Observability with Prometheus and Grafana

Logz.io Infrastructure Monitoring (Prometheus-as-a-Service)

Logz.io is an observability SaaS built on open source (ELK for logs, Prometheus for metrics, Jaeger/OpenTelemetry for traces). Its Infrastructure Monitoring product is “Prometheus-as-a-Service”: you send metrics via Prometheus remote_write or Logz.io collectors, they handle storage, scaling, and high availability. Metrics can be queried with PromQL, and are unified with logs and traces in the same platform for troubleshooting. Metrics use a multi-resolution backend that keeps recent data at high resolution and downsampled aggregates for older data. Logz.io provides both a hosted Grafana and its own UI for metrics exploration.

• Free Tier & Pricing: There is no permanent free tier for metrics, only a 14-day full-feature trial. After that, pricing is consumption-based: Infrastructure Monitoring is advertised at about $0.40 per 1,000 time series per day (≈$12 per 1,000 active series per month) with 18 months retention included. Pricing assumes an average of 6 data points per minute per series (10-second scrape). Higher scrape rates count as proportionally more series (e.g., 12 DPM ≈ 2 series). The price bundles ingestion, storage, retention, querying, and hosted Grafana; there are no separate query/retention fees. Other products (logs, traces) are billed separately, with potential bundled enterprise deals. Volume discounts and custom quotes are available, especially at large scale.

• Storage Retention: Default retention for metrics is 18 months. To support this, Logz.io uses automatic downsampling:

• Full-resolution (raw) data for ~40 hours.
• 5-minute resolution for data from 40 hours to 10 days.
• 1-hour resolution from 10 days to 18 months.

Each level stores aggregates (e.g., count/sum/min/max), so you retain visibility into spikes via min/max, but you lose per-second detail beyond 40 hours. Downsampling is automatic and cannot be disabled. Trial accounts typically have shorter retention (e.g., ~30 days); production accounts get the full policy.

• Query Language & Tools: PromQL is the main query language. You can query metrics via a hosted Grafana instance (included) or through Logz.io’s own Infrastructure Monitoring UI (“Explore Metrics”), both using PromQL. Existing Grafana dashboards can be imported and will work by pointing panels to the Logz.io Prometheus-compatible endpoint. Alerting is integrated: you define PromQL-based alerts in the Logz.io UI, and they run server-side similar to Alertmanager, with notification channels such as Slack, PagerDuty, Opsgenie, ServiceNow, and more. Cross-data correlation is supported: from a metric spike you can pivot to related logs or traces in the same UI.