Your critical application.
Consistently fast.
PSG MX is our retainer for a SQL Server application where performance is business-critical. We know its workload and its history, detect changes early and implement improvements with your team and the vendor. Month after month, with measured results.
- Goal: booking screen under 1 s95% of calls, previously 81%
- Index for month-end closeimplemented, runtime −62%
- Inventory procedureagreed with vendor, release 4.2
- Plan change since release 4.1newly detected, under investigation
Switch to MX-Care. Re-measure the effect of release 4.2 after deployment.
When performance decides the business.
PSG MX pays off for applications that revenue, core processes or many workstations depend on. And for teams that don't want to react only once users complain.
Dips without explanation
The application keeps slowing down, and every time the search starts from scratch.
Every release a risk
New versions of the application or of SQL Server change plans and runtimes. Nobody knows in advance exactly how.
Costs rise along with it
More cores, more memory, higher license costs, and the response times stay as they are.
No time for depth
Your team keeps operations running. There is no time left for a thorough analysis of workload and plans.
Only when it is on fire
Problems surface when users call or processes stop. Nobody saw them before.
Between operations and vendor
Is it the platform or the application? Without evidence, the question goes in circles.
Work through it first. Then keep it in view.
PSG MX starts with at least three months of MX-Intensive. After that, the retainer moves to MX-Care. If a release, a version change or a load peak comes up later, we plan another intensive phase of at least three months together in advance.
We work through your application in a targeted way.
We get to know the application and its workload and adapt our tools to your environment. Then we take considerably more time: the most expensive and most conspicuous queries, plans and indexes; agreeing changes with development and the vendor, implementing them and re-measuring. Plus knowledge transfer, so that your team understands the connections.
We keep your application in view.
The tools collect data continuously, we watch for changes and put anomalies in context. If a problem occurs, we investigate it with full knowledge of the application and its history instead of starting from zero. A short monthly status keeps you informed.
- JanIntensive
- FebIntensive
- MarIntensive
- AprCare
- MayCare
- JunCare
- JulCare
- AugCare
- SepIntensive
- OctIntensive
- NovIntensive
- DecCare
MX-Intensive to start, at least three monthsPlanned further intensive phase, at least three months, for example before a releaseMX-Careillustrative example
From goal to measured effect.
- 01
Set goals
Define together what “good” means for your application: which processes matter, which response times and runtimes, which costs.
- 02
Tailor and collect
During the intensive phase, we adapt our tools to your environment. After that, they record workload, plans and wait times continuously, so the history is there when it is needed.
- 03
Prioritize findings
What slows things down most, what threatens next? Each finding with evidence and an action that can be verified.
- 04
Implement and re-measure
Agree changes with your team and the vendor, deploy them with a way back, and measure the effect.
- 05
Pass on knowledge
Regular feedback to development and operations, during the intensive phase also as training on your cases.
finding Month-end close: a reporting query reads the entire bookings table. evidence Execution history since March; plan unfavorable and unchanged since release 4.1. action Filtered index on open bookings; query unchanged. approval Operations and vendor agreed, way back documented. effect Runtime in the close from 41 to 9 minutes, CPU load considerably lower. next Check again after release 4.2.
No dashboards. Answers.
A rising graph doesn't reveal its cause. PSG MX delivers what comes after the chart: an explanation, a decision and the proof that it worked.
- Goals and metrics
Agreed in writing, so that progress becomes visible. - A maintained list of findings
Prioritized, with evidence, action and status. - Implemented actions
Deployed with a way back and with measured effect. - Monthly status
What happened, what worked, what comes next. - Agreed contact path
For problems, straight to the engineers who know your application.
The tools are part of it. The data belongs to you.
With PSG MX, our own tools run in your environment. They are part of the retainer, not an additional product.
The analysis relies on traceable rules and statistical methods, local and deterministic. More about our tools →
- MX AIOps
Collects wait times, resources, jobs and events as a Windows service, without a SQL Server instance of its own. - PSG QX
Analyzes the execution history in the Query Store: which query, which plan, since when. - Data stays with you
Everything resides in your environment, with no synchronization to a public cloud. - Tailored to your environment
During the intensive phase, we adapt data collection and analyses to your application. - The history grows
The longer PSG MX runs, the better normal can be told apart from unusual.
Engineering for your application. Operations for your servers.
Both are ongoing agreements, but with different subjects. PSG MX makes a critical application faster and more stable. The PSG DBA Pool keeps your SQL Servers secure and well maintained. The two can be combined.
| PSG DBA Pool | PSG MX | |
|---|---|---|
| Focus | your SQL Server systems | a critical application and its workload |
| Guiding question | Are the servers running securely, well maintained and recoverable? | Is the application becoming faster and more stable, and does it stay that way? |
| Rhythm | every business day, plus your projects | monthly: at least three intensive months first, then ongoing |
| Typical work | System check, backup, patches, permissions, high availability | Queries, plans and indexes, actions agreed with development and vendor, measuring the effect |
| Result | a maintained, documented environment | implemented improvements with measured effect |
| Your contacts | IT operations and administration | application owners, development, vendor |
| Tiers | Bronze, Silver, Gold | MX-Intensive, then MX-Care |
Have a single question, no ongoing need? Then a performance analysis or a performance audit fits. That is often a good way into PSG MX.
Short answers.
What distinguishes PSG MX from monitoring?
Monitoring reports that something is unusual. With PSG MX, we find out why, implement the most effective change with you and measure whether it worked. We are happy to use existing data from your monitoring as well.
What happens if a problem occurs during MX-Care?
We investigate it with knowledge of your application and its history, within the agreed scope. If the scope is not enough, we talk beforehand about the additional effort or another intensive phase.
How do MX-Intensive and MX-Care alternate?
PSG MX starts with at least three months of MX-Intensive and then moves to MX-Care. We plan another intensive phase of at least three months together in advance, for example before a release or a version change. Switching freely from month to month is not part of the model.
What data do the tools collect?
Performance data such as wait times, resource consumption and the execution history of queries. Everything stays in your environment, with no synchronization to a public cloud.
When are you available?
Monday to Friday, 8 a.m. to 5 p.m. CET/CEST, public holidays as observed in Hamburg, remotely through an access that you provide. We set the contact path and priority in the contract.
Do we also need the DBA Pool?
Not necessarily. PSG MX looks after the application, not the routine operation of the servers. If you have nobody in-house for that, combine PSG MX with the PSG DBA Pool.
How is billing handled?
As a monthly retainer. We agree the scope of MX-Intensive and MX-Care months in advance, together with goals, contact path and priority.
Which application can't afford to be slow?
Tell us about your application and what slows it down today. We will suggest a starting point, often an analysis beforehand.