Configuration Drift¶

Version: v1 · 2026-08-24 · History
Prime #: 728
Origin domain: Engineering & Design
Subdomain: systems administration → Engineering & Design
Aliases: Config Drift

Core Idea¶

Configuration drift is the silent divergence between a system's intended state — specified, documented, recorded — and its actual state — what is really running — accumulating over time through ad-hoc changes that bypass the system of record. Each local change is small and locally justified: "temporarily" raise a limit, "just for now" disable a check, hand-edit a configuration to put out a production fire. The intended-state record is not updated to reflect the change. Over many such changes the divergence accumulates until nodes that should be identical are not, a deployment that should be reproducible is not, and a record that should describe reality describes only the past. The map (specified state) and the territory (running state) part company, and the longer the drift goes undetected the more expensive the reconciliation.

The arrangement carries a definite set of roles. There is a system of record describing the intended state. There is an operational state representing what is actually running. There is a change pressure on the operational state — incidents, workarounds, ad-hoc fixes. There is an asymmetry in how easily the two states accept change: operational change is cheap and immediate, while record update is expensive or simply forgotten. There is an uncoupled drift between the two over time, monotonic in the absence of reconciliation. And there is an eventual forced reconciliation — an audit, a failure, a transition, a legal challenge — that exposes the gap, followed by a two-phase response: reconcile (decide which divergences stand) and prevent (constrain future divergence).

The decisive structural force is unbounded one-way divergence under absence of coupling. If the specified and running states are not explicitly coupled by ongoing reconciliation work, they diverge, because change pressure on the running state is high while change pressure on the record is low. Drift is therefore the default trajectory, not an exception.

How would you explain it like I'm…

Map Stops Matching The Room

Imagine your toy box has a list taped on it of exactly what's inside. Over time you swap toys in and out but never fix the list. Bit by bit the list and the real box stop matching — and one day, when you actually need the list to be right, it isn't.

Plan Versus Reality

Configuration drift is when the written-down plan for how a system should be set up slowly stops matching how it's ACTUALLY set up. Each little change seems harmless: 'just for now, turn this off,' 'quickly fix this by hand.' But nobody updates the official record. Over many such changes, the gap grows: machines that should be identical aren't, and the plan describes the past, not reality. The reason it always drifts is uneven effort — changing the real system is quick and easy, but updating the record is slow and easy to forget. So unless someone keeps deliberately matching them up, drift is the DEFAULT, not the exception.

Map-Territory Divergence

Configuration drift is the silent divergence between a system's intended state — specified, documented, recorded — and its actual state — what is really running — accumulating over time through ad-hoc changes that bypass the system of record. Each local change is small and locally justified ('temporarily' raise a limit, 'just for now' disable a check), and the record is never updated to match. Over many changes the divergence accumulates until supposedly-identical nodes aren't, a reproducible deployment isn't, and the record describes only the past — the map and the territory part company. The decisive force is unbounded one-way divergence under absence of coupling: change pressure on the running state is HIGH (cheap, immediate) while change pressure on the record is LOW (expensive or forgotten), so without ongoing reconciliation work the two inevitably diverge. Drift is therefore the default trajectory, not an exception, and it ends in a forced reconciliation — an audit, a failure, a transition — that exposes the gap.

Configuration drift is the silent divergence between a system's intended state — specified, documented, recorded — and its actual state — what is really running — accumulating over time through ad-hoc changes that bypass the system of record. Each local change is small and locally justified: 'temporarily' raise a limit, 'just for now' disable a check, hand-edit a config to put out a production fire. The intended-state record is not updated to reflect it. Over many such changes the divergence accumulates until nodes that should be identical are not, a deployment that should be reproducible is not, and a record that should describe reality describes only the past — map and territory part company, and the longer it goes undetected the costlier the reconciliation. The arrangement carries definite roles: a system of record (intended state), an operational state (what is actually running), a change pressure on the operational state (incidents, workarounds), an asymmetry in how easily the two accept change (operational change cheap and immediate, record update expensive or forgotten), an uncoupled drift between them over time that is monotonic absent reconciliation, and an eventual forced reconciliation (audit, failure, transition, legal challenge) followed by a two-phase response: reconcile (decide which divergences stand) and prevent (constrain future divergence). The decisive structural force is unbounded one-way divergence under absence of coupling — without explicit ongoing reconciliation, high change pressure on the running state and low pressure on the record guarantee divergence, making drift the default trajectory rather than an exception.

