THE GOLD STANDARD OF INFORMATION · SPECIALIST STUDY
Turn a deadlock error into a reproducible explanation of who waited for whom and why
A deadlock is a cycle of dependency, not simply a slow request. Two valid update paths can each hold something the other needs, making progress impossible until the repository aborts a participant. The useful response is not to raise every timeout or retry forever. It is to preserve the transaction boundary, reconstruct a wait-for graph, separate detection from ordinary waiting, design a consistent acquisition order and prove the remedy under controlled contention.
This guide belongs to the Information subject root, the Gold Standard of Education, and the How X Works hub. It begins with a concrete decision, then tests whether the method still holds when the context changes.
Choose a reading route
Define the transaction evidence · chapters 1, 2, 3
Reconstruct the dependency cycle · chapters 4, 5
Design and test prevention · chapters 6, 7, 8
Interpret production signals · chapters 9, 10, 11
Recover and govern · chapters 12, 13
The whole problem at a glance
| Question | Evidence | Unsafe shortcut |
|---|---|---|
| Was it a deadlock? | Native cause, transaction IDs and cyclic waits | Calling every timeout a deadlock |
| Who waited for whom? | Resource or predicate, holder, waiter and interval | Keeping only the victim error |
| Why this victim? | Repository policy and work already done | Assuming the newest always loses |
| Did the repair work? | Same contention schedule and invariant check | Testing one serial request |
The victim is an outcome of cycle resolution; the cycle is the mechanism to explain.
CHAPTER 1 OF 13 · Define the transaction evidence
1. Name the logical operation and transaction boundary
Give the user intent one stable operation identifier, while each attempt and repository transaction receives its own identifier. Record request sequence, active dataset, isolation mode, client, node, start, commit or abort decision and uncertain responses. Without this separation, a retry appears to be a new job and a single logical change can be counted as both failure and success.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Place a retrying gateway in front of the repository; the logical identifier must still join every invisible attempt. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 1 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 2 OF 13 · Define the transaction evidence
2. Separate a deadlock from waiting and expiry
A deadlock is a dependency cycle in which no member can progress; a lock wait can resolve when its holder commits, and a statement or client timeout can expire without any cycle. Preserve the repository cause code and detector decision. PostgreSQL, for example, waits for deadlock_timeout before running a comparatively expensive check, so the delay is not itself evidence of the cycle.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Shorten the client timeout below the detector interval; explain why clients can cancel before a deadlock is reported. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 2 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 3 OF 13 · Define the transaction evidence
3. Model protected resources truthfully
Map native locks or optimistic guards to the smallest truthful application unit: named graph, versioned entity, materialised invariant, predicate range or internal index resource. SPARQL text alone may not reveal the physical acquisition sequence. Keep the native resource witness under access controls, and retain the application meaning needed to choose a remedy.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Change from entity version guards to predicate-range validation; rebuild the resource map rather than reusing old lock labels. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 3 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 4 OF 13 · Reconstruct the dependency cycle
4. Build a wait-for graph from timed edges
For every blocked interval record waiter, holder, requested mode, held mode, resource fingerprint, acquisition time and release or abort. Draw a directed edge from waiter to holder and search for a cycle whose intervals overlap. A list of involved transactions without direction and time cannot distinguish a true cycle from a queue.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Introduce lock conversion from shared to exclusive; represent conversion edges instead of treating them as ordinary new locks. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 4 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 5 OF 13 · Reconstruct the dependency cycle
5. Preserve the victim and the survivors
Capture all cycle members, not only the aborted transaction. Record victim-selection information when documented, work already performed, side effects outside the transaction and which survivors committed. PostgreSQL explicitly warns that the aborted participant is difficult to predict; application safety therefore cannot depend on a favourite transaction always surviving.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Upgrade the repository and observe a different victim; correctness should remain unchanged even when wasted work moves. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 5 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
Worked wait-for graph
| Transaction | Holds | Requests | Dependency |
|---|---|---|---|
| A | Guard:account-7 | Guard:account-9 | A waits for B |
| B | Guard:account-9 | Guard:account-7 | B waits for A |
| Detector | Finds A → B → A | Aborts B | Cycle is broken |
| Retry | Acquires 7 then 9 | Revalidates and commits | Order is consistent |
The retry begins a fresh transaction and recomputes its SPARQL match.
