liveBook · Manning

Time series

As explained in Designing AI Systems

A time series is an ordered sequence of timestamped values showing how measurements change over time.

A time series stores numeric measurements over time so systems can reveal trends and aggregate behavior that individual traces cannot. In observability, services record metrics independently as requests flow through the system, then store them for later analysis. section 7.2.3

Time-series metrics support operational questions such as:

  • Whether p99 latency has climbed steadily during the past several hours. section 7.2.3
  • Whether error rates spiked during a particular period. section 7.2.3
  • How request rates, latency distributions, and error counts change across aggregation windows. section 7.3.1

Metrics differ from logs and traces in both storage and query behavior. Logs remain individual searchable records, while metrics form numeric series queried by aggregation window; keeping separate ingest and query operations preserves those distinctions. section 7.3.1

Observability workflow

A service records request-rate, latency, and error metrics while handling requests. The observability platform stores the resulting series, then exposes queries for dashboards and alerting rules. A trace explains one slow request; the corresponding time series shows whether slow requests are becoming a broader operational problem. section 7.2.3 section 7.3.1

How Designing AI Systems covers it

How deep it goes
Defines metric aggregation and contrasts storage and query behavior with logs, without detailing time-series storage internals.
Examples
Uses request rates, latency distributions, error counts, p99 latency trends, and error-rate spikes.
What it assumes
Assumes familiarity with requests, traces, logs, and observability services.
What you can do after
Evaluate/understand system-wide trends and distinguish isolated trace anomalies from broader operational problems.
How it connects
Connects metrics with observability, logs, traces, dashboards, and alerting.

Also explained in

Compare how 10 books explain time series →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Designing AI Systems. The section links point to the parts of the book it draws on.