RTO & RPO Drill Gap and Downtime Exposure Calculator
Compare target RTO and RPO against observed 5-stage recovery timings and replication lag, estimate downtime cost, and export CSV or PDF reports.
Measuring the Whole Outage Clock
Why do written 30-minute RTO targets fall apart so often during an actual outage? Usually because the policy author only timed how long a database snapshot takes to restore. Customers experience five sequential stages: alert detection, human triage and failover sign-off, storage restore and log replay, application smoke checks, and DNS or load-balancer cutover.
Enter your target RTO and RPO alongside observed timings from a drill, replication or backup lag, estimated hourly outage impact, and transaction volume. The calculator shows where the clock actually slips and which recovery architecture tier matches your objective.
Reading the Estimate Plainly
Downtime cost figures and architecture notes here are planning arithmetic built from the numbers you supply. They do not prove a live stack will recover in time—only a timed restore drill under realistic conditions does that. Copy the summary, export CSV or PDF, or print.
Sample Failover Scenario
Select Load payment-DB drill example above to inspect a payment PostgreSQL workload carrying a 30-minute RTO target. Even with a 38-minute cold snapshot restore, adding 6 minutes for alert detection, 11 minutes for triage, 9 minutes for verification, and 10 minutes for DNS TTL propagation brings total observed downtime to 74 minutes—a 44-minute overrun that points toward a warm streaming standby and shorter DNS TTLs.
Related Guides and Services
Frequently Asked Questions
What is the difference between RTO and RPO?
Recovery Time Objective (RTO) looks forward on the clock from the start of an outage until normal service returns. Recovery Point Objective (RPO) looks backward from the outage to the last durable backup or replicated commit, defining how many minutes or hours of recent writes the business can afford to lose.
When should the RTO clock begin during a recovery drill?
Track both total outage duration (starting when the fault occurs, including alert lag and human triage) and hands-on technical execution time after failover is called. Starting the stopwatch only when an engineer runs the restore command hides the twenty minutes spent figuring out what broke.
Why do cloud snapshot restores often run slower than raw disk throughput suggests?
Block volumes restored from object-backed cloud snapshots often hydrate data blocks lazily on first read. Until warm-up finishes—or unless fast snapshot restore or a warm replica is already running—early database queries hit high I/O latency.