Glossary · MonitoringAlert (Monitoring)A monitoring alert is an automatic notification sent when a check or metric crosses a defined condition, such as a site failing to respond.Glossary · MonitoringAlert DeduplicationAlert deduplication merges repeated alerts about the same underlying problem into one incident, so responders are notified once rather than hundreds of times.Glossary · MonitoringAlert FatigueAlert fatigue is when responders are exposed to so many low-value alerts that they begin to ignore, delay or silence them, including important ones.Glossary · MonitoringAnomaly DetectionAnomaly detection in monitoring automatically flags metric values that deviate from normal patterns, instead of relying on fixed thresholds.Glossary · MonitoringApdex (Application Performance Index)Apdex is a 0 to 1 score that summarizes user satisfaction with response times by classing each request as satisfied, tolerating or frustrated.Glossary · MonitoringAPM (Application Performance Monitoring)APM tools instrument application code to measure response times, errors and dependencies so teams can find slow transactions and failing code paths.Glossary · MonitoringAvailabilityAvailability is the share of time a service is working and reachable when users need it, usually expressed as a percentage over a defined period.Glossary · MonitoringBaseline (Monitoring)A monitoring baseline is the normal range of a metric over time, such as typical response time or traffic, used to judge whether current values are abnormal.Glossary · MonitoringBlack-Box MonitoringBlack-box monitoring tests a system from the outside, the way a user would, to detect symptoms like errors and slowness without inspecting internals.Glossary · MonitoringBlameless CultureBlameless culture treats incidents as system failures to learn from rather than individual faults to punish, so people report problems openly.Glossary · MonitoringBlast RadiusBlast radius is the scope of users, systems, or regions affected when a change or failure goes wrong, and the goal is to keep it as small as possible.Glossary · MonitoringBrownoutA brownout is a partial service degradation in which a system still works but with reduced capacity, slower responses or disabled features.Glossary · MonitoringBulkhead PatternThe bulkhead pattern isolates resources into separate pools so a failure or overload in one part of a system cannot exhaust capacity for the rest.Glossary · MonitoringCardinalityIn monitoring, cardinality is the number of unique values a metric label or log field can take, and high cardinality can overwhelm storage and slow queries.Glossary · MonitoringCascading FailureA cascading failure is a chain reaction where one component's failure overloads others, spreading until a small fault becomes a major outage.Glossary · MonitoringChaos EngineeringChaos engineering is the practice of deliberately injecting controlled failures into a system to find weaknesses before real outages expose them.Glossary · MonitoringCircuit BreakerA circuit breaker is a resilience pattern that stops calls to a failing dependency after repeated errors, so the failure cannot cascade through your system.Glossary · MonitoringConfirmation CheckA confirmation check re-tests a failed monitor a second time, often from another location, before sending an alert, to filter out brief network glitches.Glossary · MonitoringConnection TimeoutA connection timeout occurs when a client cannot establish a TCP connection to a server within the allowed time, usually because it is down or blocked.Glossary · MonitoringCorrelation IDA correlation ID is a unique identifier attached to a request and passed through every service that handles it, so logs from different systems can be linked.Glossary · MonitoringDead Man's Switch (Monitoring)A dead man's switch alert fires when an expected signal stops arriving, such as a cron job that fails to check in or a monitoring pipeline gone silent.Glossary · MonitoringDead-Letter Queue (DLQ)A dead-letter queue holds messages that could not be processed after repeated attempts, so one bad message cannot block or endlessly retry a queue.Glossary · MonitoringDegraded PerformanceDegraded performance is a status where a service still works but is slower, less reliable or missing some functions compared with its normal level.Glossary · MonitoringDistributed TracingDistributed tracing follows a single request across multiple services, recording timed spans so you can see where latency or errors occur.Glossary · MonitoringDowntimeDowntime is the period during which a website, application or service is unavailable or not working properly for its users.Glossary · MonitoringError BudgetAn error budget is the amount of unreliability a service may have in a period before it breaks its SLO, used to balance releases against stability.Glossary · MonitoringError RateError rate is the percentage of requests that fail over a time window, usually counted from 5xx responses, and a core signal for alerting and SLOs.Glossary · MonitoringEscalation PolicyAn escalation policy defines who is notified of an alert, in what order and after what delay, so an unacknowledged incident reaches someone who can act.Glossary · MonitoringFalse Negative (Monitoring)In monitoring, a false negative is a missed failure: the check reports healthy while real users