Is your server behaving normally?
SQL Monitor shows you at a glance...
When you see performance spikes in SQL Server, how do you know if you're looking at a problem or something within the range of normal behavior?
Looking at past performance data is often not enough. Ideally, you want a baseline, so you can measure your performance data against what's normal for that specific server.
SQL Monitor calculates and displays baselines for you automatically, using the performance data that it already records.
Without any extra work to gather the data, you can check a baseline for CPU, memory, and whatever else you need on each server.
SQL Monitor takes the latest performance data into account, so you're always looking at real life performance, not a baseline that's tied to a single period in the past.
See how metrics vary over time
To get a simple baseline, you can compare a metric with its performance last week, or at any period you know was normal. The two appear side by side on the same graph.
For a full overview, you can extend this baseline to cover multiple periods, such as the last seven weeks, or the last 10 Fridays at 3-4am.
If you want to get a feel for a metric — how and when it spikes, how it's changing over time — you can view each baseline period as a separate line, making it easier to spot gradual or drastic shifts.
Check if performance spikes are within expected behavior
When you need to see the range of expected behavior, SQL Monitor can combine your baselines into regions. The dark inner region shows you where the metric falls most of the time. The outer region shows the full range.
It's a handy way to see the shape and spread of your data, so you can look past the noise without the flattening and distortion that comes from a simple average.
Investigate problems with pan-and-zoom performance graphs
To spot correlations between performance spikes and see what's going on inside your servers, you can compare multiple metrics on a single performance graph.
The graphs pan and zoom, so you can drag them back and forward in time, or zero-in on hotspots, to get right to the root of performance problems.