Structural Signature¶

the system of record describing the intended state — the operational state that is actually running — the change pressure on the operational side — the asymmetry in how easily each state accepts change — the uncoupled monotonic divergence absent reconciliation — the eventual forced reconciliation exposing the gap

A system exhibits this pattern when each of the following holds:

A system of record. A specified, documented intended state against which the running state could in principle be compared.
An operational state. What is actually running, distinct from and potentially divergent from the record.
A change pressure. Ongoing demand for change on the operational side — incidents, workarounds, ad-hoc fixes — each small and locally justified.
An asymmetry of update cost. Operational change is cheap and immediate while record update is expensive or simply forgotten, so out-of-band changes accumulate unrecorded.
Uncoupled monotonic divergence. Absent explicit ongoing reconciliation work coupling the two states, they part company in one direction; drift is the default trajectory, not an exception.
A forced reconciliation. An audit, failure, transition, or legal challenge eventually exposes the gap, triggering a two-phase response — reconcile (decide which divergences stand) and prevent (constrain future divergence).

These compose so the decisive force is unbounded one-way divergence under absence of coupling, and the intervention space sorts into three families: eliminate the distinction via immutability, reconcile continuously, or periodically forgive.

What It Is Not¶

Not fading. fading is the decay or weakening of a signal, memory, or trace toward nothing; configuration drift is the divergence between two states (record vs. reality) where neither necessarily decays — the running state may be improving while the record goes stale. Fading loses a thing; drift uncouples two things.
Not gradual deterioration. gradual_deterioration is a single thing degrading; configuration drift is two states parting company under asymmetric change pressure. The running state need not deteriorate at all — it is the gap that grows.
Not maintenance. maintenance is the reconciliation work itself; configuration drift is the failure mode that appears when that work is omitted. Drift is the disease; maintenance is the (absent) cure.
Not a tipping point. tipping_points_or_phase_transitions is a sudden regime change at a threshold; configuration drift is monotonic, gradual divergence — the only sudden event is the forced reconciliation that exposes a gap built slowly.
Not refinement. refinement is deliberate, recorded improvement of a spec; configuration drift is undocumented, out-of-band change that bypasses the record. Refinement updates the map; drift changes the territory without it.
Common misclassification. Diagnosing drift where there never was a system of record, then trying to "reconcile" against a spec that does not exist. Catch it by confirming an authoritative intended-state record exists — if practice was always informal, the gap is emergent practice, not drift.

Broad Use¶

The pattern recurs wherever a system carries a record of an intended state against which the running state could in principle be compared. In systems administration — the origin — server fleets accumulate hand-applied package upgrades and divergent environment variables until the infrastructure-as-code no longer describes the actual cloud resources, with the "snowflake server" as the endpoint.^[1] In manufacturing, design drawings diverge from as-built reality through running engineering changes, supplier substitutions, and undocumented field fixes, reconciled only by as-built documentation.^[2] In regulatory compliance, the policy manual says one thing while day-to-day operation does another, the drift never reflected in the manual. In healthcare, written protocols diverge from what clinicians actually do as field workarounds for protocol gaps accumulate silently. In biology, somatic mutations and epigenetic changes cause a cell lineage to diverge from the genome of record while immune recognition continues to assume the original.^[3] In cartography and land registry, survey records diverge from fence lines and actual occupation, producing border disputes generations later.^[4] In project management, the plan diverges from delivered work through scope and staffing changes until the plan is fiction. In law, a contract diverges from the parties' course of dealing, which in many jurisdictions eventually prevails.^[5]