are affected, so no alert is sent.Glossary · MonitoringFalse Positive (Monitoring)In monitoring, a false positive is an alert that fires although nothing is actually wrong, such as a down alert for a site that is working fine.Glossary · MonitoringFeature FlagA feature flag is a runtime switch that turns code paths on or off without redeploying, letting teams release gradually and disable problem features instantly.Glossary · MonitoringFive WhysFive Whys is a root-cause technique that asks "why" repeatedly, typically around five times, to trace a problem from its symptom to a fixable underlying cause.Glossary · MonitoringFlappingFlapping is when a monitored service or check rapidly alternates between up and down states, generating repeated alerts and recoveries.Glossary · MonitoringGolden SignalsThe four golden signals, latency, traffic, errors, and saturation, are the minimum set of metrics Google's SRE book recommends monitoring for any service.Glossary · MonitoringGraceful DegradationGraceful degradation is designing a system so that when a component fails, it keeps serving a reduced but useful experience instead of failing completely.Glossary · MonitoringGrafanaGrafana is an open-source platform for building dashboards and alerts from many data sources, such as Prometheus, databases and log stores.Glossary · MonitoringHealth CheckA health check is an endpoint or probe that reports whether a service and its dependencies work, used by load balancers, orchestrators and monitors.Glossary · MonitoringHealth Check EndpointA health check endpoint is a URL, often /health or /healthz, that a service exposes so load balancers and monitors can test whether it is working.Glossary · MonitoringHeartbeat MonitoringHeartbeat monitoring expects a regular ping from your job or service and raises an alert when the ping does not arrive on time, catching silent failures.Glossary · MonitoringIncidentAn incident is an unplanned event that disrupts or degrades a service and requires a coordinated response to restore normal operation.Glossary · MonitoringIncident CommanderThe incident commander is the person who coordinates the response to a major incident, assigning tasks and decisions while others fix the problem.Glossary · MonitoringLiveness ProbeA liveness probe checks whether a process is still functioning, and an orchestrator restarts the instance when repeated checks fail.Glossary · MonitoringLog AggregationLog aggregation collects logs from many servers and services into one central, searchable store so they can be queried and correlated in one place.Glossary · MonitoringLogsLogs are timestamped records of discrete events a system produces, such as requests, errors and state changes, used to investigate what happened.Glossary · MonitoringMaintenance WindowA maintenance window is a pre-announced period set aside for changes such as upgrades or migrations, during which brief downtime is expected.Glossary · MonitoringMetricsMetrics are numeric measurements of a system, such as request rate or error count, recorded over time so trends and thresholds can be tracked.Glossary · MonitoringMTBF (Mean Time Between Failures)MTBF is the average operating time between one failure and the next for a repairable system, used with MTTR to estimate availability.Glossary · MonitoringMTTA (Mean Time to Acknowledge)MTTA is the average time between an alert being sent and a responder acknowledging it, showing how quickly on-call staff pick up incidents.Glossary · MonitoringMTTD (Mean Time to Detect)MTTD is the average time between a failure starting and the team or monitoring system detecting it, a key measure of how good alerting is.Glossary · MonitoringMTTF (Mean Time to Failure)MTTF is the average operating time before a non-repairable component or system fails, used to estimate hardware lifetime and replacement planning.Glossary · MonitoringMTTR (Mean Time to Recovery)MTTR is the average time it takes to restore service after an incident, measured from failure start to recovery; the letter R is used for several meanings.Glossary · MonitoringNagiosNagios is a long-established open-source monitoring system that runs plugins to check hosts and services and notifies staff when they fail.Glossary · MonitoringNines (Five Nines Availability)Nines describe availability by counting nines in the percentage, such as 99.9% (three nines) or 99.999% (five nines), each limiting yearly downtime.Glossary · MonitoringObservabilityObservability is the ability to understand what is happening inside a system from the data it emits, mainly metrics, logs and traces.Glossary · MonitoringOn-CallOn-call is a rotation in which a designated engineer is reachable at any hour to respond to alerts and incidents outside normal working time.Glossary · MonitoringOutageAn outage is a period when a service is completely or substantially unavailable to users, whether due to failure, attack, error or maintenance.Glossary · MonitoringPercentile (p95, p99)A percentile such as p95 or p99 is the value below which that percentage of measurements fall, used to describe response times without averaging away outliers.Glossary · MonitoringPlanned DowntimePlanned downtime is a scheduled period when a service is intentionally unavailable or degraded for maintenance, upgrades or migrations.Glossary · MonitoringPostmortem (Incident Review)A postmortem is a written review of an incident that records what happened, why, how it was resolved and what will be changed to prevent a repeat.Glossary · MonitoringProbe LocationA probe location is the geographic point or network from which a monitoring service runs its checks, which shapes the latency and reachability it reports.Glossary · MonitoringPrometheusPrometheus is an open-source monitoring system that scrapes numeric metrics over HTTP, stores them as time series and evaluates alerting rules with PromQL.Glossary · MonitoringRCA (Root Cause Analysis)Root cause analysis is a structured method for tracing a failure back to the underlying conditions that allowed it, rather than stopping at the symptom.Glossary · MonitoringRead TimeoutA read timeout fires when a connection succeeds but the server sends no data, or stops sending, within the allowed time, often signalling a slow backend.Glossary · MonitoringReadiness ProbeA readiness probe is a check that tells an orchestrator or load balancer whether an instance is ready to receive traffic, without restarting it.Glossary · MonitoringRecovery NotificationA recovery notification is the alert a monitoring system sends when a failing service becomes healthy again, closing the loop on a down alert.Glossary · MonitoringRED MethodThe RED method monitors request-driven services with three metrics: Rate of requests, Errors per second and Duration of requests.Glossary · MonitoringRetry StormA retry storm is a surge of repeated requests from clients retrying failed calls at once, overwhelming a struggling service and blocking its recovery.Glossary · MonitoringRetry with BackoffRetry with backoff re-sends a failed request after progressively longer waits, usually with random jitter, to avoid overloading a service that is recovering.Glossary · MonitoringRetry-After HeaderRetry-After is an HTTP response header telling a client how many seconds to wait, or until what date, before sending the request again.Glossary · MonitoringRollbackA rollback reverts a system to its previous known-good version after a bad deployment, restoring service faster than debugging the problem live.Glossary · MonitoringRolling DeploymentA rolling deployment replaces servers or containers with the new version a few at a time, so the service stays up while the release progresses.Glossary · MonitoringRUM (Real User Monitoring)RUM collects performance and error data from real visitors' browsers as they use a site, showing the speed and failures users actually experience.Glossary · MonitoringRunbookA runbook is a documented, step-by-step procedure for handling a known operational task or alert, such as restarting a service or failing over a database.Glossary · MonitoringSaturationSaturation measures how full a resource is, such as CPU, memory, disk, or connection pool, and is one of the four golden signals of monitoring.Glossary · MonitoringSeverity Levels (SEV1, SEV2, SEV3)Severity levels are a ranking scale for incidents, such as SEV1 to SEV4, that sets urgency, who is notified and how fast the response must be.Glossary · MonitoringSingle Point of Failure (SPOF)A single point of failure is any component whose failure takes down the whole service because nothing else can take over its job.Glossary · MonitoringSLA (Service Level Agreement)A service level agreement is a contract that promises a measurable level of service, such as 99.9% uptime, and defines the remedy when the provider misses it.Glossary · MonitoringSLI (Service Level Indicator)An SLI is a measured number that describes one aspect of service quality, such as the share of requests that succeed or finish under a latency target.Glossary · MonitoringSLO (Service Level Objective)A service level objective is an internal reliability target, such as 99.95% of requests succeeding over 30 days, that guides priorities and alerting.Glossary · MonitoringSmoke TestA smoke test is a small, fast set of checks run right after a deploy to confirm the application starts and its most critical paths still work.Glossary · MonitoringStatus PageA status page is a public page showing the current health of your services and past incidents, so customers can check before contacting support.Glossary · MonitoringStructured LoggingStructured logging writes log events as machine-readable key-value records, usually JSON, so they can be searched, filtered, and aggregated reliably.Glossary · MonitoringSynthetic MonitoringSynthetic monitoring runs scripted checks against your site on a schedule from set locations to detect outages and slowdowns before real users report them.Glossary · MonitoringThreshold AlertA threshold alert fires when a metric crosses a fixed limit, such as response time above 2 seconds or error rate above 5 percent, for a set duration.Glossary · MonitoringUptimeUptime is the share of time a website or service is available and working, usually expressed as a percentage over a month or year, such as 99.9%.Glossary · MonitoringUptime SLAAn uptime SLA is a written promise from a provider to keep a service available for a stated percentage of time, with credits or remedies if it falls short.Glossary · MonitoringWhite-Box MonitoringWhite-box monitoring uses metrics, logs and traces exposed from inside a system to explain why it is slow or failing, not just that it is.