Questions to ask each cloud DR vendor before a trial

Questions for every vendor

1. Which resource types do you restore, per cloud, and in what form: direct changes, generated code or a replica?
2. What RTO and RPO do you state, for which tier, and how were they measured?
3. How do you detect drift, and what happens to a resource that exists in the account but not in your record?
4. Can we recover into a clean, isolated account, and how do we fail back afterwards?
5. How do we run a recovery test without touching production, and what does the test report show?

Send these five before the first call and ask for written answers. A vendor that answers them in writing has given you a baseline to hold the trial against, and the answers are easy to compare side by side.

Firefly

• Backup and DR is in the Enterprise tier, which is custom priced. What does it cost for our asset count, and what does Essential include for recovery, if anything?
• The RTO under 1 hour: for what environment size, and does it include applying the generated Terraform?
• Google Cloud is not in the Backup and DR coverage table. Is it planned, and when?
• The docs say restoration depends on Terraform provider support. Which of our resource types fall outside it?
• Is there a recovery test workflow, and is failback supported?
• Which data backup products do customers usually pair with it?

Arpio

• Drift detection and a change history are not published. How would we recover from a bad configuration change that replication has already copied?
• No RTO figure is published. What recovery time do you state for Standard and Premium?
• Recovery does not produce infrastructure as code on the pages we read. Can we export the recovered environment as code?
• Google Cloud is named as an expansion goal in the June 2026 Series A announcement. Is there a date?
• Which of our components count as core resources?

Commvault Cloud Rewind

• Per-instance prices are on AWS Marketplace only. What do Azure and Google Cloud cost, and what counts as an instance?
• No RTO or RPO is published. What do you state?
• Cross-account recovery is currently AWS only. When will it cover Azure and Google Cloud?
• Commvault markets Cloud Rewind as a way to skip complex IaC, while the G2 description says recovery environments are programmed with IaC. Can we export the recovered environment as Terraform?
• Drift analysis appears only in a third-party quote on your page. Is drift detection and remediation available?

ControlMonkey

• No RTO or RPO figure is published. What do you state?
• A recovery test or drill workflow is not described. How do we test a recovery?
• Which tier includes replication to a secondary region or account?
• You protect configuration, not data. Which data backup products do customers pair you with?

Cohesity

• Cloud Rebuild is stated for select AWS workloads in Terraform-provisioned environments. Which workloads, and what happens to resources not in Terraform?
• Are Azure and Google Cloud covered for recovery?
• No price or recovery target is published. What do you state?
• How do the clean room and Cloud Rebuild work together after an attack?

StackGuardian

• Is any backup or recovery capability planned, or should we plan to pair it with a recovery tool?
• How do your codified outputs work with the recovery tools we are considering?
• What do enterprise plans cost?

Gambit Security

• Which clouds does Balens support?
• What unit does the $1,000,000 AWS Marketplace contract cover?
• Which recovery tools does it hand off to for the rebuild?

Scalr

• Is any backup or recovery capability planned?
• How would Revert Infrastructure handle a bad change made through code rather than drift?

Questions about the trial itself

• Which application should we use? Pick one that is critical but small, with networking, identity and a database, so every layer is tested.
• What is the success test? Agree before you start: the service answers requests in the recovery target, within a time you name.
• Who runs it? A vendor engineer running the trial proves the tool works; your own team running it proves your team can use it during an incident. Do both if you can.

How to use the answers

Ask for answers in writing, then test the two or three that matter most to you in the trial on one of your own applications. Add what you learn to compare several by reading each score's reason against the vendor's answer. If a vendor's answer contradicts what we publish, send us the public source through the contact page.