Kernfunctie van het platform
Zie hoe uw software
verbonden is.
Zof brengt uw diensten, API’s en afhankelijkheden in kaart, zodat uw team weet wat het moet testen wanneer er iets verandert.
- Verificatiebewijs en controle
- Menselijke goedkeuring waar beleid dat vereist
- Works with GitHub
Tijdgebonden
wat bestaat, wat veranderde, wanneer en door wie
Impactgebied
bepaald voordat de actie wordt uitgevoerd
Beleidsregels
gekoppeld aan de systemen die ze beschermen
Afgebakend
tot goedgekeurde systemen en toegang
Entities, relationships, and time
Services, endpoints, datastores, afhankelijkheden en eigenaren uit uw repositories.
- 01
Relationships that carry consequence
DEPENDS_ON, CALLS, READS, WRITES, DEPLOYS_TO, OWNED_BY, PROTECTED_BY, VERIFIED_BY, AFFECTS, GENERATED_BY, APPROVED_BY, USES_MODEL, USES_TOOL, SATISFIES, and VIOLATES.
- 02
Temporal by design
What exists, what changed, when it changed, who or what changed it, what depended on it, what evidence existed, what policies applied, what decision was made, and what happened afterward.
- 03
Blast radius before action
Every proposed change resolves to affected systems, criticality tier, applicable controls, historical incidents, and the verification required.
- 04
Shared by the whole loop
Verification, evidence, and control all read from the same graph, so every decision is grounded in one understanding of the enterprise.
Resultaten die teams meten
Size the blast radius before shipping
See exactly which systems, data, and customers a change can affect, and how critical they are, before it is allowed to run.
Tier 0
criticality resolvedRequire the right verification
Controls mapped to affected systems determine the evidence a change must produce, so verification is targeted rather than exhaustive.
Policy
selects the evidenceReconstruct any decision
Walk back from an incident to the change, the evidence that existed, the policies that applied, and who approved it.
Temporal
history stays queryableGive leadership one risk view
One graph shows where risk concentrates, where controls apply, and what changed recently.
One
source of truth
The substrate of the control loop
The System Graph is not a standalone feature. Verification reads it to decide what must be proven, Evidence writes to it, and Control relies on it to size risk.
Verification Engine
Selects the evidence each change requires from the systems and controls the graph resolves.
Evidence
Every bundle, approval, and outcome is written back to the graph, so history is queryable.
Control Plane
Policy decisions are grounded in criticality, affected systems, and incident history from the graph.
Beheerste validatie met menselijke goedkeuring bij elke kritieke stap.
- Continue verificatie onder uw besturingslaag
- System Graph brengt afhankelijkheden in kaart voordat wijzigingen worden uitgerold
- Auditwaardig bewijs voor nalevingsbeoordelingen
Detailed guide · English
In practice
Turn dependency context into a testing decision
The useful output of a graph is an explanation of why a change deserves attention. Start with a specific revision and follow the relationships that connect it to behavior your customers depend on. Treat missing relationships as an investigation task.
01
Establish the starting point
Identify the repository, base revision and proposed revision. Include configuration and dependency changes in the review scope, even when the application code diff is small.
02
Follow the consequential paths
Compare the reported affected areas with what the service owner expects. Ask which shared libraries, API consumers and background jobs could experience different behavior. Connected data sources determine which of these relationships are available.
03
Choose checks and retain the rationale
Use the dependency paths to select tests and reviewers. Record why each critical path was included. Keep a broader regression suite when unresolved dependencies make a narrower selection unreliable.
Illustrative scenario · not a customer result
A shared authentication library
A token parser changes in a library used by a web application and a background worker. A graph can help identify consumers represented in its inputs. Test valid, expired and malformed tokens in those consumers. If the worker relationship is absent, add it to the investigation rather than treating it as unaffected.
Questions to resolve before you begin
Is this a complete map of production?
Coverage depends on the connected sources and the relationships they describe. Static imports do not reveal every runtime call. Review source freshness and unresolved edges before relying on the map.
Does the graph replace observability?
Dependency context helps decide where to investigate and test. Production telemetry answers what is happening at runtime. Use both when assessing a release or investigating an incident.
Volgende stap
Bekijk uw eigen System Graph
Boek een Reliability Sprint en Zof begint één systeem in kaart te brengen waarvan u zich geen uitval kunt veroorloven, met identificatie van kritieke paden en blootstelling aan risico’s.
- Toegang en bereik beoordelen
- Eén kritiek systeem evalueren
- Opties voor private deployment
Gerelateerd
Ontdek hoe de System Graph andere platformfuncties aandrijft
Weet wat er is veranderd. Weet wat u moet testen.
Topologiebewuste impactanalyse koppelt servicewijzigingen aan getroffen workflows en beveelt validatieplannen aan op basis van systeemafhankelijkheden.
Gebruik systeemcontext om validatie te prioriteren.
