1. Home
  2. Documentation
  3. Observability
  4. Glossary

Glossary

Definitions of the main terms used in Sensedia Observability.

A

Term
Definition
AttributeKey and value pair that describes a trace, a span, or a
log. For example, the HTTP method or the service name.

C

Term
Definition
Collection endpointURL to which the metrics are sent.
It is dedicated to your environment.
CredentialSecret that authorizes the sending
of metrics to Sensedia Observability.

G

Term
Definition
Golden signalsSignals used to follow the health of a
service: latency, traffic, errors, and saturation.
GuardrailAI Gateway rule that evaluates AI requests and
responses and can mask or reject content.

L

Term
Definition
LatencyTime between the start and the end of a request.
LogMessage that a service records during execution.

M

Term
Definition
MCPModel Context Protocol. Standard that
allows AI agents to use external tools.
MetricNumeric measure followed over time, such
as number of requests, error rate, or latency.

O

Term
Definition
OpenTelemetryOpen standard for generating and transmitting
traces, logs, and metrics. It is the format of the
data that Sensedia Observability receives.

P

Term
Definition
Percentile (p50, p90, p95, p99)Value below which a share of the measurements
falls. A p99 of 2 seconds indicates that 99% of
the requests took up to 2 seconds.

R

Term
Definition
ResolutionTime interval represented by each point of a
chart. On the screen it appears as Resolution.
Resource attributeAttribute that describes what generated the data,
such as the service, the version, and the environment.
On the screen it appears as Resource attributes.
RetentionPeriod during which data stays available for querying.

S

Term
Definition
SeverityLevel of importance of a log: Trace, Debug, Info, Warning, Error, or Fatal.
SpanA step of a trace, with a name, start, duration,
and status. A span can have child spans.
Span IDUnique identifier of a span.

T

Term
Definition
TokenUnit of text processed by an AI model. The
consumption and cost of calls are measured in tokens.
TPOTOutput tokens per second. Measures the speed
at which the model generates the response.
TraceRecord of the path of a request through the
services, made up of one or more spans.
Trace IDUnique identifier of a trace. Links the
spans and logs of the same request.
TTFTTime to first token. Time until the model
returns the first token of the response.

How happy are you with this page?

We use cookies to enhance your experience on our site. By continuing to browse, you agree to our use of cookies.Learn more