Clarity¶

Configuration drift replaces the vague complaint "the docs are out of date" with a recognisable structural dynamic: there is a single source of truth (intended state) and an operational reality (running state), coupled only by explicit reconciliation work, and without that work they drift monotonically apart. The reframing makes visible that drift is the default trajectory rather than an exception — that the question is not "how did this happen?" but "what reconciliation work was supposed to be holding the two states together, and why was it absent?"

The frame also forces a clean three-way separation that casual usage fuses: the system of record (what should be), the operational state (what is), and the reconciliation process (how the two are kept in sync). Once those are named, the interventions become legible rather than improvised, because each targets a distinct part of the structure — constrain the operational side, instrument the gap, or update the record. The clarifying move is to stop treating drift as an accumulation of unrelated mistakes and start treating it as the predictable output of an uncoupled pair of states under asymmetric change pressure.

Manages Complexity¶

The pattern compresses a sprawling family of "reality stopped matching the documentation" problems into a single structure with three named parts and one governing force, so that an analyst confronting a novel instance does not have to model the full system but only locate the record, the operational state, and the reconciliation channel, then ask whether change flows only through the record or whether out-of-band changes accumulate. The diffuse worry "things have gotten out of sync" becomes a determinate question about the coupling between two states.

The compression also organises the intervention space into three mutually exclusive families. Eliminate the operational/state distinction by making the system immutable from the operational side — immutable infrastructure, read-only filesystems, locked-down runbooks — so change must go through the record. Continuously reconcile via scheduled diffs, drift detectors, and compliance scanners that flag and revert or escalate any divergence.^[6] Or periodically forgive — accept that operational reality has diverged, capture the as-built state, retire the now-fictional record, and adopt the new state as the intended one. Naming the three lets the analyst choose deliberately rather than defaulting to the reflexive and usually wrong "just update the manual."

Abstract Reasoning¶

Configuration drift instantiates the broader map-versus-territory divergence pattern and connects it to several others. It connects to entropy: drift is the disordered direction in any system that carries a record, the state space of "record matches reality" being vastly smaller than the space of "record and reality differ." It connects to maintenance: reconciliation is preventive work, and drift is precisely the failure mode that appears when that work is omitted. And it connects to the single source of truth as a positive prescription whose absence permits drift.

The frame licenses a specific inferential discipline. For any system, identify the intended-state record, the operational reality, and the reconciliation channel, then ask whether change flows only through the record (good) or whether out-of-band changes accumulate (drift); when drift exists, ask whether reconciliation is cheap (audit plus update) or expensive (forced by failure). The decisive test for whether the prime applies is whether there is a specified state against which the running state could in principle be compared — if there is none, the system has no record and the relevant pattern is something else, such as emergent or informal practice. The inference is substrate-neutral: only the artefacts change, while the structure of two states under asymmetric change pressure is invariant.

Knowledge Transfer¶

Configuration drift transfers as a complete protocol because its roles map cleanly across substrates: the system of record maps to infrastructure-as-code, design drawings, the policy manual, the written protocol, the genome, the survey record, the project plan, or the contract; the operational state maps to the running fleet, the as-built assembly, actual practice, the cell lineage, occupied land, or the delivered work; and the reconciliation channel maps to drift detection, as-built documentation, compliance scanning, field surveys, or course-of-dealing recognition. Because the roles correspond, the three structural interventions — eliminate the distinction via immutability, reconcile continuously, or forgive periodically — are the same moves in every domain.

