# Monitoring > Synthetic probes from your regions, anomaly detection that catches drift, and severity-aware routing to every channel your operators already live in. *Monitoring* ## See every heartbeat. Catch every incident. [Request a demo](https://calendly.com/apinizer/15min) · [Read the docs](https://apinizer.com/developers/docs) **Highlights** - **Detection** — Uptime · Anomaly · Platform - **Math** — EMA + Bollinger - **Channels** — 9 actions · 15+ connectors --- ## Capabilities ### 01 · Probe like a customer. From everywhere they live. Synthetic checks run from the regions your gateway already does — and validate not just a 200 OK, but the right body, headers, and latency budget the SLA promises. - Three targets — managed proxy, backend API, external endpoint - Schedule · 1 / 5 / 10 / 30 minutes or a cron expression - Assertions · status, body, XPath, JsonPath - Retry on fail with configurable count and delay - Same monitor on every environment — dev, test, production **Concepts:** `3 target types` · `Status · body · XPath · JsonPath` · `Retry on fail` · `Multi-environment` ### 02 · Catch issues no threshold could. EMA with Bollinger bands evaluates the trend, not just the value. Two sliders tune sensitivity; one toggle picks fire-per-event or once-per-series. - Any Elasticsearch query as the source - EMA with upper and lower Bollinger bands - Sensitivity · data points + standard deviation multiplier - Execute mode · per event or once per series - Charted detection — metric, EMA, bands, every marker **Concepts:** `Threshold` · `EMA + Bollinger` · `Per event / series` · `Trend-aware` ### 03 · Pods, nodes, certificates — every alert your operators want. Ten triggers ship with the platform — Kubernetes, Elasticsearch, certificates, and logs. Each comes with a threshold, a schedule, a severity, and the same action chain you use everywhere else. - Kubernetes · pod health, node health, node CPU - Elasticsearch · cluster health, CPU, disk - Certificates · SSL and JWK expiry windows - Logs · application log count and traffic log presence - Same severity grades and action chains as everything else **Concepts:** `10 triggers` · `Kubernetes` · `Elasticsearch` · `SSL · JWK` · `Log presence` ### 04 · CPU and memory. Per component. Sustained. Manager, Gateway Worker, Cache, and Log — each has its own threshold and duration window. A one-second spike never pages anyone; eighty percent for thirty seconds does. - Per-component thresholds for Manager, Worker, Cache, Log - Duration window — sustained-over-time, not a spike - Same severity grades as API alerts - Same action chains as the rest of the platform - Audit of every breach and every resolution **Concepts:** `Manager` · `Worker` · `Cache` · `Log` · `Threshold + duration` ### 05 · Right person. Right channel. Right now. Severity decides where the alarm goes — P1 to on-call, P3 to the daily digest — across nine action types and fifteen-plus connectors. - Severity-aware routing · P1 to on-call, P3 to digest - Nine action types · email, webhook, API, DB, script, SNMP - Connector library · Slack, Teams, ServiceNow, and 15+ more - Grouping, deduplication, acknowledgment, escalation - Acknowledge from Slack — no context switch **Concepts:** `Slack · Teams` · `ServiceNow · Email` · `Webhook · API · DB` · `Script · SNMP` · `Group · dedup · escalate` ### 06 · Four grades. Every probe stored. Full audit. Info, warning, error, critical — every detection is graded, grouped, deduplicated, and audited. Retention is configurable per surface; exports go straight to PDF, CSV, or Excel. - Four severity grades · info, warn, error, critical - Grouping, deduplication, series for repeated detections - Retention per surface · uptime, anomaly, alerts, audit - Acknowledged · escalated · resolved as state transitions - Export to PDF, CSV, Excel — auditor-ready **Concepts:** `Info · Warn · Error · Critical` · `Group · dedup · series` · `Configurable retention` · `PDF · CSV · Excel` ### 07 · One screen. All four lanes. Monitors, results, anomalies, notifications — one screen, no console switching. Same screen across dev, test, and production. - KPI strip · monitors, success ratio, average response, open alerts - Scope tree by project and monitor type - Result grid with coloured status, response time, timestamp - Notification panel with unread badge and severity-coloured cards - Same screen across every environment **Concepts:** `KPI strip` · `Scope tree` · `Result grid` · `Notification panel` · `Multi-env` --- ## Use cases ### Probe the way the SLA reads Pick the target, the interval, the assertions. Every probe is stored, the success ratio is charted, and the action chain fires on the first real failure. - Managed proxy · backend · external API - Status · body · XPath · JsonPath - Retry on fail with configurable count and delay - Per-environment enablement and retention ### Catch behaviour changes thresholds miss EMA with Bollinger bands evaluates the trend, not just the value. Fire per event or once per series. - Threshold rules for the obvious cases - EMA with upper and lower Bollinger bands - Two sensitivity sliders - Per-event or once-per-series action firing ### Stop being surprised by expiring certificates Every SSL and JWK certificate is tracked. Pre-expiry alarm windows. Action chains to renewers, tickets, and operators. - SSL expiry tracking per host - JWK expiry tracking per signing key - Configurable pre-expiry windows - Action chains to renewers and tickets ### Operate the platform itself, not just the APIs Manager, Gateway Worker, Cache, and Log — CPU and memory rules with a duration window. Same playbook as the API surface. - Per-component thresholds - Sustained-over-time duration window - Same severity and action chains as alerts - Audit of every breach and resolution --- ## What ships in the box ### Detection - Uptime probes — 3 targets, 4 assertion types, retry on fail - Anomaly detection — EMA, Bollinger bands, sensitivity tuning - Ten platform alert triggers ready on day one - System health — CPU + memory + duration window per component - Query + filter + condition pipeline over Elasticsearch ### Response - Nine action types — email, webhook, API, DB, script, SNMP - Connector library — Slack, Teams, ServiceNow, and 15+ more - Four severity grades with per-grade channel routing - Grouping, deduplication, acknowledgment, escalation - Configurable retention and audit exports (PDF, CSV, Excel) --- ## Resources - [Monitoring overview](https://apinizer.com/developers/docs) — What each surface watches and how the action chains connect. - [Uptime monitor reference](https://apinizer.com/developers/docs/uptime-monitor) — Targets, intervals, assertions, retries, retention — the full surface. - [Anomaly detector guide](https://apinizer.com/developers/docs/anomaly-detector) — EMA and Bollinger bands with sensitivity tuning and query examples. - [Platform alerts catalog](https://apinizer.com/developers/docs/alerts) — Every shipped trigger, default severity, recommended action. - [Action chains and connectors](https://apinizer.com/developers/docs/actions) — Wire one alarm to multiple channels with severity-aware routing. - [Operations dashboard](https://apinizer.com/developers/docs/monitoring-dashboard) — One screen for monitors, results, anomalies, and notifications. --- ## Next step *Operate confidently* **Stop fighting fires. Start preventing them.** Uptime, anomaly, alerts, and severity-aware fan-out — wired to your gateway in 30 minutes. [Book a Demo](https://calendly.com/apinizer/15min) · [Read the docs](https://apinizer.com/developers/docs) --- ## Links - Products: https://apinizer.com/products - AI Gateway: https://apinizer.com/products/ai-gateway - Solutions: https://apinizer.com/solutions - Pricing: https://apinizer.com/pricing - Developers: https://apinizer.com/developers - Documentation: https://docs.apinizer.com/index-en - Blog: https://apinizer.com/blog - Contact: https://apinizer.com/company/contact © 2026 Apinizer. All rights reserved.