Recently updated terms
New definitions and material revisions to the reliability glossary.
2026-07-17
- Uptime monitoringUptime monitoring repeatedly checks whether a website, API, or service is reachable and behaving as expected from outside the system.
- Website monitoringWebsite monitoring checks a public web page or site on a schedule to confirm it responds successfully for visitors.
- API monitoringAPI monitoring checks HTTP endpoints on a schedule to confirm status codes, timing, and response content still match expectations.
- Endpoint monitoringEndpoint monitoring is watching a specific URL or route. Teams use the term to keep checks, alerts, and reviews precise.
- HTTP monitoringHTTP monitoring is checking services over HTTP. Teams use the term to keep checks, alerts, and reviews precise.
- HTTPS monitoringHTTPS monitoring is checking services over HTTPS with TLS. Teams use the term to keep checks, alerts, and reviews precise.
- External monitoringExternal monitoring is checking a service from outside its private network. Teams use the term to keep checks, alerts, and reviews precise.
- Service monitoringService monitoring is watching a customer-facing service for health. Teams use the term to keep checks, alerts, and reviews precise.
- Availability monitoringAvailability monitoring is measuring whether a service is usable when expected. Teams use the term to keep checks, alerts, and reviews precise.
- Synthetic monitoringSynthetic monitoring is running scripted checks that simulate user or client requests. Teams use the term to keep checks, alerts, and reviews precise.
- MonitorMonitor is a configured check that runs on a schedule. Teams use the term to keep checks, alerts, and reviews precise.
- Scheduled checkScheduled check is a single execution of a monitor on its interval. Teams use the term to keep checks, alerts, and reviews precise.
- Manual checkManual check is an on-demand test run outside the normal schedule. Teams use the term to keep checks, alerts, and reviews precise.
- Monitoring intervalMonitoring interval is how often a monitor runs its scheduled check. Teams use the term to keep checks, alerts, and reviews precise.
- Monitoring regionMonitoring region is the geographic or network location that runs a check. Teams use the term to keep checks, alerts, and reviews precise.
- Monitoring historyMonitoring history is the stored record of check results over time. Teams use the term to keep checks, alerts, and reviews precise.
- Health endpointHealth endpoint is a dedicated URL that reports whether a service is ready to work. Teams use the term to keep checks, alerts, and reviews precise.
- Health checkHealth check is a request or probe that evaluates whether a service is healthy. Teams use the term to keep checks, alerts, and reviews precise.
- Service healthService health is the overall condition of a service relative to expected behavior. Teams use the term to keep checks, alerts, and reviews precise.
- Availability checkAvailability check is a probe that asks whether a dependency or service is available now. Teams use the term to keep checks, alerts, and reviews precise.
- RetryRetry is repeating a failed check or delivery attempt before escalating. Teams use the term to keep checks, alerts, and reviews precise.
- TimeoutTimeout is the maximum time allowed for a request or check before it fails. Teams use the term to keep checks, alerts, and reviews precise.
- Response-time thresholdResponse-time threshold is the maximum acceptable duration for a successful response. Teams use the term to keep checks, alerts, and reviews precise.
- IncidentIncident is a tracked period of degraded or failed service.
- Incident managementIncident management is the process of detecting, coordinating, resolving, and reviewing service failures.
- Incident detectionIncident detection is discovering that a service likely failed or degraded.
- Incident verificationIncident verification is confirming a failure is real before treating it as an incident.
- Incident responseIncident response is the coordinated actions taken while a service is degraded or down.
- Incident severityIncident severity is a ranked label for how bad an incident is.
- Degraded performanceDegraded performance is a state where a service works partially or slowly but is not fully down.
- Partial outagePartial outage is a failure that affects some features, regions, or customers but not the whole service.
- Major outageMajor outage is a severe failure that blocks core customer journeys.
- Service disruptionService disruption is an interruption to normal service behavior, whether partial or complete.
- Recovery confirmationRecovery confirmation is evidence that a service has returned to healthy behavior after a failure.
- Incident resolutionIncident resolution is closing an incident after the service is healthy again and work is complete.
- Incident reopeningIncident reopening is returning a resolved incident to an active state when failure returns.
