Alerts that actually reach a human in time

A bolted-together Grafana stack technically has alerting, but its evaluation engine polls your time-series database on a fixed interval, so a fault that starts and clears between scrapes is simply never seen. Routing is an afterthought, multi-channel delivery means stitching together extra services, and there is no per-reading context, so by the time someone notices the red panel the bearing is already running hot or the batch is already out of spec. Alerting becomes a maintenance project of its own instead of a safety net.

• Interval polling misses short-lived spikes and transients entirely
• Acknowledgements, escalation, and de-duplication are DIY or absent
• Routing to email or chat needs separate plugins and contactpoint config
• No tamper-evident record of who was notified, when, or whether they acted

Sentrel evaluates its rule engine on every single reading as it lands over MQTT or HTTPS ingest, not on a scrape timer, so a threshold breach is caught the instant it occurs. When a rule trips, Sentrel pushes the event over a live WebSocket to anyone with the dashboard open, raises an in-app toast, POSTs a structured payload to your webhooks, and sends email all at once. Each alert carries the device, the offending value, the rule that fired, and the integrity hash of the source reading, so responders see exactly what happened without hunting through a chart.

• Rules evaluated per reading at ingest, not on a polling interval
• WebSocket live updates so open dashboards react in real time
• In-app toasts plus webhooks plus email from one rule definition
• Every alert links back to the tamper-evident reading that triggered it
• Scoped API and MQTT keys keep alert delivery locked to the right tenant

The difference is whether an alert is a feature or a project. Sentrel ships per-reading evaluation, four delivery channels, and an auditable trail in the box, configured from the same plain-English thresholds you used to describe your equipment. A Grafana-plus-exporter assembly makes you build, host, and babysit each of those pieces, and still leaves you with interval blind spots and no compliance-grade record. With Sentrel, the operators and integrators who actually run the plant get notified through the channel they already watch, with the context to act, and a WORM-grade log proving the alert went out.

• No polling gaps: catch transient faults a scrape interval would miss
• No glue code: toasts, webhooks, email, and live updates from one rule
• Compliant by default: notification events captured in the audit trail
• White-label ready: alerts carry your branding for resale by OEMs and integrators