The Software Vendor Touching a Customer's Production: Where Should the Evidence Live?
The evidence should live on the customer's own network under the customer's control, but the supplier must be able to reach it.
Read →
Blog
Writing on Database Governance
These articles were written to help teams facing the same problems, not to praise a product. Each one ends with a concrete checklist you can apply.
To the DBA Who Says "We Do Not Want Prevention": Is Control a Burden or an Alibi?
Evidence is the only thing that protects the DBA after an incident. The person with the most access is the easiest to blame.
Read →The Life of a File That Leaves Production: Where Is That Customer List Now?
Most organisations manage the moment of exit. Once the file has left, no system follows its life.
Read →A Million Row UPDATE Does Not Go Through the Pipeline: Whose Record Is a Bulk Data Fix?
The riskiest work is the least recorded work. The fix is not to switch on blocking but to record the decision and evidence while the DBA does the job.
Read →The Key in the Agent's Hand: A Coding Agent Deleted Production in 9 Seconds
"The AI made a mistake" is the wrong sentence. The right one: the key handed to the agent made the mistake. Three layers: enforce, route through a gate, monitor.
Read →Six Correlation Rules for Your SIEM: What DAM and Governance Catch Together
The rule is written on your own platform with your own risk appetite. The product supplies three things: the expected accounts report, the session tag and start and end events.
Read →No Privileged Accounts. So Who Got In at 03:00?
Removing standing privilege is easy; removing emergencies is impossible. The emergency path should be made provable rather than abolished.
Read →DAM Sees Everything but Does Not Know Which Access Was Approved
DAM sees everything but does not know what was approved. The process system knows what was approved but does not see the rest. The two gaps are two halves of the same story.
Read →They Had Access, They Had No Reason: Why Access Management Is Not Enough
Access answers one question: what can they do. Justification answers another: why did they do it. In most organisations the second is recorded nowhere.
Read →The DBA Who Pasted a Query Result Into an AI Assistant: Who Is Accountable?
There is no bad intent and no record either. No product can stop the paste, but it can reduce what is available to paste.
Read →Corporate Sign In: Where Should the Password, the Role and the Emergency Account Live?
Authenticating tells who someone is. It does not say what they may do.
Read →The Ticket Was Approved. Is the Script That Runs Tied to It?
A ticket approves the intent. The text that runs lives somewhere else.
Read →What Should a Product Send to the SOC?
A system's records must also exist somewhere the people who run that system cannot change.
Read →Is Your Installation Audit Ready?
A control written in a procedure but not enforced by the system is not a control, it is a wish.
Read →Whose Change Is Your Script About to Overwrite?
A script does not say change this line. It says this object is now the following. The object may have changed in between.
Read →How Does an Audit Finding Actually Close?
If a gate has never turned anyone away, there is no way to know whether it is locked.
Read →How Long Should an Audit Trail Be Kept?
An audit record is kept long because it is evidence; a query result file is kept briefly because it is risk.
Read →Which Controls Were Relaxed Before the Audit?
A rule enabled two weeks before the audit produces no evidence for a full year period.
Read →"We Have Backups" Does Not Answer the Audit Question
A backup is a photograph of data. It does not contain the decision taken before the photograph.
Read →Why Is the Audit Sample Always Twenty Items?
Twenty items means as many as can be reviewed in the time allotted. That is a budget decision.
Read →Log Management Does Not Produce Audit Evidence
A log platform tells you an event happened. It cannot tell you whether it was permitted.
Read →Where Should the Key That Signs Evidence Live?
If the key is stored in a table, the person who wants to alter the chain is already there.
Read →Who States the Date of a Record? RFC 3161 Timestamping
Someone who can change the server clock can produce evidence that passes every check and still carries a false date.
Read →External Digest: What Shows That a Record Is Old?
The digest that counts is one that has sat in a locked store since the day it was written.
Read →How Does an Auditor Verify an Evidence Package?
If the party making the claim is the party keeping the records, the claim is not evidence.
Read →Database Activity Monitoring Is Not Governance
A monitor is a camera, governance is a turnstile. Needing both does not mean either one is inadequate.
Read →How Does AI Written SQL Reach Production?
A developer writing an unqualified delete hesitates. An assistant producing the same statement does not.
Read →How Many Requests Were Rejected Last Quarter?
