Testing program
SnapStream runs its own drill campaign before any of this touches a production cluster. Each tier has pass/fail criteria set before the drill runs, and every result ships with a sealed evidence manifest.
Three tiers
Dry runs against a simulated cluster. No live nodes are touched.
Controlled drills on two real nodes, with restarts we trigger ourselves.
A private 20-node cluster on infrastructure we own, where restarts are genuinely induced rather than simulated.
How a drill is graded
Pass/fail criteria are set before a drill runs, not chosen afterward to fit the result. Evidence from each run goes into a manifest sealed with a SHA-256 hash, so it can't be edited after the fact. Failures get published too. That's the policy.
Tier 1 evidence
Tier 1 is complete. Its replay is public on our drill deck, along with the sealed manifest behind it.
Open the drill deck →