In this video we will provide some guidance and best practices for improving the performance of your network discovery. You can add nodes individually or you can discover a range of IP addresses using the built-in Network Sonar.
To get to the Network Sonar Wizard you select the Network Discovery under the settings menu.
The first thing you will see is a list of existing discoveries that have already been completed or are scheduled.
To start a new discovery, select the Add New Discovery link.
You'll now be presented with Network Sonar Wizard.
The first step in discovery is to outline your network. This is the first place that you can impact discovery performance. You can define a range of IP, specific IP addresses, or subnets. We recommend breaking up your discovery jobs into small subnets, or IP ranges. Remember that you can create as many separate discovery jobs as you'd like.
The next place that you can improve performance is through credentials. Yes, you can add multiple credentials to a discovery job, however, Orion will need to try all of those credentials against every device it finds within the range, So the more credentials you have the slower the job. If you do require multiple credentials you can add as many as you'd like but make sure the most important are at the top of the list as Orion will test the credentials in order.
The third way to split up jobs to improve performance is based on latency.
If you’re scanning a network where there is significant variance in the amount of latency, you’re probably going to want to split up your discovery job. For instance, let’s say that most of what you’re monitoring is in your central office along with your polling engine, but you’ve also got one small remote office with significant latency out across the WAN.
To run a reliable discovery for the remote office, you might have to increase your timeouts to catch everything. But now everything, including IPs that would be on the local network, are going to be affected by the increased thresholds and the whole job will run slower.
It’s better then to split up the discovery job and have one for the local network and one for the remote office.
Finally, if you have multiple polling engines, you will always have to split up your discovery jobs. You cannot have a single discovery job split between multiple polling engines. Make sure if you do have multiple polling engines that you are pointing your discovery jobs at the right one.
Now one of the questions that we are often asked is how do these discovery jobs affect performance and when should I schedule them to run? Will they have an impact on regular polling? For the most part, they really shouldn’t. Orion is really good at managing multiple tasks, and polling takes priority over everything else. So when you kick off a discovery job, Orion looks at that and says “OK, here’s this list of IPs I need to go scan.” But if you have a polling engine that’s pretty heavily taxed, polling might affect how long it takes the discovery job to finish running. That’s part of why we recommend that you run the discovery jobs immediately rather than in the middle of the night for the first time to see how long it takes the job to run. And that job time should be fairly consistent whether you're running the job in the middle of the day or the middle of the night. Orion polls at exactly the same rate 24 hours a day, seven days a week. It doesn’t matter what’s happening on the network or whether anyone is in the office.
If you have multiple discovery jobs you can use that information to offset the job start times to improve performance. If not, and you have multiple discovery jobs trying to run at once, all you're doing is eating up more RAM and slowing down the discovery.
So make sure that once you’ve gone to the trouble of breaking up your discovery jobs, you also pay attention to your job scheduling. Offset the job start times enough that one job has time to finish before the next job starts. If not, then all your hard work is for nothing.
So in summary, to get the best performance out of Orion split up your discovery jobs into logical, manageable chunks and pay attention to the job scheduling.
Disclaimer: Please note, any content posted herein is provided as a suggestion or recommendation to you for your internal use. This is not part of the SolarWinds software or documentation that you purchased from SolarWinds, and the information set forth herein may come from third parties. Your organization should internally review and assess to what extent, if any, such custom scripts or recommendations will be incorporated into your environment. You elect to use third-party content at your own risk, and you will be solely responsible for the incorporation of the same if any.