An approval step means something because it can say no. If it never does, it is not a gate but a counter.
Read →Real Data in Test Environments: The Quietest Compliance Risk
A customer asked for erasure. It was deleted from production. What about the copies in test?
Read →More Than Half of Outages Come From Changes
A script can be entirely correct and still dangerous enough to stop production.
Read →Who Did This? The Attribution Problem of Shared Accounts
Without its third property, what you hold is an event log rather than an audit trail.
Read →Why Your CI/CD Pipeline Does Not Fully Cover Database Change
A pipeline knows what it carried and stays silent about what it did not. That silence reads as nothing happened.
Read →Who Is Accountable When a Vendor Connects to Your Database?
An account left from a project that ended years ago is still open and still privileged. Nobody knows whose it is.
Read →For the CIO Who Says "We Have a Process": Seven Questions Your Stack Cannot Answer
A process may be working, but if it produces no evidence the audit outcome is the same.
Read →COBIT BAI06 and ITIL 4: Five Break Points at the Database Layer
The frameworks include it in scope; the design was built around application releases. That is where it breaks.
Read →Could We Just Build This Ourselves?
The question is not whether you can. It is whether this is your core business and whether you will maintain it for five years.
Read →When the Change Board Becomes a Bottleneck: The Moment Work Escapes
If a board has rejected nothing in months, the submissions are not flawless. The board has no time.
Read →What Is Database Change Management? An End to End Guide in Eight Steps
The eight steps of the process, who does what in each, and four maturity levels.
Read →Which Column Gets Masked? How the Masking Decision Is Made
Not every eleven digit number is an identity number. Does the name decide, or the content?
Read →Permission Changes: The Quietest Gap in an Audit
A permission granted too broadly breaks nothing; it only lets someone see what they should not.
Read →The Database Side of PCI DSS: Which Controls Bind You?
The application and network sides are usually mature. Friction tends to appear in database operations.
Read →ISO 27001 and the Database: What Four Controls Mean
The control exists but the evidence does not. What four Annex A controls mean at the database layer.
Read →The Database Side of a Regulatory IT Audit: Four Questions, Four Kinds of Evidence
What stalls an audit is not missing records but records that do not answer the question.
Read →Database Schema Drift: Why It Happens and How to See It
Why does a script that passes in test fail in production? Schema drift and how to detect it.
Read →Eight Paths Into Production: Why Governing the Pipeline Is Not Enough
The pipeline is built, the scope is not. The eight paths into production and how many are really governed.
Read →The Correlation Tax: Why Jira and Database Logs Are Not Governance
Four systems, four half answers. The invisible cost of assembling evidence by hand, and the single record model.
Read →Audit Ready in Minutes: The Sealed Evidence Dossier
Why a screenshot is weak evidence, and what a per request sealed file puts in its place.
Read →Was the SQL That Ran the SQL That Was Approved?
The ticket says what was asked, the log says what ran. What connects the two?
Read →Emergency Database Change: Fast but Controlled
An emergency change is not skipping approval. A separate policy, a locked step and a declaration threshold.
Read →How to Govern Production Data Requests
Seven steps to turn the request into a process, and how the masking decision is made.
Read →Preparing the Database Side for a KVKK Audit: 12 Evidence Items
The 12 evidence items, where each comes from, and what does not count as evidence.
Read →UPDATE Without WHERE: A One Line Disaster and Four Layers of Defence
Why it keeps repeating, and how to build four independent layers of defence.
Read →SQL Rollback Strategy: Three Databases, Three Different Realities
Implicit commit, transaction boundaries and how to build a realistic rollback plan.
Read →Risk Scoring for SQL: Why Averaging Is the Wrong Idea
The band approach, non skippable rules and an explainable decision.
Read →Segregation of Duties in the Database: From Policy to Behaviour
How SoD works in small teams, the emergency exception and what auditors check.
Read →Who Can Switch Off a Control? Auditing Settings Changes
A wrong request can be undone. What a relaxed control missed while it was off cannot be recovered.
Read →How Does an Audit Trail Become Tamper Evident?
The ledger, the digest outside, independent verification and the limits worth admitting.
Read →See these ideas working
If you want to apply these checklists in your own environment, let us start with a 30 minute demo.
Book a Demo →