CHAPTER 6 OF 13 · Design and test prevention
6. Design a consistent acquisition order
The strongest ordinary prevention is a shared order for acquiring multiple protected objects. Canonicalise graph or invariant keys before issuing updates, and ensure every code path—including repair jobs—uses the same comparison. An order that depends on arrival sequence or local labels can diverge across clients and reproduce the cycle.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Move from numeric IDs to internationalised strings; define one canonical comparison across clients and locales. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 6 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 7 OF 13 · Design and test prevention
7. Shorten protected work without weakening the invariant
Move parsing, remote calls and human interaction outside the transaction when semantics permit. Narrow graph patterns and precompute immutable inputs, but revalidate every decision-dependent read after the transaction starts. Shorter transactions reduce overlap; they do not replace serialisable protection of a multi-resource invariant.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Move a remote lookup outside the transaction; revalidate its version before using it in the committed change. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 7 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 8 OF 13 · Design and test prevention
8. Use bounded fresh retries after approved causes
Retry a deadlock victim only in a fresh transaction that repeats matching and validation against current state. Cap attempts and elapsed time, add jitter and make one layer own the budget. A blind replay of previously computed DELETE and INSERT data can commit an obsolete decision even though the cycle itself was resolved correctly.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Enable retries in the driver and service mesh; prove the combined attempt bound rather than multiplying hidden limits. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 8 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 9 OF 13 · Interpret production signals
9. Instrument production without exposing RDF data
Metrics need stable low-cardinality labels such as cause, operation class, repository version and protected-unit class. Put detailed resource fingerprints in access-controlled traces or exemplars, never raw queries, credentials or sensitive RDF values. Publish cycle size, victim work, wait duration, retry amplification and final logical outcome.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Add tenant-specific graph names; hash or map them under privacy rules instead of placing them in metric labels. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 9 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
Signal interpretation matrix
| Signal | Can support | Cannot prove alone |
|---|---|---|
| Deadlock abort | A cycle was detected | Which application path was wrong |
| Long lock wait | Contention or blocking | A cycle exists |
| Statement timeout | Work exceeded a limit | The blocker or victim policy |
| Retry success | A later attempt committed | The original history was harmless |
| Lower abort rate | The symptom changed | The invariant is protected |
Correctness, progress and latency need separate evidence.
CHAPTER 10 OF 13 · Interpret production signals
10. Interpret timeouts as censored observations
A transaction cancelled at 400 milliseconds tells us its wait exceeded that boundary, not how long it would have lasted or whether a cycle existed. Analyse detector reports, lock waits, statement limits, network deadlines and client abandonment separately. Changing a timeout alters what becomes observable and can make a dashboard improve while unfinished work merely disappears sooner.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Reduce the gateway deadline by half; rebuild the latency and abort denominators before comparing periods. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 10 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 11 OF 13 · Interpret production signals
11. Reproduce the contention schedule
Start from a minimal graph and two complete update paths. Use barriers to force the intended order: A holds resource 1, B holds 2, then both request the other. Preserve repository and client versions, isolation, plans, timestamps, outcomes and invariant queries. Run enough repetitions to detect nondeterministic acquisition and prove the repaired order cannot form the cycle.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Run on a multi-node service where scheduling differs; preserve barriers at the application boundary and report residual nondeterminism. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 11 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 12 OF 13 · Recover and govern
12. Test limits, fallbacks and overload
A single-writer queue can prevent cycles yet create starvation; coarse locking can protect correctness yet destroy throughput. Test ordinary and urgent traffic, large updates, cancellations, failover and detector load. Define what happens when wait-for telemetry is unavailable, and never weaken isolation automatically to meet an availability target.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Double the bulk-update size under peak load; verify bounded waiting and correctness, not only aggregate throughput. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 12 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
CHAPTER 13 OF 13 · Recover and govern
13. Operate a deadlock review and retirement loop
Review recurring cycles by operation pair and protected unit, rank wasted work and participant harm, assign a design owner and verify each fix with the same witness. Retire instrumentation only when its decision purpose ends; preserve stable cause mappings and representative tests through repository upgrades so an apparent trend is not merely a renamed error code.
