What you will learn in Cloud Observability in Action
Chapter by chapter
Concepts explained in this book
- monitoringLogging is the practice of recording textual or structured events produced by a system or application.
- performanceA metric is a numerical observation, typically sampled over time, that describes a system property or behavior and may include identifying metadata and a timestamp.
- data sourceA data source is a configured backend from which an application retrieves data for querying or visualization.
- instrumentationInstrumentation is the practice of adding software mechanisms that emit telemetry about a system’s behavior and state.
- signal correlationMetric-log correlation is the association of a metric with log lines from services contributing to its aggregate value.
- OpenTelemetry CollectorThe OpenTelemetry Collector is a vendor-neutral service for receiving, processing, and exporting telemetry data.
- continuous profilingA continuous profiling tool is software that continuously collects and analyzes execution profiles from running applications or systems.
- OpenTelemetryOpenTelemetry is an open-source framework and set of APIs, SDKs, and tools for generating, collecting, and exporting telemetry.
- observabilityObservability is the capability to infer and continuously discover a system’s internal state from externally available signals in order to influence that system.
- PromQLPromQL is the query language used to select, aggregate, and calculate time-series data in Prometheus.
- OTLPThe OpenTelemetry Protocol is a telemetry data-delivery protocol defining signal encoding, transport, and delivery between sources, collectors, and destinations.
- spanA span is a timed record of an individual operation or service segment within a distributed trace.
- GrafanaA Grafana dashboard is a Grafana interface that presents selected observability data and visualizations.
- instrumentation costAn instrumentation cost is the labor, debugging effort, and resource consumption incurred by adding and maintaining telemetry instrumentation.
- service-level objectiveA service level objective is a target value or range for a service reliability or performance measure over a specified time period.
- samplingSampling is the technique of selectively capturing or discarding telemetry records according to a decision policy.
- alertingAn alert is a notification or triggered condition indicating that a monitored service or infrastructure condition requires attention.
- exporterAn exporter is a software component that exposes telemetry from an external system in a format that a monitoring system can collect.
- pprofnet/http/pprof is a Go package that exposes runtime profiling data through HTTP handlers.
- log routerA log router is software that receives, processes, and forwards log records between sources and destinations.
- service-level indicatorA service level indicator is a quantitative measurement of a service aspect relevant to evaluating a service level objective.
- processorAn OpenTelemetry processor is a Collector component that transforms, filters, batches, or otherwise manipulates telemetry within a pipeline.
- receiverAn OpenTelemetry receiver is a Collector component that accepts telemetry from an external source or protocol.
- profilingProfiling is the collection and analysis of measurements describing a program's execution behavior, resource usage, or function activity.
- columnar dataColumnar storage is an arrangement of tabular data that stores values from the same column together for efficient analytical access.
- dependency graphA service map is a graphical representation of the services in a distributed system and their communication relationships.
- eBPFeBPF is a technology for safely running sandboxed programs in the Linux kernel to observe or extend system behavior.
- context propagationContext propagation is the transmission of request-identifying context across processes and network boundaries.
- PrometheusPrometheus is an open-source monitoring system and time-series database for collecting and querying metrics.
- B2I ratioThe B2I ratio is the number of lines of code after instrumentation divided by the number of lines of code before instrumentation.
- flame graphA flame graph is a hierarchical visualization of collected stack traces in which horizontal extent represents call frequency or resource consumption and vertical position represents stack depth.
- cardinality explosionCardinality explosion is a rapid and potentially unbounded increase in the number of metric time series caused by many distinct dimension values or combinations.
- Prometheus exposition formatThe Prometheus exposition format is a text-based format for representing metrics and their labels for collection by monitoring systems.
- signalA signal is information about a system that can be observed externally.
- FluentBitFluent Bit is a lightweight open-source agent for collecting, processing, and forwarding logs and other telemetry.
- error budgetAn error budget is the allowable amount of service unreliability remaining under a service level objective during its measurement period.
- sequenceA time series is an ordered sequence of values associated with successive timestamps and a particular metric and dimension set.
- KubernetesKubernetes is an open-source system for orchestrating containerized workloads and managing their deployment, scaling, and operation.
- dashboardA dashboard is a visual interface that presents selected measurements or status information for monitoring and analysis.
- dashboard variableA dashboard variable is a configurable dashboard value that users can select to parameterize queries or other dashboard elements.
- alert fatigueAlert fatigue is the reduced ability to notice or appropriately respond to important alerts because of excessive notifications.
- AlertmanagerAlertmanager is a Prometheus component that groups, routes, silences, and suppresses alerts.
- microservice architectureA microservices application is an application composed of independently deployable services that communicate over interfaces.
- Prometheus agent modePrometheus agent mode is a Prometheus configuration optimized for scraping metrics and remotely reading and writing them for federation.
- anomaly detectionAnomaly detection is the identification of observations or behavior that significantly deviates from an established normal baseline.
- unified destinationAn all-in-one observability tool integrates multiple telemetry collection, storage, analysis, visualization, or alerting capabilities in a single product.
- logging levelA log level is a classification that indicates the nature or severity of a log event.
- analytical processingOnline analytical processing is a workload pattern involving complex queries over large volumes of historical data, typically with fewer writes than transactional systems.
- ClickHouseClickHouse is an open-source column-oriented database management system designed for analytical query workloads.
