Tagged monitoring

MTTR, MTBF, downtime and incident counts for a reporting window, from a list of incidents.

Aggregate a metric series into fixed time buckets (count, min, max, avg, sum, last), empty buckets included.

The samples of a time series from one moment up to (not including) another, checking they are in time order.

How fast an SLO's error budget is burning: the error rate as a multiple of the rate the SLO allows, in thousandths.

Multiwindow, multi-burn-rate SLO alerting: which rules fire, from event counts per window (Google SRE workbook).

How much of an SLO's error budget is allowed, used and left, from good and bad event counts or seconds.

Time-weighted uptime of a target over a window from its checks, with gaps longer than a limit counted as unknown.

A heartbeat's state as a check status for a status page: up, late as degraded, down as down, new as nothing yet.

One bar per UTC day for a status page's uptime history: seconds in each state, uptime and a colour state, oldest first.

Validate one network inventory entry: name, host, ports, optional MAC and subnet, with every problem listed.

Whether to hold an alert back: its component is under maintenance, or a related alert on it is already firing.

Evaluate a threshold alert rule over a metric series: inactive, pending, firing or no-data, Prometheus style.

Step an alert through inactive, pending, firing and resolved, and say which notification to send.

Is a new metric value an outlier against its history? By z-score or by the robust median/MAD modified z-score.

A TLS certificate's expiry state as a check status: ok and warning are up, critical is degraded, expired is down.

How long a TLS certificate has left before its notAfter, and whether that is ok, a warning, critical or expired.

Combine the statuses of one target's probes from several locations into one up, degraded or down, with a quorum.

A status page component's state from its checks: operational, degraded, partial or major outage, or maintenance.

How much a monotonic counter grew over a series of samples, counting restarts from zero, and its rate per second.

Whether a cron job that pings on a schedule is new, up, late or down, from its last ping, period and grace time.

Classify one HTTP probe as up, degraded or down, with a reason and message, from status, body and latency.

Turn a series of up, degraded and down checks into incidents, ignoring runs too short to confirm.

The scheduled maintenance window in force at a moment, if any, so alerts and uptime can be suppressed during it.

Parse one Apache/NCSA Common or Combined Log Format access log line; null for a line in neither format.

Count HTTP status codes by class (1xx to 5xx) and give the server error and client error rates in basis points.

The overall state, headline and affected components at the top of a status page, from each component's state.

Classify a device as up, degraded or down from its latency summary and loss, p95 and jitter thresholds.