Jani K. Savolainen, Founder & CTO, SQL Governor & DB Pro Services Oy.
In SQL Server environments, costs do not usually arise solely from having too many servers. An equally significant problem is over-provisioning.
In SQL Server environments, costs do not usually arise solely from having too many servers. An equally significant problem is over-provisioning. CPU cores, memory and storage resources are provisioned in greater quantities than the actual workload requires, simply to be on the safe side. This is directly reflected in infrastructure costs and, in particular, SQL Server licensing costs. On the other hand, under-provisioning jeopardises the service level and availability of the production environment, although this phenomenon is not quite as common as over-provisioning.
SQL Governor tackles the problem by combining proactive SQL Server monitoring, performance analysis and capacity optimisation into a single solution. Long-term measurement data enables capacity decisions to be made based on actual workloads, peak loads and trends, rather than on the basis of individual, momentary measurements.
A key differentiating factor is that SQL Governor’s methods relating to capacity planning, consolidation and predictive alerts for workload profiles are internationally patented. This illustrates that this is not merely a matter of visualising traditional monitoring data, but rather a technology based on advanced analytics and forecasting methods.
Right-sizing and consolidation without guesswork
One of the most significant sources of cost savings is the right-sizing of the SQL Server environment. SQL Governor analyses CPU, memory and storage capacity requirements at server, instance and database level.
In capacity planning, the system utilises historical time-series data, performance benchmark data derived from Spec.org, and advanced analytical methods to determine the optimal configuration for the target environment. At the same time, it is possible to model consolidation – that is, which workloads can be placed on the same servers or instances without compromising capacity limits or service levels in the short or long term.
It is precisely in this context that SQL Governor’s capacity and consolidation analyses, based on patented methods, which aim to make workload placement and environment sizing as cost-effective, systematic and data-driven as possible, whilst taking into account business requirements and any technical constraints.
From a DBA’s perspective, the benefit is clear. For example, the decision to replace a 32-core server with a 16-core server is not based on average CPU load or an Excel estimate, but on numerous different variables, such as performance comparisons between different processor types, the long-term behaviour of workloads and actual load peaks, as well as a systematic assessment of potential technical constraints and opportunities in relation to the planned topology.
For the IT service manager, this translates into euros. Fewer servers, CPU cores and unnecessarily reserved capacity mean lower infrastructure and licensing costs, without compromising on platform availability and performance.
From monitoring to predictive monitoring
Traditional monitoring provides an effective way of finding out what is happening in the environment right now. In cost-effective operations, however, it is also important to know what is going to happen next.
SQL Governor establishes baseline levels for workloads and utilises capacity forecasts, predictive alerts and anomaly detection. This means, for example, that an impending capacity limit for the CPU, disk space, database query execution plans or any other resource can be identified before it causes a disruption visible to users.
Predictive alerting mechanisms are also part of SQL Governor’s internationally patented methodology. The aim is not merely to react to an anomaly that has already occurred, but to identify the phenomenon and its trajectory early enough for the DBA or service team to take action before service levels deteriorate.
This also changes the capacity procurement model. There is no need to increase resources ‘just to be on the safe side’; instead, investments can be timed in line with actual forecast demand.
Predictive performance optimisation also reduces the need for capacity
Not all capacity problems can be solved by increasing capacity.
If the CPU load is caused by a poorly designed SQL statement, a missing index or a changed execution plan, purchasing a new server or additional processors will merely mask the actual problem. SQL Governor combines capacity data with SQL statement-level analytics, the plan cache, wait statistics and performance diagnostics, enabling a distinction to be made between actual capacity requirements and performance issues that can be optimised.
At their best, performance and capacity optimisation create a virtuous circle. More efficient SQL workloads require less computing capacity, which in turn enables more aggressive right-sizing and consolidation.
Capacity should be managed as an ongoing process
SQL Server FinOps is not a one-off consolidation project. Workloads, business requirements and infrastructure are constantly changing.
Therefore, the most effective approach combines continuous monitoring, forecasting, performance optimisation and regular right-sizing. SQL Governor’s patented capacity, consolidation and forecasting methods are designed to bring a systematic approach and automated analytics to this process. This gives the DBA technical visibility into the behaviour of the environment and enables the IT service manager to justify capacity and licensing decisions using measured data.
The bottom line is simple. In an SQL Server environment, you do not have to pay for capacity that is not being used, nor do you need to purchase capacity to solve a problem that could be resolved through optimisation.
