Skip to content
Compliance Automation Tools

Home / Guides / Continuous monitoring means testing controls on a set schedule and alerting on drift, not watching a screen

Guide, read 13 Nov 2026

Continuous monitoring means testing controls on a set schedule and alerting on drift, not watching a screen

Framework Coverage MatrixAll guides

NIST defines continuous monitoring as ongoing awareness, set by risk

NIST SP 800-137, published in September 2011, defines information security continuous monitoring (ISCM) as maintaining ongoing awareness of information security, vulnerabilities, and threats to support organizational risk management decisions. It does not define it as real-time. The control that implements the idea in NIST SP 800-53 is CA-7, Continuous Monitoring. Its guidance in the Rev 5.2.0 catalog says the terms continuous and ongoing imply that organizations assess and monitor their controls and risks at a frequency sufficient to support risk-based decisions, and that different types of controls may require different frequencies.

This matters when you buy or build. A product that checks every control every hour is not more continuous than one that checks a policy acknowledgment monthly and a cloud storage setting hourly. The test is whether each control has a stated frequency and a reason for it. See continuous control monitoring for the short definition.

What a monitoring program actually tests

SP 800-137 lists six steps: define an ISCM strategy, establish an ISCM program, implement it, analyze data and report findings, respond to findings, and review and update the strategy and program. The CA-7 statement in SP 800-53 lines up with them. It asks for metrics to be monitored, frequencies for monitoring and for assessment of control effectiveness, ongoing control assessments, correlation and analysis, response actions, and reporting status to named recipients.

In practice a tool tests three kinds of things. Configuration state, such as encryption on a storage bucket. Population membership, such as every active employee has a completed training record. And activity, such as privileged actions are logged. Process controls, like a board-approved policy, tend to change rarely. SP 800-137 notes that common controls that are process-related do not tend to be volatile and typically do not lend themselves well to automation. It also says all implemented controls must be regularly assessed even if monitoring cannot be automated. On automation, it says ISCM is most effective when automated, while recognizing that many aspects of an ISCM program are not easily automated.

That sets the honest scope for any product. Expect automated tests for configuration, identity and log sources, and expect manual attestations for training content, tabletop exercises and policy review. A tool that claims to monitor everything is describing its checkbox count, not a monitoring program.

What is testedExample checkTypical cadence logic
Configuration stateStorage encryption enabled on all bucketsHourly to daily, since it can change with one click
Population membershipEvery active user has completed trainingDaily to weekly, tied to joiner and leaver events
ActivityPrivileged use is logged (AC-6(9))Continuous log collection, reviewed on a schedule
Process artifactPolicy approved in the last 12 monthsMonthly check of the date, human review yearly

How often: the frequency is a decision you record

SP 800-137 gives the factors that set frequency. The list in its text is security control volatility, system categorization or impact level, threat information, risk assessment results, the output of monitoring strategy reviews, and reporting requirements. It gives a reporting example: if OMB policy requires quarterly reports on unauthorized components, the organization would monitor for them at least quarterly. It also says frequency may differ by individual requirement within one control, and that organizations may raise frequencies as automation is added.

Write down, for each control, the frequency, who set it, and which of those factors drove it. When two frameworks ask for different cadences, take the shorter one, as described in the multi-framework guide. The record of that decision is itself something an assessor can read.

SP 800-137 works across three tiers: the organization, the mission and business processes, and the information systems. For a shared control set the useful part is its remark on common controls. Because common controls are inherited by a large number of systems, their aggregate criticality may require more frequent assessment than a similar control protecting a single system. A control used to answer five frameworks deserves a tighter schedule for the same reason.

Drift is a passing control that stops passing between tests

Drift is the gap between the state you last proved and the state now. A control can pass at the audit sample date and fail three weeks later when someone disables a setting, a new system joins without a connector, or an employee changes role and keeps old access. Control drift is the term this site uses. Monitoring exists to shorten the time that gap stays open, and the number worth tracking is the time between a control failing and a person learning of it.

The size of the gap is easy to compute. A control checked weekly can be broken for up to 7 days before the next test sees it, and one checked quarterly for up to about 90 days. That is the cost of a lower frequency, stated in days. Drift has three common causes. A change to the monitored thing. A change to the population, such as new accounts. And a change to the test itself, such as a rule revised or a connector losing permissions and returning empty results. The third is the one teams miss: a test that returns no data can read as a pass. Count a test with no data as an error, not a pass.

What an alert should contain

An alert that says only control failed gets ignored. Include the control ID and plain name, the asset or account, the value found and the value expected, the last time the control passed, the time of this failure, the frameworks affected, the named owner, and the due date for a response. SP 800-137 describes the response step as responding to findings with technical, management, and operational mitigating activities or acceptance, transference/sharing, or avoidance/rejection. So an alert should leave room for that decision: fix, accept, transfer or avoid, with the choice and the person recorded.

Add a second tier for review, since a test that fails once and recovers inside an hour is different from one failing for nine days. Keep the alert count in range. If a team gets more alerts than it can close, set a stated sampling or severity rule and write it down. NIST says organizations should adjust frequencies as resource availability changes, which gives you the cover to do it openly. Then track the remediation time per alert, because that number shows whether the program works.

Where this shows up in the index

On continuous monitoring and integrations, which carries 20 of 100 points, Sprinto scores highest at 9.5 of 10 and Scytale scores 8.5.

Sources

  1. nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-137.pdf, read 12 Nov 2026
  2. raw.githubusercontent.com/usnistgov/oscal-content/main/nist.gov/SP800-53/rev5/json/NIST_SP-800-53_rev5_catalog.json, read 12 Nov 2026
  3. nvlpubs.nist.gov/nistpubs/CSWP/NIST.CSWP.29.pdf, read 12 Nov 2026

Last reviewed 13 Nov 2026.