The documented transfers run cleanly. The two-phase response — first reconcile (decide which divergences stand and which revert), then prevent (constrain future divergence) — applies identically to an infrastructure team discovering its clusters have drifted from spec in dozens of distinct ways and to a hospital ICU discovering that its ventilator-weaning practice has diverged substantially from the written protocol: reconcile by documenting and either adopting or reversing the workarounds, prevent by training and electronic enforcement. The same logic applies to a regulatory agency that finds examiner practice has diverged from the published manual by years of accumulated custom; the configuration-drift-informed response is not the reflexive "update the manual" but to separate the manual (spec) from examination practice (reality), build a reconciliation channel (an annual field survey with formal incorporation or reversal), and decide deliberately which divergences are adopted and which reverted. Across these the structural force is the same — unbounded divergence under absence of reconciliation — and the intervention set is the same. The transfer carries an institutional flavour the bare structure does not: because the pattern presupposes a record-keeping institution and treats the specification as authoritative, it is partly framed rather than purely structural, and the normative weight ("the spec should win, or be deliberately retired") travels with it. But the load-bearing dynamic — two states under asymmetric change pressure diverging monotonically until a forced reconciliation — is what does the diagnostic work in every substrate, from server fleets to land registries to cell lineages.

Examples¶

Formal/abstract¶

A cloud server fleet managed by infrastructure-as-code is the origin instance and exposes every role. The system of record is the declarative configuration (a Terraform state or Ansible playbook) specifying the intended state of every node — package versions, environment variables, open ports.^[1] The operational state is what is actually running on the live instances. The change pressure is the steady stream of production incidents: at 3 a.m. an engineer SSHes into one node and hand-raises a connection limit "temporarily" to clear an outage, and never updates the playbook.^[7] The asymmetry of update cost is the engine — the operational change took ten seconds and worked immediately, while updating the record is a separate pull-request with review, easily skipped under pressure.^[8] Absent a reconciliation channel the divergence is monotonic and one-directional: each incident adds an unrecorded change, never subtracts one, so the state-space argument bites — the configuration space where record-matches-reality is vanishingly small relative to the space where they differ, and entropy carries the fleet toward the "snowflake server" that no one can reproduce. The forced reconciliation arrives as a failure: a node dies, is rebuilt from the (stale) record, and silently lacks the dozen hand-fixes, so the rebuilt node behaves differently and a new outage exposes the accumulated gap. The two-phase response follows the prime: reconcile (diff every node against spec, decide which of the out-of-band changes to fold into the record and which to revert) then prevent (move to immutable infrastructure so nodes are rebuilt from the record and never hand-edited, eliminating the operational/record distinction entirely).^[1]

Mapped back: the IaC declaration is the system of record, the running instances are the operational state, the 3 a.m. hotfix is the change pressure, the cheap-edit-versus-expensive-PR gap is the asymmetry, the accumulating snowflake state is the uncoupled monotonic divergence, and the failed rebuild is the forced reconciliation triggering reconcile-then-prevent.

Applied/industry¶

A hospital ICU's ventilator-weaning protocol runs the same dynamic in a clinical-governance substrate. The system of record is the written, committee-approved weaning protocol in the policy binder.^[9] The operational state is what bedside clinicians actually do — the real sequence of pressure-support reductions and spontaneous-breathing trials. The change pressure is the daily reality that the protocol has gaps and edge cases, so experienced nurses and respiratory therapists develop field workarounds for patients the protocol does not cleanly cover, each locally justified and clinically sound. The asymmetry is acute: adopting a better bedside practice is immediate and spreads by word of mouth, while amending the formal protocol requires reconvening a committee, so the record is updated rarely and the practice drifts ahead of it.^[10] Over a couple of years the divergence is large and silent — the binder describes a procedure the unit no longer follows. The forced reconciliation is an accreditation audit or an adverse-event review that compares charted practice against the protocol and finds them far apart.^[11] The prime prescribes the non-reflexive response: not "retrain everyone back to the binder" and not "just rewrite the binder," but the two-phase move — reconcile by documenting the actual practice and deciding, divergence by divergence, which workarounds are improvements to adopt into the protocol and which are unsafe drift to reverse; then prevent by building a standing reconciliation channel (a periodic practice-versus-protocol review with formal incorporation) and electronic order-set enforcement for the parts that must not drift. The same reconcile-then-prevent logic governs a regulatory agency whose examiner practice has diverged from its published manual by years of accumulated custom, and a manufacturer whose as-built assemblies have drifted from the design drawings through undocumented field substitutions.^[2]

Mapped back: the written protocol is the system of record, bedside practice is the operational state, protocol-gap workarounds are the change pressure, the committee-amendment cost is the update asymmetry, the silent two-year divergence is the uncoupled drift, and the accreditation audit is the forced reconciliation — the same structure operating across systems administration, clinical governance, and regulatory compliance.

Structural Tensions¶

T1 — Spec-Authoritative versus Practice-Authoritative (sign/normative). The prime's framed lineage treats the record as authoritative — "the spec should win" — but the two-phase response includes forgiving, where reality is right and the record is the fiction to retire. The failure mode is reflexively reverting every divergence to spec, reversing improvements that operational reality discovered. Diagnostic: for each divergence, ask whether the spec or the practice better serves the goal — drift is not automatically error, and some workarounds are upgrades to adopt.

T2 — Asymmetric Update Cost (scalar). The engine of drift is that operational change is cheap and immediate while record update is expensive or forgotten, so out-of-band changes accumulate one-directionally. The failure mode is blaming individual mistakes ("people keep forgetting to update the docs") rather than the cost asymmetry that makes forgetting rational under pressure. Diagnostic: compare the cost of the operational change to the cost of recording it — where the gap is large, drift is the default trajectory, not a discipline problem.

T3 — Monotonic Divergence versus Bounded Coupling (temporal). Absent explicit reconciliation work, the two states part company monotonically — drift is the default, not an exception. The tension is that the only thing holding them together is ongoing, easily-omitted reconciliation. The failure mode is assuming the states stay synced "for free" and asking "how did this happen?" instead of "what reconciliation was supposed to run and why was it absent?" Diagnostic: locate the reconciliation channel; if there is none, divergence is guaranteed and accumulating right now.

T4 — Three Intervention Families as Mutually Exclusive (scopal). The intervention space sorts into eliminate-via-immutability, reconcile-continuously, or forgive-periodically — and choosing reflexively defaults to the usually-wrong "just update the manual." The failure mode is mixing or mis-selecting: making a system immutable that genuinely needs operational flexibility, or continuously reverting drift that should have been forgiven. Diagnostic: decide deliberately which family fits the substrate's tolerance for operational change before acting — the reflexive update is rarely the right structural move.

T5 — Forced-Reconciliation Cost Grows with Delay (temporal). The longer drift goes undetected, the more expensive the eventual reconciliation, because the gap compounds and the forcing event (audit, failure, legal challenge) arrives at the worst time. The failure mode is deferring reconciliation as cheap-to-postpone when its cost is rising monotonically. Diagnostic: estimate how long the states have been uncoupled — a long quiet divergence is not stability but a growing reconciliation debt that a forced event will call in.

T6 — Specified State Exists versus Informal Practice (scopal). The prime applies only where a specified intended state exists against which the running state could in principle be compared; with no record, there is no drift, only emergent or informal practice, and a different pattern governs. The failure mode is diagnosing drift where there never was a system of record, then trying to "reconcile" against a spec that does not exist. Diagnostic: confirm there is an authoritative intended-state record — if practice was always informal, the gap is not drift and the reconcile/prevent machinery does not apply.

Structural–Framed Character¶

Configuration Drift sits in the mixed-framed band of the structural–framed spectrum — mixed-framed, aggregate 0.5. Every one of the five diagnostics reads exactly 0.5: there is a genuine relational skeleton underneath, but each diagnostic registers an equal, real pull toward the framed pole, and they sum to a prime that imports as much interpretive context as it recognizes. Conceding that frame honestly is the point of this section.

