Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials

How do you manage your management process?

This is a very tough question to search, so please pardon me if this has been asked before.

Basically, if you are a dedicated Solarwinds Administrator for example in charge of monitoring nodes for the different business units in your organization, how do you manage the process? I am interested in how you become notified of changes, what the SLA is for turnaround, managing of maintenance modes, etc. Yes I know the answer could be "whatever worked best for you and your organization" but I am curious on some best practices others employ.

Thank you!

Find more posts tagged with

Accepted answers

All comments

dragoon231

For the most part it is really up to the management to tell you what they want and you to make it happen as the SolarWinds Admin. So when you add a server (node), we will mark it as a Tier 1 - 3. What defines the tier level are things such as how critical the system is, what the response time needs to be (do i need to call someone at 2 am about this or can it wait a few hours for morning staff).

Specifics such as drive alerts, or CPU you may want to set a general alerting, like, 90% warning, 95% critical. That would really depend on the app owner though, at what point does drive capacity mean something? If it is at 90% and alerting, but the app owner allows it to sit for 2 weeks, clearly it is not an issue. As the admin I try to drill that into people, I want to avoid as much as possible, alert fatigue. That is a dangerous place to be because people responsible for the nodes will miss important stuff.

Another challenge is explaining this to people that don't understand the fundamentals or the purpose of monitoring. Some people think the purpose of monitoring is to send an alert out for a problem. Honestly, this should be looked at as a last resort. At the end of the day, the purpose of monitoring is the organization's bottom line. Do you want the downtime to control the organization, or do you want the organization to control the downtime? What I mean by that is, do you want to know a drive is going bad, have the part, have a 30 minute downtime to replace it before it totally konks out? Or have it just go bad, have a down system for a day?

Sorry, I tend to ramble. I hope this helps and if you have any other questions, let me know.

john.ta

Yes, that did help a little bit...but maybe I did a poor job of explaining what I am looking for. If a department adds 20 servers, what is the process you follow to receive that information and the SLA before it is put into active monitoring for example. Do you receive that information via helpdesk ticket, via e-mail, via post-it note? I work for a large company with multiple groups of IT spread across different physical areas. I want to make sure I have a solid process of receiving and processing this information to ensure we don't get backlash for "well why weren't you monitoring it, we have had it in production for 4 months!"

m_roberts

This is a good question and one with more than just the one answer I am afraid. There are a number of factors that will influence how this is handled in Orion and therefore the business review needs to take place, but and I stress this, dont be led by the current business process, but use the importance of monitoring and managing the infrastructure as a driver to make it work for the business.

If you have a CMDB and the current process is that all infrastructure teams must put devices into that as part of the commissioning process, then this is ideal, as you can use this data source to align with Orion. From a simple comparison report method i.e. what objects exist in only one system (Orion and CMDB), is the operational state in the CMDB reflected in Orion e.g. use of Unmanage. Also be aware you can automate this, so activity in the CMDB is automagically applied to Orion or just data syncronisation to ensure data quality is good. The Orion API will allow you to do this and for us is one of our most interesting areas we work in.

At the very least you must have a central location where all devices in your organisation are registered, their operation state, owner, importance, alert method, commission date, service info etc. This source is then authoritative for everyone to reference and understand what is in place, what does it do etc. Again - sync this data to Orion!

At the end of the day, everyone in the organisation responsible for devices, servers, apps must have a process to follow to keep this central inventory up to date. You can even use Orion on its own for this, utilising the custom properties feature.

Mark Roberts

Prosperon - UK SolarWinds Partners

Installation | Consultancy | Training | Licenses

pratikmehta003

Well I guess there are lot of factors that come into picture when we talk about best practice. Depends on what resources you have, like any CMDB is there or not, how is that managed, is it updated on time, etc.. From an Solarwinds admin perspective, I believe you should get the requests into your queue if the there is any CMDB tool, this will also help in tracking your effort and also an artifact that through a defined process and approvals activity was performed.

For SLA, It again depends on what has been decided for your project and based on that it should be completed.

You can also do some testing of alerts and confirm with the concerned team and if they are fine then you can configure it. And if you want to compare how many are getting monitored or not, its simple comparison job which you will have to work with concerned teams and then at one shot configure for all.

dragoon231

Overall I would recommend setting a "commissioning and decommissioning" process set in place. That could be something as basic as a work order submitted with server name. The person who submits it needs to know what information you need and this could be used for multiple groups. So some basic information you may want in this could be:

Server name

IP address

virtual/physical

warranty information

application

monitoring type

backup retention policy

In the end the person responsible for the server needs to tell you want to monitor, hard drive, memory, all the way down to a service or application and how they want to be informed and how quickly. where i put monitoring type, you could put something like 'basic' and define it with like cpu usage 100% for 1 hour then alert, drive alert, warning 95% critical 98%, etc.

There is a lot to the process when putting it in place that will require growing pains (as it does for any new system). Hope this helps

john.ta

How do you receive this request though? Post-it note on desk, helpdesk ticket, this is really what I'm struggling with...deciding the best flow to these requests...