# Time Series

A time series is a sequence of observations indexed by time that represents how a measured quantity changes.

Cloud concept. Updated 2026-10-29. Page: https://blowstack.com/tools/cloud-concepts-hub/time-series

> Time series organize measurements such as CPU utilization, request counts, and response latency so engineers can analyze trends, detect anomalies, and evaluate operational thresholds. Effective cloud monitoring depends on identifying each series correctly, choosing suitable collection and aggregation intervals, and balancing historical detail against storage and query costs.

Practice exams covering Time Series: [DOP-C02 AWS Certified DevOps Engineer - Professional](https://blowstack.com/practice-exams/aws-certified-devops-engineer-professional).

## 6 key aspects of Time Series

Reliable time-series monitoring combines stable identities, appropriate temporal granularity, meaningful statistics, controlled cardinality, and explicit retention and missing-data policies.

1. **Timestamped observations**: Every observation is associated with a time, allowing comparisons across intervals and correlation with deployments or incidents. Use consistent timestamp conventions and synchronized clocks to reduce misleading ordering and alignment.
2. **Explicit series identity**: A metric name and its identifying attributes distinguish one series from another. In CloudWatch, namespaces and complete dimension sets are central to that identity within an account and Region.
3. **Resolution and collection cadence**: Fine-grained storage only helps when producers supply sufficiently frequent observations. Choose collection cadence, storage resolution, and alarm evaluation periods together so short incidents are observable.
4. **Meaningful aggregation**: Choose statistics according to the measurement: sums for interval counts, rates for cumulative counters, and averages or percentiles for suitable distributions. Aggregation across time and across resources must preserve the intended meaning.
5. **Cardinality management**: Every distinct identifying attribute combination can create another series. Use bounded operational dimensions and keep high-cardinality event identifiers in logs unless the monitoring backend and budget explicitly support them.
6. **Retention and missing-data semantics**: Downsampling reduces storage requirements but removes historical detail. Treat absent samples separately from measured zero values, especially when configuring alarms and diagnosing collector failures.

## Cloud services for Time Series

CloudWatch collects AWS operational metrics and custom measurements for dashboards and alarms, while Amazon Managed Service for Prometheus supports Prometheus-based collection and querying. Amazon Managed Grafana can present both sources in shared dashboards, helping engineers compare application and infrastructure behavior. Amazon S3 can retain exported or independently collected measurements when historical analysis requires a different retention policy. Across these systems, consistent timestamps, compatible units, and deliberate dimension or label mappings are necessary; displaying series together does not automatically make their aggregation statistically valid.

- **[Amazon CloudWatch](https://blowstack.com/tools/cloud-services-hub/amazon-cloudwatch)**: Stores operational metrics as time series identified by namespaces, metric names, and dimensions, and evaluates them through statistics, dashboards, metric math, and alarms. Custom metrics support standard or high resolution, while retention progressively reduces historical granularity.
- **[Amazon Managed Service for Prometheus](https://blowstack.com/tools/cloud-services-hub/amazon-managed-service-for-prometheus)**: Stores and queries Prometheus-compatible metric time series identified by metric names and labels. PromQL supports operations such as counter rates and aggregation across selected labels, making label design and cardinality control important.
- **[Amazon Managed Grafana](https://blowstack.com/tools/cloud-services-hub/amazon-managed-grafana)**: Queries data sources such as CloudWatch and Amazon Managed Service for Prometheus to visualize time series in operational dashboards. It provides a visualization and analysis layer rather than serving as the primary metric storage backend.
- **[Amazon S3](https://blowstack.com/tools/cloud-services-hub/amazon-s3)**: Can retain exported measurements or independently collected time-series datasets for longer-term analysis. Objects require a separate query or processing layer and are not directly equivalent to an operational metric store with built-in alarm evaluation.

## Time Series exam questions

Real practice exam questions on Time Series, with the correct answer and its explanation.

### Question 1 (DOP-C02)

**Question:**

A company operates a custom monitoring collector at a remote processing facility. Every minute, the collector records a queue-depth measurement and its UTC observation timestamp. It publishes the measurements to Amazon CloudWatch under the namespace Company/Processing, metric name QueueDepth, and dimensions Facility=west and Queue=priority.

After a 6-hour network outage, the collector has a durable backlog of measurements. Multiple upload workers will replay the backlog while another worker publishes current measurements, so CloudWatch can receive observations out of chronological order. All buffered observations are less than 12 hours old.