The relational skeleton is real and worth naming first: two states under asymmetric change pressure diverging monotonically until a forced reconciliation. That bare dynamic — uncoupled divergence in the absence of explicit coupling work — is what does the diagnostic labor in every substrate, and it connects to substrate-neutral patterns like entropy (the record-matches-reality region of state space is vanishingly small) and map-versus-territory divergence. But the prime cannot be told without the frame, and that is why all five diagnostics land at the midpoint rather than zero.

vocab_travels is 0.5 because the home lexicon — "system of record," "infrastructure-as-code," "snowflake server," "reconcile," "drift detection" — is systems-administration vocabulary that must travel and translate to reach manufacturing as-builts, clinical protocols, or land registries. evaluative_weight is 0.5 because the prime carries a normative load the structural primes lack: it presupposes that the specification is authoritative — "the spec should win, or be deliberately retired" — so divergence reads as a problem to be reconciled, not a neutral fact, even though the two-phase response does allow forgiving where practice is the better state. institutional_origin is 0.5 because the pattern's origin and operating assumption is a record-keeping institution: it requires a maintained system of record, an artifact produced by human or organizational practice, not a relation that exists in bare physics. human_practice_bound is 0.5 because, while there is a partial biological instance (somatic mutation diverging from the genome of record), the prime's core cases — server fleets, policy manuals, contracts, project plans — presuppose a human or institutional practice of specifying and recording an intended state, and the asymmetric update cost that drives drift is largely a fact about human processes (a cheap hotfix versus an expensive review PR). And import_vs_recognize is 0.5 because invoking the prime partly RECOGNIZES a real divergence dynamic already running and partly IMPORTS the whole record/reality/reconciliation framing, including the normative judgment about which state ought to govern. The honest reading is a prime whose structural core is genuine but whose every facet carries an inherited institutional, normative, and practice-bound frame — squarely mixed-framed, exactly as the aggregate of 0.5 records.

Substrate Independence¶

Configuration Drift is strongly substrate-independent — composite 4 / 5 on the substrate-independence scale. Its domain breadth is broad (4): the pattern of a documented intended state diverging from the running state through accumulated out-of-band changes recurs with the same structural force in systems administration (fleets diverging from infrastructure-as-code into "snowflake servers"), manufacturing (design drawings drifting from as-built reality via running engineering changes), compliance, healthcare, biology (somatic changes accumulating against the germline blueprint), cartography, project plans, and contracts. Its structural abstraction is high (4): the bare skeleton — a reference record of intended state, a running state, and an accumulating gap fed by undocumented changes the record never captured — is medium-neutral and stated identically whether the reference is a config file, a drawing, or a genome. The transfer evidence is concrete (4): the diagnostic and remedial moves — detect divergence by comparing record to reality, re-baseline or reconcile, close the out-of-band change path — carry across sysadmin reconciliation, manufacturing as-built audits, and compliance attestation. Its slight cap below the ceiling reflects that the strongest cases involve a maintained human record, though the somatic-mutation analogue shows the pattern also runs where the "record" is a biological template with no agent in the loop. Within its range the prime is recognized rather than translated wherever a system carries a record against which its running state could be compared.

Composite substrate independence — 4 / 5
Domain breadth — 4 / 5
Structural abstraction — 4 / 5
Transfer evidence — 4 / 5

Relationships to Other Abstractions¶

Current abstraction Configuration Drift Prime

Parents (1) — more general patterns this builds on

Configuration Drift presupposes, typical Traceability Prime

Drift is the divergence between a system-of-record (intended state) and the running state; it presupposes a maintained authoritative record against which reality could be compared — the traceability infrastructure whose absence-of-reconciliation it names.

