Alerts · SLOs
Stop firefighting. Start predicting.
Alerts across metrics, logs, APM, RUM, SLOs, events, and traces. Anomaly detection, forecasting and outlier detection catch the drift before it pages anyone. Burn-rate windows on any signal.
Alerts across metrics, logs, APM, RUM, SLOs, events, and traces. Anomaly detection, forecasting and outlier detection catch the drift before it pages anyone. Burn-rate windows on any signal.
Spot something in service details, log analytics, APM analytics, or events? Create an alert right there — no context switching to a separate configuration page.
Rolling Quantile, SARIMA, Seasonal Decompose and Prophet for anomalies; Linear and Seasonal forecasting; DBSCAN for outliers — all running on your data, inside your VPC.
Latency, conversion, cache hits, custom signals, log-derived rates.
An alert nobody acts on costs more than no alert at all. These are the controls that decide how much of your alerting reaches a human.
Suppression windows silence planned maintenance for a set interval, matched on labels with equals, not-equals and regex operators — so a deploy window mutes the deploy, not the estate.
Routing rules match on labels, severity, service or team ownership, and notify multiple channels at once across eight contact points — Slack, PagerDuty, OpsGenie, Teams, Email, Telegram, Discord and webhooks.
Every alert carries the query that fired it, the time window, and one-click drill into the matching logs and traces. The first minute of an investigation is already done.
Threshold, anomaly, forecast, outlier, and burn-rate detection — evaluated by one rule engine over metrics, logs, traces, APM, RUM, SLOs, and events. Two clicks from investigation to monitoring.
Kloudfuse evaluates alerts across metrics, logs, APM, traces, RUM, SLOs, and events through a single rule engine. Create alerts directly from service detail pages, log analytics, APM analytics, or events views — two clicks from investigation to monitoring.
| Signal | Threshold | Anomaly | Forecast | Outlier | Burn-rate |
|---|---|---|---|---|---|
| Metrics | ✓ | ✓ | ✓ | ✓ | ✓ |
| Logs | ✓ | ✓ | ✓ | ✓ | ✓ |
| APM | ✓ | · | · | · | ✓ |
| Traces | ✓ | · | · | · | · |
| RUM | ✓ | · | · | · | ✓ |
| SLOs | ✓ | · | · | · | ✓ |
Static thresholds miss slow degradations and seasonal shifts. Kloudfuse ships anomaly detection, forecasting, and outlier detection that adapt to your data patterns and alert on deviations without manual tuning.
Kloudfuse implements multi-window, multi-burn-rate SLOs following the Google SRE Workbook methodology. Burn-rate alerts catch both fast burns and slow leaks before your budget is exhausted.
Alerts route to Slack, PagerDuty, OpsGenie, Microsoft Teams, Email, Telegram, Discord, or any webhook. Every alert includes the underlying query, the time window, and one-click drill to the matching logs and traces.
Eight contact points, one routing layer. Match on labels, severity, service, or team ownership — and suppress during planned maintenance windows.
Kloudfuse ships with preinstalled infrastructure alerts for compute, storage, networking, and Kubernetes workloads. Teams already running Datadog or Wavefront import their alert definitions directly.
100+ rules live on day one — prebuilt infrastructure coverage plus your imported Datadog and Wavefront definitions, thresholds and routing preserved.
Kloudfuse supports 12 alert types across two dimensions. Signal-based alerts cover Metrics, Logs, APM, Traces, RUM, and SLOs. Detection methods include Threshold, Change, Anomaly (4 algorithms), Forecast (Linear and Seasonal), and Outlier (DBSCAN). Availability varies by signal; the matrix above lists the detection methods supported for each alert type.
Kloudfuse offers 4 anomaly detection algorithms: Basic (Rolling Quantile), Agile (SARIMA), Robust (Seasonal Decompose), and Agile-Robust (Prophet). Each algorithm is suited to different data patterns. Configure sensitivity using bounds of 1, 2, or 3 standard deviations. Availability varies by signal — the matrix above lists the detection methods supported for each.
Yes. Kloudfuse supports direct alert import from Datadog and Wavefront. Use the built-in catalog service for bulk migration or Python scripts for custom workflows. Alert definitions, thresholds, and notification routing are preserved during import.
Kloudfuse implements SLOs using multi-window, multi-burn-rate alerting based on the Google SRE Workbook methodology. Three SLO types are supported: Availability (error rate), Latency and Custom (PromQL-based numerator/denominator). Error budgets are tracked in real time, and burn rate alerts fire before your budget is exhausted.
Kloudfuse routes alerts through a Contact Points system supporting Email, Slack, PagerDuty, OpsGenie, Microsoft Teams, Webhook, Telegram, and Discord. Routing rules can be configured based on alert labels, severity, service, or team ownership.
Connect your data sources, set up your first cross-signal alert, and configure notifications. No proprietary agents required.
Read the guideHow to define service level objectives using any metric with PromQL — tracking business outcomes, not just infrastructure health.
Read the blogDetection methods, SLO implementation, notification routing, and pricing. A side-by-side breakdown for teams evaluating both.
See the comparisonThirty minutes on your telemetry. The cause, before the call ends.