Frequently Asked
Questions

LogNexis is an enterprise-grade API observability platform. Unlike traditional logging tools that just store raw text files, LogNexis actively monitors every request, streams metrics via WebSockets, and uses Machine Learning to detect latency anomalies and payload structural errors before they cause system-wide outages.

Yes! Our generous free tier allows up to 100,000 API requests per month with 7 days of data retention. You get access to 3 projects, real-time log streaming, and basic ML baselining. No credit card is required to sign up.

Datadog and New Relic are fantastic generalized infrastructure monitoring tools, but they require significant manual configuration and can be overwhelmingly complex. LogNexis is laser-focused purely on API health. By restricting our scope to APIs, we provide zero-config Machine Learning anomaly detection specifically trained on HTTP time-series data.

Integration takes less than 3 minutes. Simply install our SDK (npm install lognexis-node) and attach it as middleware to your Express, Fastify, or Koa application. The SDK runs asynchronously and uses less than 1ms of overhead per request.

We offer official SDKs for Ruby (Rails), Python (FastAPI, Flask, Django) and Go (Golang). The Java / Spring Boot SDK is currently in development. You can also integrate any backend technology using our cURL/REST API guides.

No. The LogNexis SDK processes telemetry data entirely off the main thread. It aggregates request metrics and flushes them to our ingestion servers asynchronously, guaranteeing zero impact on your application's critical path.

We utilize an unsupervised Machine Learning algorithm called the Isolation Forest. Instead of requiring you to set manual threshold rules (e.g., alert if latency > 500ms), the algorithm establishes a dynamic baseline of normal variance based on seasonality. It automatically flags statistical outliers without manual configuration.

The initial Learning Phase takes roughly 24 to 48 hours, or a minimum of 10,000 requests. During this time, the model maps out diurnal seasonality (day vs. night traffic patterns) and request distributions. You can still view live logs instantly, but ML alerts will begin firing after this phase.

Yes. Inside your Project Settings, you can configure the Anomaly Sensitivity Score. A high sensitivity (0.85) will alert you to minor latency drifts, while a low sensitivity (0.99) will only trigger on massive, undeniable system failures.

Security is our top priority. All telemetry is encrypted in transit using TLS 1.3 and at rest with AES-256. We utilize strict UUID v4 API keys for ingestion, and you can enable IP allowlisting for dashboard access.

By default, we collect routing metadata: the endpoint path, HTTP method, status code, latency, timestamp, IP address, and User-Agent. We do NOT collect request/response body payloads unless you explicitly opt-in for debugging purposes.

The LogNexis SDK includes built-in payload scrubbing. You can configure a blocklist of keys (e.g., 'password', 'credit_card', 'ssn', 'email'). The SDK will automatically redact these fields from headers or logged bodies before the data ever leaves your servers.

If you exceed your tier's limit, we will not block or drop your traffic immediately. We provide a 10% grace buffer. If you consistently exceed limits, we will pause new log ingestion until the next billing cycle, but your historical data remains fully accessible.

Yes. Upgrades are prorated and applied immediately so you can handle sudden traffic spikes without data loss. Downgrades will take effect at the start of your next billing cycle.