Resources
Observability glossary
Monitoring and observability terms explained clearly, from alert fatigue to uptime. Each term has its own page with examples and best practices.
40 terms
A
B
- BaselineThe learned picture of normal behavior for a metric, against which deviations and anomalies are detected.
- Blue-Green DeploymentA release strategy that runs two identical environments and switches all traffic from the old (blue) to the new (green) version at once.
- Burn RateHow fast a service consumes its SLO error budget, relative to the rate that would use it up exactly at the end of the SLO window.
C
- Canary DeploymentA release strategy that exposes a new version to a small share of traffic first and expands only if it stays healthy.
- CardinalityThe number of unique values of an attribute, or unique label combinations (time series) of a metric.
- Chaos EngineeringThe practice of running controlled fault-injection experiments to find weaknesses in a system before they cause outages.
- Context PropagationPassing trace and request context across process boundaries so spans from different services join one trace.
D
E
- eBPFLinux kernel technology that runs verified, sandboxed programs on kernel and user-space events without changing kernel code.
- Error RateThe share of requests or operations that fail within a time window, one of the four golden signals.
- ExemplarA sample measurement attached to a metric data point, carrying the trace ID of a request that produced it.
F
G
H
I
L
M
N
O
- On-CallA duty rotation in which engineers are reachable outside normal work to respond to production alerts and incidents.
- OpenTelemetry CollectorA vendor-neutral service that receives, processes and exports traces, metrics and logs through configurable pipelines.
- OTLPThe OpenTelemetry Protocol, a vendor-neutral wire format for sending traces, metrics and logs over gRPC or HTTP.
P
- Percentile (p95, p99)A value below which a given share of observations falls, e.g. p99 latency: 99% of requests are at least this fast.
- PostmortemA blameless written review after an incident that documents impact, timeline, contributing factors and follow-up actions.
- PromQLThe functional query language of Prometheus for selecting, aggregating and calculating with time series data.
R
S
- SamplingKeeping only a subset of traces or events to control volume and cost while preserving the data needed for analysis.
- Semantic ConventionsOpenTelemetry's shared naming standard for attributes, span names and metrics, so telemetry means the same everywhere.
- Service MeshAn infrastructure layer of proxies that handles service-to-service traffic, security and telemetry outside application code.
- SpanA single named, timed operation within a distributed trace, such as an HTTP request, a database query or a message publish.
- Structured LoggingWriting log events as machine-readable key-value records, usually JSON, instead of free-form text lines.
T
- TelemetryThe data software and infrastructure emit about their own behavior: metrics, logs, traces, profiles and events.
- ThroughputThe amount of work a system completes per unit of time, such as requests per second, messages per second or bytes per second.
- ToilManual, repetitive, automatable operational work without lasting value that grows linearly as a service grows.
U
V
Z
No term matches your filter.
Guides
Put these concepts into practice
Try Obseria free for 14 days, with OpenTelemetry and no proprietary agents.