Reporting Questions

[jtata]

We are noticing some issues with the built in reports that makes it somewhat difficult to generate accurate SLA type reports. The two biggest concerns are:

-If you run an Availability Report only against a single host/service, Scheduled Downtime is accounted for and not held against your total availability. If the same report is run against a host or service group Scheduled Downtime is NOT accounted for, resulting in lower availability %. (I've actually seen this in Nagios Core as well going all the way back to 2.x).

-Generated reports aren't "pretty" enough to use for presentation to stakeholders.

We really need a way to have more control over what is contained in each report and how its presented One avenue I'm looking at is using the MySQL ODBC connector to pull data direct from the MySQL DB, however I don't know how to find exactly what queries give me the data I need.

Is there a guide to customizing/modifying reports out there? Failing that is there a way to look at the code behind the existing reports so I can see how they arrive at their numbers? That could at least get me moving in the right direction.

[tonyyarusso]

I don't know that I can reply to this directly with current version of XI, but I can tell you that improved reporting is a very high priority for the next major release of XI. I'll point this thread out to the person working on that, so if you can provide as much detail as possible about what your needs are I'm sure he'll be glad to take that into account while implementing things.

[jtata]

I'm glad to hear reporting is being looked at. Nagios is a great platform from a functionality standpoint, but its reporting is only barely adequate in enterprise terms. It takes way too long to get data out of the system that should be accessible with a couple clicks.

My main priorities would be:
-Reports have to work and be accurate, especially with regards to Scheduled and Program downtime. If I have only scheduled downtime during a month my SLA should be 100% in ALL reports, not 100% on a host and 99.x on a hostgroup.
-I need to be able to modify the look/feel of a report. For troubleshooting and back end use its fine to have generic bitmap looking graphs etc, but I need to be able to change things like chart type, layout, colors etc. to preset to my higher-ups. Custom Branding would be nice too (change logo, etc.).
-More granularity regarding sources. Currently I can report on a host, service, or groups. I'd like to be able to create an availably report for Host X, Y and Z and services A and B on each, regardless of their host/group status. Likewise filtering, exclude
-Ability to save report definitions. Once I create the report above I should be able to save it so i can run it again any time with a click.
-Report delivery. There are plugins that do this but they're clunky and outside the system. I should be able to have a saved report show up in my inbox or file server automatically and change the report delivery parameters via the web UI.
-Documentation or API for external reporting. There is no way you as developers can anticipate every possible reporting need. If my boss tells me he needs a pie chart of alert acknowledgments correlated to staff vacation days you probably aren't going to have a prebuilt report that can do that...but if I can easily push my Nagios data into SSRS or Crystal then I can probably build what I need.

Many other enterprise monitoring packages have this level of functionality in their reporting already, though they are either vastly more expensive or not flexible as Nagios.

[mguthrie]

This is excellent feedback. I know some of these features are underway and will be available sometime around the new year, but I don't know off hand what all of them will be. All of the reports will be exportable to either pdf or csv, and several other major improvements will be taking place. We've sometimes struggled to get concrete feedback as to what customers need in their reports, so we really do appreciate you being specific with what you're looking for.

[jtata]

I did figure out a workaround for the scheduled downtime issue. What I did was create a time period called SLA that excludes our normal maintenance windows. If I specify that time period when running my availability reports the numbers are correct. This obviously is not ideal because many environments have scheduled downtime that does not fall into a pre-defined window, but I thought I'd post it anyway in case someone else runs into the same issue and needs a quick fix.