Hierarchy paths (2) — routes to 2 parentless roots

Configuration Drift → Traceability → Observability

Show alternative path (1)

Neighborhood in Abstraction Space¶

Configuration Drift sits among the more crowded primes in the catalog (16^th percentile for distinctiveness): several abstractions describe nearly the same structure, so a description that fits it will tend to fit its neighbors too — transporting it usually means disambiguating within this family rather than landing on it exactly.

Family — Distributed Consistency & Shared State (11 primes)

Nearest neighbors

Computed from structural-signature embeddings · 2026-09-10

Not to Be Confused With¶

The embedding-nearest neighbor is fading, and the surface similarity is real — both describe something slipping away over time, and "the docs faded out of relevance" is a tempting gloss. But they are structurally different. Fading is the decay of a single thing: a signal, a memory, a trace weakening toward zero, where the content itself is being lost. Configuration drift is the divergence of two coupled states: a record and a running reality parting company, where neither need decay — in fact the running state is often improving (clinicians discovering better practice, engineers applying superior fixes) while only the correspondence between record and reality erodes. The decisive difference is that drift is about a gap between two things, not the diminution of one. This matters because the fading frame suggests "restore the thing that is fading" (re-document, re-record), whereas the drift frame insists on a three-way structure — record, reality, reconciliation — and a deliberate choice among reconciling, forgiving, or eliminating the distinction. Reading drift as fading collapses the running state and the record into one "thing," missing that the running state may be the more correct of the two.

A second genuine confusion is with gradual_deterioration. Both are slow, monotonic, and silent until something forces attention. But gradual deterioration is the degradation of a single system — it gets worse along some quality axis. Configuration drift is the uncoupling of two states, and crucially the operational state may not be deteriorating at all; it may be perfectly healthy and even superior to the spec. What grows is the divergence, driven by asymmetric change pressure, not the badness of either state. The tell is whether one thing is getting worse (deterioration) or two things are getting further apart while each may be fine (drift). The intervention differs accordingly: deterioration calls for repair or replacement of the degrading thing; drift calls for reconciliation work coupling the two states, plus a normative decision about which state should win.

A third confusion, and a more instructive one, is with maintenance. The two are intimately linked but are figure and ground of each other. Maintenance is the ongoing reconciliation work that holds the record and reality together — drift detection, as-built documentation, periodic review. Configuration drift is precisely the failure mode that appears when that maintenance is omitted. They are not rivals but complements: drift names the disease, maintenance names the cure, and the whole diagnostic payload of the configuration-drift frame is to make visible that the absence of maintenance work has a predictable, named consequence. The error the distinction guards against is treating drift as a series of unrelated mistakes ("people keep forgetting to update the docs") rather than as the structural output of missing reconciliation coupling — which reframes the fix from exhortation ("be more careful") to building a maintenance channel.

For a practitioner the cuts route to different responses. If a single thing is weakening toward nothing, that is fading — restore or refresh it. If one thing is degrading in quality, that is deterioration — repair or replace it. If the concern is the work that keeps states synced, that is maintenance — schedule and resource it. Configuration drift specifically names the gap between an authoritative record and a running reality under asymmetric change pressure, and its response is the three-family choice: reconcile continuously, forgive periodically, or eliminate the distinction via immutability.

Solution Archetypes¶

Solution archetypes in the catalog that build on this prime — directly (this prime is a source ingredient) or as a related prime.

Built directly on this prime (3)

Also a related prime in 5 archetypes

Dependency-Aware Change Notification: Warn the parties who actually depend on a changing system early enough, and specifically enough, that they can prepare before the change binds them.
Enacted-Control Verification and Closure: Verify controls as enacted, not merely as documented, and close the gap when paper controls and real operating practice diverge.
Latent Constraint Preservation Audit: Treat a persistent structure as possible evidence of a hidden constraint: understand its function, dependencies, and failure-prevention role before removing or simplifying it.
Managed Retreat: Withdraw or relocate an exposed subject into a viable receiving zone—and release or move blocking boundaries—before an advancing front closes the remaining corridor.
Moving-Target Tracking: Treat the objective as a time-varying reference and jointly tune target governance, sensing, prediction, planning, and response so cumulative tracking error remains bounded while the target moves.