- FlappingFlapping is rapid oscillation between healthy and failed states.
- False positiveFalse positive is an alert or incident that fired when the service was actually fine.
- False negativeFalse negative is a real failure that monitoring failed to detect.
- Incident timelineIncident timeline is the ordered record of detection, updates, and recovery for an incident.
- Incident acknowledgmentIncident acknowledgment is an explicit signal that a human has seen the incident and is engaging.
- Incident assignmentIncident assignment is naming the person or role responsible for driving an incident.
- Incident recapIncident recap is a short after-action summary of what happened and what changed.
- Root-cause analysisRoot-cause analysis is structured investigation into why an incident happened.
- Post-incident reviewPost-incident review is a blameless review of detection, response, and prevention after an incident.
- AlertAlert is a message that tells someone a monitored condition needs attention.
- Outage alertOutage alert is a notification that a service has entered a failed or down state.
- Recovery alertRecovery alert is a notification that a previously failing service is healthy again.
- Alert channelAlert channel is a configured destination for alert delivery such as email or Slack.
- Alert routingAlert routing is rules that decide which channels receive which events.
- Alert ruleAlert rule is a specific condition that maps events to channels or suppression behavior.
- Alert fatigueAlert fatigue is desensitization caused by too many low-value alerts.
- Alert deduplicationAlert deduplication is suppressing repeated identical alerts while a condition remains open.
- Quiet hoursQuiet hours is a window when non-critical alerts are delayed or suppressed.
- Alert retryAlert retry is sending an alert again after a delivery attempt failed.
- Alert deliveryAlert delivery is the attempt to send an alert to a channel and record the outcome.
- Dead-letter queueDead-letter queue is a holding place for alert deliveries that exhausted retries.
- Webhook alertWebhook alert is an alert delivered as an HTTP POST to a URL you control.
- Slack alertSlack alert is an alert delivered into a Slack workspace channel or conversation.
- Discord alertDiscord alert is an alert delivered into a Discord channel via webhook integration.
- Email alertEmail alert is an alert delivered to one or more email addresses.
- Status pageStatus page is a public or private page that communicates current service health and incidents.
- Hosted status pageHosted status page is a status page served on infrastructure managed by the monitoring provider.
- Public status pagePublic status page is a status page anyone can open without logging in.
- Private status pagePrivate status page is a status page limited to authorized viewers.
- Status-page componentStatus-page component is a named part of a product shown on a status page with its own health.
- Component groupComponent group is a collection of related status-page components shown together.
- Public incidentPublic incident is an incident projection written for customers on a status page.
- Status-page subscriberStatus-page subscriber is a person who opts in to receive status updates by email.
- Operational statusOperational status is the healthy state indicating a component is working as expected.
- Scheduled maintenanceScheduled maintenance is planned work communicated in advance on a status page.
- Maintenance windowMaintenance window is the time range when planned maintenance is expected to run.
- Service-status historyService-status history is the historical record of component and incident states over time.
- Uptime historyUptime history is a presentation of past availability for a service or component.
- Status badgeStatus badge is a compact embeddable indicator of current status.
- Custom status-page domainCustom status-page domain is serving a status page on a domain you control such as status.example.com.
- APIAPI is an interface that lets software request work from other software over a defined contract.
- API endpointAPI endpoint is a specific URL and method that performs one API operation.
- API health checkAPI health check is a request that reports whether an API is ready to serve traffic.
- HTTP status codeHTTP status code is a three-digit code describing the result of an HTTP request.
- JSON responseJSON response is an HTTP response body encoded as JSON.
- JSON pathJSON path is an expression that selects a value inside a JSON document.
- Response headerResponse header is metadata returned by a server with an HTTP response.
- Request headerRequest header is metadata sent by a client with an HTTP request.
- WebhookWebhook is an HTTP callback that delivers an event to a URL you provide.
- Webhook payloadWebhook payload is the body of a webhook HTTP request describing an event.
- Webhook signatureWebhook signature is a cryptographic proof that a webhook came from the expected sender.
- HMACHMAC is a keyed hash used to authenticate a message such as a webhook body.