Build the evidence record around the decision, not around the convenience of the software. Name the unit observed, the version or policy in force, the time boundary, the numerator and denominator, exclusions, missing cases and the person authorised to act. Preserve raw observations separately from interpretation. That separation lets a later reviewer distinguish a changed world from a changed judgement.
Convert the finding into a bounded action. Predefine the trigger, permitted response, owner, expiry, rollback and verification window. Keep one worked case that passes and one that fails. A system which stores only successful cases cannot demonstrate that its control detects the failure it was designed to contain.
Evidence discipline: record the observation, version or time boundary that supports this chapter’s conclusion. Preserve the rejected alternative as well as the selected action. That makes a later review able to distinguish a changed world from an undocumented change of mind.
Changed-context transfer: Change native deadlock codes during an upgrade; revalidate mappings before comparing historical rates. The method transfers only if the decision question and measurement meaning survive the change. If either changes, rebuild the comparison rather than copying the earlier threshold or weight.
Chapter 13 check · what should survive an audit
- State the decision and measurement unit in one sentence.
- Preserve version, timing, denominator, exclusions and missing-case evidence.
- Connect the finding to a bounded action, rollback and later verification.
Contents · Previous chapter · Return to the Information guide
Verification lab: work the decision twice
Force the A-holds-7, B-holds-9 schedule, preserve the cycle, let one victim abort, then retry by acquiring 7 before 9. Prove one logical operation has two attempts and only one final outcome.
Set a client deadline shorter than the repository deadlock detector. Explain which executions become client cancellations, which can still be detector victims and why one timeout counter cannot separate them.
Add a bulk repair path that acquires keys in descending order while interactive work uses ascending order. Predict the wait-for cycle, then redesign and test the shared canonical order.
Limits and misconceptions
A timeout is not proof of a deadlock
It is a bounded observation of unfinished work unless a detector records a cycle.
Victim retry does not repair the design
It restores progress but can waste work and recur indefinitely.
Lock order protects only covered paths
One repair job using a different order can recreate the cycle.
Lower deadlock counts do not prove correctness
Weaker isolation or early cancellation can hide the symptom.
Questions readers ask
What is a wait-for graph?
- A directed graph whose edge points from a waiting transaction to the transaction holding what it needs.
Should the newest transaction always be the victim?
- No. Victim policy is implementation-specific and must not be part of correctness.
Can SPARQL requests deadlock on disjoint triples?
- Yes, when internal indexes, graphs or shared invariants create dependencies not visible in a simple triple diff.
Should deadlock victims be retried?
- Only approved causes, in a fresh transaction, with re-evaluation and a bounded shared budget.
Why record survivors?
- They reveal the full cycle, final history and displaced work.
What proves a fix?
- A deterministic contention witness, invariant check and repeated changed-context test.
Practise, verify and continue
Build a two-client test with stable logical IDs, timed wait-for edges, one forced cycle, a fresh bounded retry, an invariant query and a dashboard separating waits, detector victims, timeouts and final outcomes.
Repeat through a gateway and driver that both retry, then change key collation and add a descending-order maintenance job. The prevention rule transfers only if acquisition order and total attempt budget remain explicit.
Practical evidence check
Before implementation, write a one-page evidence map with the decision, unit of analysis, version or time boundary, inputs, outcomes, uncertainty, authority and stop rule. Attach one worked case that passes and one that fails. A reviewer should be able to reproduce the reasoning without relying on the memory of the person who built the system.
Run one changed-context rehearsal before approval. Alter a mechanism rather than a cosmetic detail, then identify which calculation, safeguard or governance gate must change. Record any conclusion that does not transfer. This short rehearsal is often the fastest way to expose a hidden assumption before it becomes an operational dependency.
Related eduKate reading
SPARQL transaction abort telemetry and retry budgets
SPARQL Update isolation and anomaly witnesses
SPARQL replay safety and uncertain completion
Authoritative sources and scope
PostgreSQL explicit locking and deadlocks
PostgreSQL viewing locks
PostgreSQL lock-management configuration
W3C SPARQL 1.2 Update
The sources support SPARQL update boundaries and general lock, wait and deadlock behaviour. Repository mappings, cycle examples, thresholds and incidents are illustrative and must be verified against the deployed RDF store, driver, isolation and privacy controls.