References¶

[1] Morris, Kief. Infrastructure as Code: Managing Servers in the Cloud. Sebastopol, CA: O'Reilly Media, 2016. Defines configuration drift and the "snowflake server," and prescribes immutable infrastructure and continuous reconciliation against a declarative system of record. registry ↩a ↩b ↩c

[2] SAE International / Electronic Industries Alliance. EIA-649C: Configuration Management Standard. SAE International, 2019. The national consensus configuration-management standard covering configuration identification, change management, status accounting, and verification/audit — including as-built vs. as-designed reconciliation of products to their record after engineering changes and substitutions. registry ↩a ↩b

[3] Alberts, Bruce, Alexander Johnson, Julian Lewis, et al. Molecular Biology of the Cell, 4^th ed. New York: Garland Science, 2002. Somatic mutation and epigenetic change cause a cell lineage to diverge from the germline genome of record, with immune surveillance keyed to the original. registry ↩

[4] American Land Title Association and National Society of Professional Surveyors. Minimum Standard Detail Requirements for ALTA/NSPS Land Title Surveys. ALTA/NSPS, 2021. The land-title survey standard governing how recorded boundaries are reconciled against physical occupation (fence lines, lines of possession), the gap that produces adverse-possession and boundary disputes when survey record and occupied reality diverge. registry ↩

[5] American Law Institute & Uniform Law Commission. Uniform Commercial Code § 1-303 (Course of Performance, Course of Dealing, and Usage of Trade). Course of dealing/performance may supplement or qualify written contract terms, so practice diverging from the contract can eventually prevail (subsuming the former § 2-208). registry ↩

[6] Amazon Web Services. Detecting Unmanaged Configuration Changes to Stacks and Resources (CloudFormation Drift Detection). AWS Documentation. Drift-detection and continuous-compliance tooling that compares running cloud resources against a declarative system of record and flags divergences for reconciliation or remediation. registry ↩

[7] Allspaw, John, and Jesse Robbins (eds.). Web Operations: Keeping the Data on Time. Sebastopol, CA: O'Reilly Media, 2010. Operational accounts of ad-hoc production hotfixes and out-of-band changes that bypass the system of record under incident pressure. registry ↩

[8] Beyer, Betsy, Chris Jones, Jennifer Petoff, and Niall Richard Murphy (eds.). Site Reliability Engineering: How Google Runs Production Systems. Sebastopol, CA: O'Reilly Media, 2016. Documents change management and the review overhead of reconciled changes versus the speed of an ad-hoc operational fix (toil, configuration-as-code, progressive rollout) — the asymmetric cost that drives drift. registry ↩

[9] Ely, E. Wesley, Allan M. Baker, Daniel P. Dunagan, et al. "Effect on the Duration of Mechanical Ventilation of Identifying Patients Capable of Breathing Spontaneously." New England Journal of Medicine, vol. 335, no. 25 (1996): 1864–1869. Protocol-directed (vs. usual bedside) ventilator weaning in the ICU, establishing the written-protocol system of record against which actual weaning practice can diverge. registry ↩

[10] Grol, Richard, and Jeremy Grimshaw. "From Best Evidence to Best Practice: Effective Implementation of Change in Patients' Care." The Lancet, vol. 362, no. 9391 (2003): 1225–1230. Documents the gap between written clinical guidelines and routine bedside practice and why changing practice (and updating guidance) is slow and effortful — the asymmetric-update-cost driver of protocol/practice drift. registry ↩

[11] The Joint Commission. Sentinel Event Policy and Procedures. The Joint Commission. Accreditation and adverse-event (sentinel-event) review that compares actual charted practice against an organization's written policies/protocols via root-cause analysis, the forced reconciliation exposing protocol-vs-practice drift. registry ↩