A DevOps engineer must ensure that, after CloudWatch makes the historical data available, the existing dashboard shows the backlog measurements in the minutes when they were observed, alongside the current measurements in the same metric. The solution must not require another data store or changes to the dashboard's metric selection.

Which publishing configuration meets these requirements?

**Correct answer:** Publish each measurement with its original observation time in the PutMetricData Timestamp field. Keep the existing namespace, metric name, and complete dimension set for both buffered and current measurements.

**Explanation:** CloudWatch places a datum according to its supplied timestamp, not its arrival order. The original observation timestamps place buffered measurements in their historical intervals. Keeping the namespace, metric name, and dimensions unchanged makes buffered and current measurements part of the same metric. These timestamps are within CloudWatch's supported window for publishing historical data, although historical data can take additional time to become available.

## Practice Time Series in a full exam

Questions on Time Series in each practice exam, with every answer option and an explanation for each:

- [DOP-C02](https://blowstack.com/practice-exams/aws-certified-devops-engineer-professional): 3 questions

## What is Time Series

A time series associates observations with timestamps. In operational monitoring, each observation usually contains a numeric value, while identifying attributes describe what was measured and where it originated. For example, CPU utilization sampled every minute for one virtual machine forms a time series. The same measurement for another virtual machine forms a different series. Samples can arrive regularly or irregularly; evenly spaced measurements are common, but they are not required by the concept.

Series identity is fundamental. Monitoring systems typically combine a metric name with attributes such as service, instance, environment, or region. In Amazon CloudWatch, within an AWS account and Region, a metric is uniquely identified by its namespace, metric name, and complete set of dimensions. A namespace groups related metrics, while dimensions are name-value pairs that distinguish series. For example, a custom namespace named Checkout, a metric named RequestLatency, and dimensions Environment=production and Operation=payment identify one metric. Changing or adding a dimension creates a different metric rather than updating descriptive metadata on the existing one.

This distinction affects aggregation. CloudWatch does not automatically aggregate custom metrics across dimension combinations. Publishing separate series for individual instances does not automatically create a dimensionless fleet-wide metric. Engineers must explicitly publish suitable aggregate metrics or query the individual series using supported mechanisms such as metric math or Metrics Insights. Some AWS services publish metrics with multiple dimension combinations, but this behavior should not be assumed for custom metrics.

Collection interval, storage resolution, and query period describe different choices. Collection interval determines how often a producer measures or sends data. Storage resolution determines the supported granularity of stored measurements. A query period groups observations into time windows for statistics. CloudWatch custom metrics support standard resolution of 60 seconds and high resolution of 1 second, selected through StorageResolution. Selecting high resolution does not make a producer collect more frequently; the producer must actually supply finer-grained measurements.

Aggregation changes what a chart or alarm reveals. A five-minute average can hide a ten-second CPU spike, whereas a maximum can expose it. For request counts published as interval totals, Sum usually answers how many requests occurred. For a cumulative counter, calculate an appropriate increase or rate while accounting for resets rather than summing successive counter readings. Latency percentiles can expose slow requests that averages conceal, but an average of instance-level percentiles is not a valid fleet-wide percentile. Preserve raw observations or a suitable distribution representation when percentile analysis is required.

Retention also changes available detail. CloudWatch retains data points with periods below 60 seconds for three hours, one-minute data for 15 days, five-minute data for 63 days, and one-hour data for 455 days. Older data is aggregated into coarser periods, so a long-term graph cannot recover the original second-level behavior. Export or independently retain measurements when investigations require longer access to fine-grained history.

Dimension design introduces a cost and scalability trade-off called cardinality: the number of distinct series. Stable attributes such as deployment environment and operation are often useful, while request IDs or unbounded customer identifiers can create excessive numbers of metrics. Keep detailed event context in logs and extract bounded operational measurements for metrics. Finally, missing data is not automatically zero: it might indicate an idle workload, a failed collector, delayed delivery, or a stopped resource. Align timestamps, synchronize clocks, and configure alarm missing-data behavior according to the measurement's meaning.

## Learn Time Series in lessons

1 short lesson on Time Series in 1 certification course.

- DOP-C02: Time Series: modeling CloudWatch measurements (Premium), https://app.blowstack.com/lessons/time-series-modeling-cloudwatch-measurements-dop-c02
