Privacy by Design. Evidence Without Exposure.
We believe telemetry should inform engineers without compromising confidential application data, customer tokens, or environment credentials.
1. Recursive Redaction Engine
Before any payload is persisted or processed by AI models, Lazers runs deep recursive sanitization across JSON payloads, HTTP headers, stack frames, and error messages.
- ✓ Automatic scrubbing of Authorization, Cookie, and X-Api-Key headers
- ✓ Pattern matching for JWTs, bearer tokens, private keys, and session cookies
- ✓ Redaction of Lazers API keys (
laz_live_...andlaz_ci_...)
2. Dual-Layer Credential Separation
Browser and mobile telemetry use write-only client DSNs (laz_ci_) that are cryptographically restricted to signal ingestion. Client DSNs cannot read operational history, query incidents, or alter project configurations. Backend servers and OpenTelemetry pipelines use secret server API keys (laz_live_).
3. Bounded Telemetry & Tenant Isolation
Every organization and project operates in an isolated tenant boundary. Queries and AI analysis contexts enforce strict database-level project filters. Telemetry payloads have enforced byte ceilings and rate limiters to protect against runaway loops.
4. Responsible Disclosure Policy
We welcome responsible security research. If you discover a vulnerability or security concern, please report it to our dedicated security team: