Plain-language monitoring glossary

Monitoring basics

Monitor
A saved instruction that tells Uptime what service to check, how to check it, and what counts as healthy.

Check or observation
One attempt to reach a service from one region at one time. One failed check is evidence, but it may not become an incident.

Target
The website address, hostname, IP address, or service being checked. Public status pages should normally show a friendly component name instead of the private target.

Region
A geographic location from which a check runs. Multiple regions help distinguish a broad outage from one local internet route problem.

Interval
How often a check is scheduled. A five-minute interval means Uptime normally checks every five minutes from each selected region.

Timeout
How long Uptime waits for a response before the check fails. The timeout should reflect what your customers consider usable.

Incident policy

Failure threshold
The number of failed checks in a row required before one region is considered failing.

Regional confirmation or quorum
The number of regions that must agree before Uptime confirms an incident. “2 of 3” means any two selected regions must reach the failure threshold.

Investigation
The period when failure evidence is accumulating but confirmation has not yet been reached.

Incident
A tracked service problem with a start, state, evidence, activity timeline, and resolution.

Acknowledgment
A team member indicates that they have taken ownership of the incident. It does not mark the service healthy.

Recovery threshold
The number of successful checks in a row required before a failing region is considered stable again.

Resolved
The configured recovery conditions have been satisfied and the incident lifecycle is complete.

Availability and performance

Uptime or availability
The percentage of measured time a service satisfied its healthy policy. Always read it with the selected time range.

Response time or latency
How long a service took to reply, usually shown in milliseconds (ms). One thousand milliseconds equals one second.

Average
All response times added together and divided by the number of checks. Unusual slow checks can pull the average upward.

Median
The middle response time when results are placed in order. It represents a typical observation with less influence from extremes.

p95
A response time that 95% of observations were at or below. It helps reveal slower experiences that an average may hide.

Time bucket
A group of observations summarized together for a chart. Longer ranges use larger buckets so a 90-day chart remains fast.

Previous-period comparison
The immediately preceding range displayed for context. Comparing seven days uses the seven days before the current seven-day window.

Website checks

HTTP
The normal protocol used to request websites and web APIs. This is the best default for most website monitoring.

Status code
A short result returned by a website. 2xx generally means success, 3xx a redirect, 4xx a request or access problem, and 5xx a server problem.

Redirect
The website sends a visitor from one address to another. For example, an old page may redirect to its replacement.

Keyword assertion
A rule requiring the response to contain particular text. Use stable text that is genuinely part of a healthy response.

SSL or TLS certificate
The digital certificate that lets browsers create and trust a secure https:// connection.

DNS
The system that translates a domain name into records computers use to find websites, mail services, and other destinations.

TCP
A basic connection to a service on a specified port. It proves the connection was accepted, not necessarily that a full customer action succeeded.

ICMP or ping
A simple network reachability test. Many healthy public websites intentionally block it, so use HTTP for websites.

Communication and workspace terms

Maintenance window
A planned time range when work may affect a service. It remains visually distinct from an unexpected incident.

Status-page component
A customer-facing service name such as Checkout or Public API that can be backed by an internal monitor.

Subscriber
A customer who asks to receive status-page incident or maintenance updates and verifies their address.

Delivery attempt
One attempt to send a notification to a person or integration. Its record can be queued, delivered, skipped, retrying, or failed.

Integration
A connection between Uptime and another service, such as a Slack destination.

Workspace or organization
The shared boundary containing monitors, incidents, members, integrations, status pages, and billing.