Capacity planning - where to start? Whose task is this? Working tools?

Question

Iron

Good afternoon Colleagues,
There is a task to deal with IT resources and, as a result, set up the capacity planning process. I have already done this, but it was in an organization where I worked for a long time and knew a lot, as a result, responsible for the main services were appointed, they wrote standard documents describing the service, its main characteristics, as well as the resources it consumes + built growth trends. Further, at least once every six months, the data was updated. The process worked because firstly, the owners of the services understood well what was where, well, they controlled the growth. Secondly, the services were built quite transparently and understandably.
Now I have a more difficult task. New organization and highly obfuscated services that often use the same resources. In addition, there are no people who would know all the details - on the one hand, there are developers who receive tasks from the business, but they do not always even know the names of the servers that serve their application. On the other hand, there are administrators, they have a list of servers / databases / moons, but there is no attachment to applications or developers. File resources create a separate problem, there are several shared folders where information is exchanged, who participates in it is not clear, all developers have access rights (due to the fact that they flow between applications). The virtualization system also creates problems, because on the same moon there are quite a lot of virtual machines for different tasks, it is not clear who should be considered the owner.
Previously, they conducted audits, threw a common cry, everyone quit their jobs and in a week or two they generated several tables in Excel with more or less relevant information, which quickly ceased to be relevant. I don't want to repeat this mistake.
In total, the questions are:
1. What software is used for capacity planning? CMDB is still quite static for this, Excel is certainly universal, but somehow it is not an enterprise solution. I want something with the ability to integrate into the monitoring system (there is such a task too).
2. How to parse shared resources? So far, only one idea is migration to a new place, but this will take years, unfortunately.
3. Who should be responsible for shared resources? Who uses or provides them?
4. If there is no OLA / SLA between departments, how to oblige to turn on the brain when requesting and using resources? Now no one thinks when they upload the same amount into a 100GB database (although the database was created a year ago and they asked for 100GB moons, because then it weighed 10GB).
Thank you.
PS If I chose the hubs incorrectly, please correct me, but in general I went through them for a long time and did not find the ITIL option (but I saw 999 hubs for individual programming languages).

Answer 1

The task is solvable and in the presence of FROM even almost free of charge.
The first thing to do is to write out all the noticed services, both for business (SLA) and interconnected (OLA). So it will be possible to identify and process consumers in the future.
Then, in front of each item (service), list the known resources that provide it. Everything is also simple, go from iron up to the OS and the final butt. It’s too early to take into account the network or life support here, because. in the end, they are still divided into all, and in order to divide correctly, you need to separately measure the loads, the contours of the rotation of information, classify it, in general, mess around ...
By the way, these two steps are convenient to do in any mindmap program.
At the output, we get the prototype of the R-C model, but it is still uncontrollable. There are actually no methods here: all the resources that create the service are assets, you will have to create and keep records of them. What can be discovered is to choose a software that can be discovered and regularly detect (analogous to an accounting inventory), which is not automatically caught, but it is important to see - to take impressions with your hands.
Even an automatically assembled configuration - I'm sure - most of the resources for most of the services from the catalog will significantly unload and give time to parse hard-to-reach resources.
Carry out the initial filling of the CMDB, link what you found with the catalog in the OT and voila: it remains to link the consumers and the picture is almost ready.
Capacity is not maintained without taking into account configurations, this was obvious from the second ITIL.
As far as resources are concerned, it is not clear what they are talking about. File? Here, if you have money, you can buy good tools, the key phrase is data governance, but in general, methodically, you now see the so-called. unstructured data, and the goal is to make them structured. This is a big project, no less than capacity, but almost identical in approach: count, classify, identify consumers, log usage and then, after collecting the actual data, rebuild the folder structure / ACL and the rules for working with them.
On the last question: I don’t advise you to get into the topic of internal SLA / OLA (there are also underpinning contracts, UC) at your current moment in time. Accounting should get used to, you should understand that you can catch any unauthorized configuration changes or resource usage and how transparently and quickly.
As soon as the changes can be tracked at least within the next 24 hours, incitement can begin: inform the consumer of the service and the culprit that the load has come to the service and its allowable capacity is almost exhausted. They themselves will figure it out several times, and then get tired and ask for a single order. That's when it will be possible to paint all these agriments.
Something like this…

Answer 2

1-2) There is a set of Service Catalog Management applications. I will not advertise, but they are well sought after.
This process describes the hierarchical structure of services, from the finished product to the hardware level, and allows you to keep track of all dependencies and resources.
3) What does it mean to answer? There is an owner of the resource, there are users, each with their own rights and obligations, which are regulated by the business process.
4) Improve communication between departments.

Answer 3

1-2) Difficult to explain within a commentary.
Department A provides services within the company. Let's say it organizes block storage. This service, in turn, is provided by some material resources (servers, storage, cookies, etc.) and other services (electricity, ventilation to some extent, support from the vendor, personnel, security)
. So this process describes all the resources necessary for rendering services
What it gives:
1) The cost of the service. It is always clear how much this or that service costs towards the consumer.
2) Accounting for resources. You can see how many services are provided, and how many resources are used.
What size organization are we talking about? How much equipment?

Answer 4

1. Conduct a total audit of the system. Long, expensive, dreary, but every component at every level will be known.
Start documenting - yes, at least in 5 minutes of a deployed mediawiki.
2. when the dependencies are drawn in the same Visio, it will already be possible to understand
3. what, how and in what sequence to monitor, track and whom to kick if anything.