- Webhook retryWebhook retry is sending a webhook again after the receiver failed or timed out.
- Webhook idempotencyWebhook idempotency is designing receivers so duplicate webhook deliveries do not cause duplicate effects.
- Idempotency keyIdempotency key is a unique key that lets a system recognize and ignore duplicate requests.
- HTTP timeoutHTTP timeout is the client-side limit on how long to wait for an HTTP response.
- HTTP redirectHTTP redirect is an HTTP response that tells the client to continue at another URL.
- SSL certificateSSL certificate is a digital certificate commonly used to enable HTTPS for a hostname.
- TLS certificateTLS certificate is a certificate used to authenticate a host during a TLS handshake.
- SSL certificate monitoringSSL certificate monitoring is watching certificates for upcoming expiration and validity problems.
- Certificate expirationCertificate expiration is the moment after which a certificate should no longer be trusted.
- Certificate chainCertificate chain is the sequence of certificates from a server cert up to a trusted root.
- Hostname mismatchHostname mismatch is when a certificate does not cover the hostname being visited.
- TLS handshakeTLS handshake is the negotiation that establishes a secure TLS session.
- DNSDNS is the system that resolves human hostnames into addresses machines use.
- DNS resolutionDNS resolution is the process of looking up records for a hostname.
- DNS recordDNS record is a typed piece of DNS data such as A, AAAA, CNAME, or TXT.
- CNAME recordCNAME record is a DNS record that aliases one hostname to another.
- TXT recordTXT record is a DNS record that stores text, often for domain verification.
- Domain verificationDomain verification is proving control of a domain, often via DNS or HTTP challenges.
- Managed TLSManaged TLS is automatic certificate provisioning and renewal for a hostname you control.
- HTTPS certificateHTTPS certificate is the certificate presented by a site serving HTTPS.
- LatencyLatency is the delay between starting a request and observing a response milestone.
- Response timeResponse time is how long a request takes until a complete response is received.
- Time to first byteTime to first byte is the time until the client receives the first byte of the response.
- Error rateError rate is the share of requests that fail within a period.
- UptimeUptime is the time a service was available during a period.
- DowntimeDowntime is the time a service was unavailable during a period.
- AvailabilityAvailability is the share of time a service was able to fulfill its intended function.
- ReliabilityReliability is the consistency with which a service meets its expected behavior over time.
- Service availabilityService availability is availability measured for a named service or customer journey.
- Uptime percentageUptime percentage is availability expressed as a percentage of eligible monitored time.
- Mean time to detectMean time to detect is the average time from failure start to detection.
- Mean time to recoveryMean time to recovery is the average time from failure to restored service.
- Mean time between failuresMean time between failures is the average time from one failure to the next for a system.
- Service-level indicatorService-level indicator is a quantitative measure of some aspect of service level.
- Service-level objectiveService-level objective is a target value or range for a service-level indicator.
- Service-level agreementService-level agreement is a contractual commitment about service level between parties.
- Recovery time objectiveRecovery time objective is the maximum acceptable time to restore a service after disruption.
- Recovery point objectiveRecovery point objective is the maximum acceptable amount of data loss measured in time.
- Cron jobCron job is a task scheduled to run at calendar or interval expressions.
- Heartbeat monitoringHeartbeat monitoring is expecting a periodic signal from a job and alerting when the signal is late or missing.
- Heartbeat URLHeartbeat URL is the URL a job must request to prove it is still running on schedule.
- Missed heartbeatMissed heartbeat is a heartbeat that did not arrive within the allowed schedule and grace period.
- Late heartbeatLate heartbeat is a heartbeat that arrives after the expected time but may still be within grace.
- Grace periodGrace period is extra time allowed after the expected heartbeat before alerting.
- Dead man's switchDead man's switch is a control that alerts when an expected signal stops arriving.
- Scheduled taskScheduled task is work configured to run at a future time or on a repeating schedule.
- Background job monitoringBackground job monitoring is watching asynchronous jobs for timely completion and success signals.
- Audit logAudit log is a recorded history of important actions taken in an account.
- Operational readinessOperational readiness is whether people, processes, and tools are prepared to detect and handle failures.