Graphicient
Structure Explore Query Monitor
Graph history Environment fork Access control Skills library
01Forged parts, moving prices 02One part, several brands 03Hidden talent, revealed pool 04At-risk parts, anticipated failures 05Securing the supply chain
Blog FAQ About us Careers
FR EN
Contact us
Platform

FROM SCATTERED DATA TO RELIABLE DECISIONS

Structure, explore, monitor, decide: four connected building blocks on a single graph, turning scattered data into reliable decisions.

Company

A team that comes from the field

We build Graphicient with the teams who use it, from Toulouse.

Resources

Shedding light on graph technology and its applications.

What we learn in the field, and answers to the questions we hear most often.

Platform · Rollback

Your data should survive any mistake

November 9, 2026 Graphicient 6 min read
Your data should survive any mistake: Graphicient title on a black background, next to a blue data tunnel crossed by concentric circles and circuits

“The business model is versioned.” In our previous article on versioning, we explained how Graphicient keeps every version of your data without ever duplicating it.

That memory has a very concrete consequence: whatever has been kept can always be restored.

That is what we call rollback.

01

Mistakes are inevitable

In a living data system, the question is not whether an incident will happen, but when. Incidents usually come from three sources:

INCOMING DATAWrong or incompleteA badly exported file, an interrupted feed, a source that changes format without warning.
HANDLINGA manual mistakeA deletion by error, a change applied in the wrong place.
CHANGEUnexpected side effectsAn updated rule, a restructuring with unforeseen consequences.

In a traditional system, each of these mistakes overwrites the previous data. Someone then has to rebuild by hand, dig out backups and cross-check files. Meanwhile, dashboards show wrong figures, teams wait and decisions get postponed.

What really costs money is the time during which the data can no longer be trusted.

5 minspotted at 9:00, fixed at 9:05: a minor incident
SAME MISTAKE
1 daya whole team tied up: a business problem

A mistake spotted at 9:00 and fixed at 9:05 remains a minor incident. The same mistake, if it ties up a team for a whole day, becomes a business problem: decisions made on wrong figures, meetings postponed, trust in dashboards eroded. And the longer the rebuild takes, the higher the risk of introducing a new error while fixing the first one.

Without rollback: data availability and company productivity drop at the incident and stay low throughout the impact period, until resolution
The impact of a data incident on the business, without rollback: until the data is rebuilt, productivity stays at its lowest.
02

Just another Monday morning

Monday · 8:00

The weekly load starts as usual. But the connection to the data lake has failed: one of the source tables arrived empty. The load finished with only part of the bills of materials.

Monday · 10:00

Teams build their comparison dashboards. Links are missing: some items are no longer attached to their bill of materials, and the comparisons are skewed. Without history, someone would have to identify what is missing, rerun a full reload of the context, then wait.

A few minutes later

With Graphicient, the team administering the context brings it back, from the administration panel, to its state before the load. The bills of materials are complete again and the comparison dashboards recover all their links.

Later that day

The connection to the data lake is restored and the table checked. The data is included in the next load.

Nobody worked on incomplete comparisons, and nobody waited for a full reload.

03

Back to a known, reliable state

At Graphicient, every version of the data is kept along with its validity period. A change never replaces the previous state: it is added to it.

Going back therefore does not mean rebuilding. All it takes is to find the exact state of the data just before the incident, then restore it.

1 · STABLE VERSIONThe data is reliableThe reference state, kept with its validity period.
2 · INCIDENTAn error is introducedIncomplete load, deletion, modified rule.
3 · ROLLBACKThe stable version is backThe previous state is restored as it was.

Whatever caused the problem, the answer is the same. In practice, the rollback applies to a given date, for all the nodes and links of a specified context. The source is then fixed during the day, and the corrected data is available from the next load.

ANY DATESince day oneRollback is possible to any date in the context's history.
ADMINISTRATION PANELIn the team's handsIt is triggered by the team administering the context.
REVERSIBLEA rollback of the rollbackIf a restore needs to be undone, it is just as simple.
With Graphicient rollback: after the incident, data availability is restored immediately, in under 5 minutes, and the impact on productivity is avoided
With rollback: the restore is immediate and the impact on the company's productivity is avoided.
04

Consistent rollback on MongoDB and Neo4j

The Graphicient platform relies on two complementary databases: MongoDB and Neo4j.

In a two-database architecture, restoring only one of them is not enough. If the graph goes back to Friday evening while the rest of the data stays on Monday morning, the two databases no longer tell the same story. The incident then turns into an inconsistency, often harder to detect than the original error.

ContextMongoDBNeo4jSame date

That is why Graphicient rollback applies to both databases at once. For the context concerned, MongoDB and Neo4j are brought back together to the same date. The restored data is therefore reliable as a whole, not just in one of its parts.

For business teams, this mechanism stays invisible: they do not need to know which database holds which piece of information. They simply get back a complete, consistent state, the one that existed just before the incident.

05

What rollback changes in practice

≈ 3 hto fully reload the context, without rollback
5 M NODES15 M LINKS
≈ 2 minto restore the same context with rollback
RECOVERYA few minutesRestoring a context of 5 million nodes and 15 million links takes about 2 minutes. Without rollback, it would have to be fully reloaded: about 3 hours, not counting investigation time. Three hours business teams simply cannot afford to wait.
TRACEABILITYA clear audit trailYou know what changed, when, and what was restored.
PEACE OF MINDThe right to make mistakesAn unfortunate manipulation is no longer a disaster. Teams can evolve their rules and models without fearing that an error will be irreversible.
06

The real benefit: time to fix things properly

Once the reliable state is back, the pressure drops. Teams can take the step back they need to understand where the problem came from: which source, which load, which rule.

That time changes the nature of the fix. Instead of a patch applied in a rush, at the risk of creating a new error, the fix follows a controlled sequence:

1 · RESTOREIn a few minutesThe context goes back to its last reliable state.
2 · UNDERSTANDNo rushThe team identifies the source, load or rule at fault.
3 · FIXWithin the dayThe source is properly corrected and checked.
4 · RELOADAt the next loadThe corrected data becomes available.

At no point was the business blocked during the fix.

And it is this regained time that builds trust. When everyone knows a mistake can be fixed calmly, data becomes a working tool again rather than a source of worry.

07

At the core of the platform, not an option

Versioning and rollback are not features bolted on after the fact. They sit at the heart of how we design Graphicient. Data is only valuable if it is available when you need it, so we wanted getting back to a reliable state to be simple, fast and complete, across the whole platform.

Rollback does more than restore data: it gives you time back.

The right to make mistakes, on your own data

We will show you how Graphicient versions your contexts and brings them back to a reliable state in minutes, on your own scope.

Book a demo
Newsletter

Subscribe to our newsletter

Insights on knowledge graphs, straight to your inbox.

Thanks — you're on the list.