- 01
Market
Saudi Arabia
- 02
Written for
CIO, COO
- 03
Decision
Find out whether the DR plan on file would actually work, before an incident asks
- 04
Languages
EN · AR
The game day
The rehearsal that turns a DR document into a DR capability
A game day is a scheduled, controlled failure: pick a scenario, declare the primary lost, and run the plan exactly as written — because the gaps you want are the ones the written plan contains.
Agree what recovery must achieve
Before anything is restored, the business states its tolerances per system: how much data loss is survivable, and how long an outage is survivable. These are the pass marks the drill is scored against. Targets the business signed, per system.
Declare the failure, follow the paper
The scenario begins and the team recovers using only the documented plan — no shortcuts through production, no asking the one engineer who remembers. If a step only exists in someone’s head, the drill exposes it now rather than the incident exposing it later. The plan executed as written, gaps and all.
Restore the backups into isolation
Real backups are restored into an isolated environment and verified — does the data open, is it complete, does the application start against it. This is the step that surprises nearly every first drill, and it is the entire reason to run one. A restore proven on real data, off production.
Time it, score it, fix it, repeat
The measured recovery time is set against the agreed target, every gap becomes an owned fix, and the drill goes on the calendar — because the estate keeps changing, and a plan proven once decays quietly from that day on. A result, a fix list, and a recurring date.
The gap
The plan passed the audit; the restore has never been asked to pass anything
A DR document can be complete, well written and approved, and still describe a recovery that would fail in its first ten minutes — because documents are reviewed for coverage, and recovery fails on specifics. The credentials for the backup store that only the primary environment held. The restore procedure written for a database two versions ago. The dependency on a service that would be down in the same disaster. The step that says "restore the database" and stands for six unwritten hours of work. None of this is visible to a reader; all of it is visible to a rehearsal, which is why an audit-approved plan and a working recovery are different assets, and only one of them helps during an incident.
Within that gap, one step deserves its own warning: the restore. Backup jobs succeed for years — the metric watched is whether the copy was written, and it was. Whether the copy can be turned back into a working system is a different property, tested by nothing in normal operation: the backup that silently excluded a schema, the archive that takes a day to retrieve, the encryption key that rotated, the restore that works but produces a database the application refuses to start against. Every one of these is a routine first-drill finding, cheap to fix on a Tuesday afternoon and unaffordable during the real event. Restore testing is not a refinement of a backup strategy; it is the half of it that was never tried.
Before the drill
What has to exist before a recovery drill can be trusted
A drill run without these produces a number that means nothing. Each condition is settled before the game day so that what the drill measures is the recovery, not the gaps in its own setup.
Targets the business set, not IT
Recovery time and recovery point are statements of what an outage and a data loss cost the organisation — tolerances only the business can weigh. A drill scored against targets IT invented proves nothing to anyone.
What it makes possible: A pass mark with authority behind it.
An isolated restore environment
Restores are rehearsed somewhere that cannot touch production, so the drill can run to full completion — data verified, application started — instead of stopping at the safe parts.
What it makes possible: A full rehearsal with zero production risk.
A recovery site inside the data’s legal boundary
For estates under residency constraints, the failover location must satisfy the same rules as the primary — a recovery that moves regulated data outside its permitted boundary trades an outage for a violation.
What it makes possible: Continuity that keeps its compliance.
Credentials and keys that survive the disaster
The access needed to recover — backup-store credentials, decryption keys, account access — must live outside the environment being recovered, or the plan depends on the thing it assumes lost.
What it makes possible: A recovery that can begin when the primary is gone.
Our role
What we rehearse alongside your team and what stays rehearsed after us
- 01
The targets, brokered
We facilitate the conversation where the business puts real tolerances on data loss and downtime per system — the conversation most untested plans skipped, which is why their targets are guesses.
- 02
The drill, designed and run
We build the isolated restore environment, design the failure scenario, and run the first game day with your team — timed, observed, and scored against the targets, with every finding logged as a fix.
- 03
The cadence, installed
We leave a drill your operations team can rerun without us and a schedule that keeps it running, because the capability is not the successful drill — it is the practice of drilling as the estate changes.
What the business can tolerate losing, and for how long, are decisions only you can make; we turn them into targets, prove or disprove the plan against them in rehearsal, and leave your team able to keep proving it.
Common questions
What buyers ask
Our DR plan passed the audit. Does that mean it would work?+
It means the document is complete; it says nothing about whether the recovery is real, because those are different properties checked by different instruments. An audit verifies that a plan exists, covers the required scenarios and assigns responsibilities. A drill verifies that the backups restore, the credentials are reachable when the primary is gone, the steps work as written and the clock beats the target. Plans fail on specifics no reviewer can see — a key that rotated, a procedure written for an old version, a step that hides hours of unwritten work — and rehearsal is the only instrument that checks specifics. The audit is worth passing. It is just not the same test.
What usually breaks the first time a backup is actually restored?+
The findings repeat across organisations with remarkable consistency. Access first: the credentials or decryption keys needed for the restore lived in the environment that is presumed lost, or rotated since the procedure was written. Scope second: the backup covered the database but not the object storage, the configuration, or a schema someone added last year, so the restored system is incomplete. Time third: retrieval from cold storage and the restore itself take vastly longer than anyone assumed, because the duration was never measured. And last, the subtlest: the restore completes, the data is present, and the application will not start against it — a version mismatch, a missing migration, a dependency that was never in the backup at all. Every one of these is cheap to discover in a drill and brutal to discover in an incident.
Can we test disaster recovery if our data must stay in the country?+
Yes — residency changes where recovery happens, not whether it can be rehearsed. The constraint has to be honoured twice: the failover location must sit inside the same legal boundary as the primary, and so must the isolated environment the drill restores into, because a test that copies regulated data somewhere the regulation forbids has created a violation in the name of preparedness. In practice this means recovery designs built on in-country regions or facilities, and it makes the drill more valuable, not less: a residency-constrained recovery has fewer location options, so whether the permitted site actually works — capacity, connectivity, restore time — is exactly the kind of assumption that should be proven in rehearsal rather than trusted during an incident.
In this cluster
Rehearse it with us
A first game day scoped and scheduled: which system, which failure scenario, which targets it is measured against — so the DR plan stops being a belief and becomes a result.
Start the conversation
