Some general questions

lcsw2013

1. What’s the best upgrade path from 12.1 to current on npm? I just took over a new environment that needs lots of work. Right now I’m just auditing and trying to make decisions on on best ways to deploy a better system and design.

2. Theoretical limits. I know SolarWinds has a soft cap of 12,000 elements per node. All dependent on several factors. But how would you go about getting the best return of investment on all the hardware and license you already have? For instance I have a beefed up server. I throw everything at it and it barely breaks a sweat but software wise I get close to 12 k and SolarWinds likes to whine about it. Would license stacking take care of this issue to basically make use of the hardware assigned to do the polling?

3. This environment is extremely heavy on SAM. 90% of all monitoring is software monitoring. I know Sam has its limits too. How can I get Sam to respond correctly to high heavy loads? For example 13 pollers and all with an average of 12k components each being monitored.

The whole point here is consolidation into the smallest amount of machines I need to make it work. I don’t want to have an expansive environment all because the software is capped. This is the part of forecasting that always gets me frustrated because I want to go over kill on specs but what’s the use if the software can’t handle it? See what I mean?

As I said I’m auditing and creating reports to understand what’s the main purpose this environment will be used for. Then working on resolving all issues found and upgrade as needed.

Ipsla and netflow aren’t even being used due to the modules currently being handicapped. Site performs is down the drain. And I’ve been task of turning this around to give it value and purpose.

any help suggestions, tips, tricks, lottery numbers...all appreciated haha

Find more posts tagged with

fix

npm

upgrade

Accepted answers

All comments

serena

lcsw2013 wrote:
1. What’s the best upgrade path from 12.1 to current on npm? I just took over a new environment that needs lots of work. Right now I’m just auditing and trying to make decisions on on best ways to deploy a better system and design.
2. Theoretical limits. I know SolarWinds has a soft cap of 12,000 elements per node. All dependent on several factors. But how would you go about getting the best return of investment on all the hardware and license you already have? For instance I have a beefed up server. I throw everything at it and it barely breaks a sweat but software wise I get close to 12 k and SolarWinds likes to whine about it. Would license stacking take care of this issue to basically make use of the hardware assigned to do the polling?
3. This environment is extremely heavy on SAM. 90% of all monitoring is software monitoring. I know Sam has its limits too. How can I get Sam to respond correctly to high heavy loads? For example 13 pollers and all with an average of 12k components each being monitored.
The whole point here is consolidation into the smallest amount of machines I need to make it work. I don’t want to have an expansive environment all because the software is capped. This is the part of forecasting that always gets me frustrated because I want to go over kill on specs but what’s the use if the software can’t handle it? See what I mean?
As I said I’m auditing and creating reports to understand what’s the main purpose this environment will be used for. Then working on resolving all issues found and upgrade as needed.
Ipsla and netflow aren’t even being used due to the modules currently being handicapped. Site performs is down the drain. And I’ve been task of turning this around to give it value and purpose.
any help suggestions, tips, tricks, lottery numbers...all appreciated haha

1. We support direct upgrades from NPM 12.1 to the latest. You simply need to be prepared with the correct infrastructure. This document will help How to Migrate Orion to a New Server With Zero Downtime

2, I recommend reviewing our scalability guidelines, because in NPM 12.3 we drastically increased what we can support Scalability Engine Guidelines for SolarWinds products

mesverrum

To find your upgrade path just download the latest installer and run it, the first thing it does is the preflight checks where it will tell you if there are any intermediate steps that need to be done to get you up to date.

So the 12000 element limitation is somewhat artificial, you can stack additional APE licenses on the primary and additional pollers to increase the limit and get more use from the hardware you have in place.

Also, the official capacity of sam components is supposed to be around 10,000 components per engine, but that is a very rough estimate. For example the Appinsight components do things like run a single scripted query and then parse the results of that and turn them into hundreds of separate components. A single appinsight for SQL instance will often have 800 or more components spin up from it, but the load on your server is nothing like if you had added 800 individual components to some monster template. You really do just need to figure your SAM loads based on more vague things like keeping an eye on the % they show on the engines page, plus monitoring the perfmon counters on the orion server itself to make sure it isn't queuing up. The impacts from unknown/misconfigured components tends to be pretty painful too, because things like time outs and retries tie up more resources for longer than just getting the normal up/down statuses.

lcsw2013

Serena,

Can you guide me towards some further guides to help me learn more about newer capabilities please? Right now I'm in a data gathering mode and research. I just took over a brand new environment. And right now I'm building diagrams and design plans to improve this aging old handicapped environment.

And I'm more used to the old versions. And old limitations. I haven't been able to get caught up to speed on the new versions. You guys move quicker than I can learn haha.

The problem I have is that we want to consolidate but with the current component count we have its near impossible.

Any chance you or anyone might possibly have suggestions on some reports I could run to maybe find out ways to reduce component count? One of our goals is that we want to make full use of our resources we assign to our servers. Meaning if our hardware can support 50 k devices we want to find ways to allow our poller to support fully support said number of devices. And if there is way to increase limits on other module we'd like that too.

I want to strike that harmony between resource and sw limitations without needing to go over kill. It's all part of preparation for future deployment. Design prep. figuring out load balancing, etc.

This way I can draw out some visio diagrams that can offer return of investment. A tight system that is forecasted for growth. Is optimized for top level performance, can handle our needs, and can provide the data we are looking. Because the team here is finding reasons to throw out solarwinds. And we don't want to do that. I know how powerful the software is when it works. And I want them to see it too.

So if you can guide me to relevant doc's along with the information here I can study up and see the best paths for me to take.

mesverrum Didn't forgot about you.... I thank you for the input.