Principles
Flat Pricing Still Needs Honest Telemetry Units
Why Vight separates plan allowances, retention, and operational limits instead of hiding infrastructure complexity behind host-based pricing.
· Vight team · 5 min read
Product principles
Price the signal, not the fleet.
Observability pricing becomes hard to predict when every host, container, CPU, custom metric, indexed field, and viewer can change the bill independently.
Vight uses workspace plans with included telemetry because the unit should follow the evidence customers retain, not the topology they happen to run this month.
01
Architecture should not be a billing multiplier
Splitting a service into more containers should not turn the same workload into a more expensive investigation. Neither should adding developers who need to participate in incident response.
Telemetry volume is not perfect, but it is auditable. It connects the customer-visible allowance to the data Vight accepts and retains.
02
Limits must fail visibly
A hard cap is only trustworthy when the dashboard shows the unit, current usage, remaining allowance, and what happens at the boundary.
Silent sampling protects a vendor's margin by changing the customer's evidence. Explicit limits preserve the decision: increase the allowance, change collection policy, or stop at the cap.
03
Efficiency keeps the model durable
Simple pricing depends on disciplined infrastructure. Bounded queries, batched writes, materialized rollups, and self-telemetry controls are product work because they keep the cost per retained signal predictable.
The price stays understandable when the system behind it treats compute as a budget, not an abstraction.
A pricing unit earns trust when customers can see what it measures and engineers can explain what it costs to serve.