the other options. Don't worry about going through and selecting all the contents of the third box; the script does this for you when you submit it.

To add every interface from the left box to the right one, click "Add All." If you want to clear out the contents of the third box, click the "Clear All" button. The "Clear Deselected" button will do what it says it does except in Unix versions of Netscape; there, it will simply crash your browser. I'd have thought that in all the years Netscape has made a Unix client, they'd have fixed all the Javascript bugs in it, but apparently not...

A note about interfaces. If you have an interface configured to be monitored but it isn't responding to SNMP queries, it won't show up on this page; only interfaces with at least one collected entry will be on here. If you think about it, this could be a really useful diagnostic feature; trust me, it's not a bug... :)


There are five checkboxes and five dropdown lists to the right of the interface selection section. Pick which report types you want out of the first four checkboxes: bytes transferred, etc. A separate graph is generated for each type of info. The default is to just view the "bytes transferred" report.

The fifth checkbox turns on "multiple" reports; that is, ones in which you want the transfer averages of every interface you have selected added together and put on one graph. It will still make separate graphs for bytes, packets, drops, and errors as usual, but it will only generate a multi-IF graph for bytes transferred, not packets or drops or errors. It defaults to off.

The text box just below the "Multiple" checkbox is the averaging interval to use when making the multi-IF graph. You should keep this the same as your collection interval for best results (it defaults to that), but it can be any number of seconds you want. If you set it to less than 15 seconds, it will be set to 15 seconds for you; if you were to set it to, say, 1 second, the report would never finish generating and would just suck up all your RAM. It's ignored if the "Multiple" checkbox isn't selected.


Report Time Configuration
The first select box is the most confusing option. It has some entries with an "X" in them, and some without. Right next to it, you'll see a text box that defaults to "1" labeled "X=". Whenever you select an option in the dropdown list with an "X" in it, whatever is in the text box will be used in place of it by the report generator. So if you select "The past X days" and put "7" in the text box, you'll get a report covering the last week. If you select one of the options without an "X" in it, the "X=" box will be ignored.

Selecting "The entire contents of the database" will show you every stat collected for each interface you selected above, back to the beginning of time. You should take care with this option; a report with a million entries to be analyzed could take quite some time to generate, and suck up all available RAM. It is of course necessary to analyze each and every entry to get accurate averages, you see.

Selecting "A date/time range, set below" will enable the From and To date/time boxes. You should set each of these to whatever date and time range you want the report to cover and make sure you have "A date/time range, set below" chosen in the first select box; otherwise these dates will never be looked at.


And then there's the averaging variable. This only affects the text-only report tables (which aren't shown by default; see the Report Data Configuration section). It specifies how often you want to see a row in the report table, to put it in simple terms. The smaller this time increment, the more rows you'll get. If you tell it every five minutes and get a three-year report, your browser may never finish loading it. :) But of course, the smaller this value, the more accurate the averaging will be; hence the need to balance which one you desire more.... small report size, or high accuracy. Life is full of such balances for this very reason; it's almost Zen-like. Oh; this value also doesn't affect how much data is retrieved for a report; no matter what you have the options set to on this page, the report generator will always pull every stat collected on the selected interface(s) during the requested time period and use those precise counts and timespans to calculate everything.


Graph Configuration
The first two options are self-explanatory: the width ("X") and height ("Y") that you want the graph sizes to be. If you set the width to less than 650 pixels, the labels at the bottom will run together and become unreadable; if the height is less than 180, the same happens to the labels on the left side. You could hack the NISCA source code and change the font sizes and such, or change how many X- and Y-axis gridlines or labels there are, I suppose... just another checkmark on the "Pro" side of the Open Source half of the software preference checklist.

If you make a graph, say, 2000x1000, you'll be amazed at the detail you can see. Try it on a 1-day graph and you'll see what I mean. You can make a 1-week graph that's 4000x300 and it's like getting 7 1-day graphs all at once. This is a really cool thing to do if you want to compare, say, all the interfaces on a switch to each other (if you have enough memory to display that many images that large); then you can scroll left and right, find large peaks, and scroll up and down to see which interface(s) match(es) it. Play with it; go wild...

The other seven options select which color you want to use on the graphs for the graph's background, the plain text on it, the light and dark borders around its edges (gives it a kinda drop-shadow look), the little circle that marks the peaks, and the incoming and outgoing plot lines. These should be standard hex color codes; you know, like FF0000 for red, 00FF00 for green, 0000FF for blue, 00FFFF for cyan, FFFF00 for yellow, FF00FF for magenta, FFFFFF for white, 000000 for black, 888888 for grey, .... you get the idea. The page will actually update the "Sample Text" displayed as you change values, so you know what you'll get before you even get it.


Report Name Option
Here is where you have the opportunity to create a stored report. To do so, just enter a name for it in this box; this is the only thing you have to do. This will save every option on this page as a stored report, and then display that report. If the name entered already exists in the stored reports database, it will not be replaced. So if you have Javascript disabled and save a new report and view it and then return to the index page with your Back button, the report name field will still have a value in it and if you don't clear it out, it will try to create another report with that same name and give you an error about it. Just so you know.


Then there's just the other "Run It!" button and the "Clear" button, which just resets the form to the values all the fields had when you first loaded it. Happy reporting...


Main Help Index