- grouping and throttlingGrouping and throttling is the consolidation and rate limiting of alert notifications to reduce redundant delivery.
- inhibitionInhibition is the suppression of notifications for alerts when related alerts are already firing.
- notificationA notification is a message delivered to a person or system about an alert or monitored condition.
- notification pipelineA notification pipeline is an ordered sequence of alert-processing phases that suppresses, routes, groups, and delivers notifications.
- silencingSilencing is the deliberate suppression of notifications for selected alerts, typically for a known maintenance or operational condition.
- open standardAn open standard is a publicly available specification developed through an open process and implementable without exclusive control by a single vendor.
- pprof formatA profiler data format is a representation specification for storing or exchanging execution-profile data.
- baselineA baseline is a reference value or range representing normal system behavior against which observations are compared.
- incidentAn incident is unexpected behavior of a service or platform that negatively affects its function or performance, including a recurrence for which prior learnings have not yet been applied.
- telemetryTelemetry is the process of collecting signals from sources, processing them, and ingesting them into destinations for consumption.
- semantic conventionA semantic convention is a standardized name and meaning for telemetry attributes, events, and other observability data.
- backpressureBackpressure is a flow-control mechanism in which a downstream component signals upstream producers to reduce or regulate data transmission when capacity is constrained.
- recording ruleA recording rule is a Prometheus rule that precomputes an expression and stores its result as a new time series.
- OpenSearch DashboardsOpenSearch Dashboards is an open-source web interface for searching, analyzing, and visualizing data stored in OpenSearch.
- correlationTelemetry correlation is the linking of related observability signals so users can move between corresponding records and views.
- counterA counter is a metric type whose value represents a quantity that generally increases over time, apart from resets.
- service-level agreementA service level agreement is a contractual commitment by a service provider specifying promised service performance or availability and possible customer remedies for failing to meet it.
- correlation pathA correlation path is a sequence of transitions between signal types whose observations describe the same system behavior or request.
- JaegerJaeger is an open-source distributed tracing system for collecting, storing, querying, and visualizing trace data.
- protocol bufferProtocol Buffers is a language-neutral, platform-neutral mechanism for serializing structured data using schema-defined messages.
- protocprotoc is the Protocol Buffers compiler and decoder for compiling .proto definitions and serializing or deserializing protobuf messages.
- stack profileA stack profile is a collected representation of call-stack samples and their associated execution or resource measurements.
- log driverA logging driver is a runtime component that transfers container-generated log records to a configured logging destination or collector.
- GoGo is a statically typed compiled programming language designed for efficient systems and network software development.
- request pathA request path is the sequence of components and service interactions through which an individual request travels.
- head-based samplingHead-based sampling is a tracing sampling strategy that makes a capture decision before a trace is complete.
- tail-based samplingTail-based sampling is a tracing sampling strategy that defers the capture decision until a trace is complete.
- developer observabilityDeveloper observability is the practice of providing developers with telemetry-derived insight into code changes and their effects on applications.
- Jaeger UIThe Jaeger UI is a web-based interface for searching, analyzing, and visualizing traces.
- root spanA root span is the first span representing the start of a distributed trace's request processing.
- observability systemAn observability solution is an integrated product or service for collecting, analyzing, visualizing, and acting on telemetry from software systems.
- burn rateA burn rate is the rate at which a service consumes its error budget relative to the permitted rate.
- ingestion costAn ingestion cost is a charge for receiving or writing observability signals into a backend.
- query costA query cost is a charge for processing or scanning stored observability data to answer a query.
- waterfall visualizationA waterfall visualization is a time-aligned display showing the durations and ordering of activities or service executions.
- OpenTelemetry SDKThe OpenTelemetry Node SDK is a JavaScript or Node.js implementation for generating, processing, and exporting OpenTelemetry telemetry.
- XOR compressionXOR compression is a compression technique that uses bitwise exclusive-or differences between values to encode related data compactly.
- ParcaParca is an open-source continuous profiling platform for collecting and analyzing profiles of running software.
- build-versus-buy decisionA build-or-buy decision is a strategic choice between developing a capability internally and procuring it from an external provider.
- LokiGrafana Loki is a log aggregation system that indexes log labels while retaining the log content for retrieval.
- debuggingDebugging is the systematic process of locating, diagnosing, and correcting defects in software.
- eventAn event is an occurrence of interest in code, including expected behavior and unexpected or erroneous behavior.
- serverless computingA serverless environment is an execution environment in which the provider manages the underlying servers and infrastructure for application code.
- OpenMetricsOpenMetrics is an open standard defining a text-based format for exposing metrics and associated metadata such as exemplars.
- time-series databaseA time series database is a database optimized for storing and querying values indexed by time.
- data gravityData gravity is the tendency for large or important data stores to attract related applications, services, and infrastructure toward their environment.
- control planeThe Kubernetes control plane is the set of Kubernetes components that expose the cluster API and maintain the desired state of cluster resources.
- data planeThe data plane is the set of worker resources that execute application workloads and carry operational traffic in a distributed system.
- Kubernetes APIThe Kubernetes API is the programmatic interface for querying and changing the state of Kubernetes resources.
- data temperatureData temperature is a classification of data based on how frequently and how quickly it must be accessed.
- full-text search indexFull-text indexing is the creation of a searchable index over the textual content of documents or log records.
- label-based indexingLabel-based indexing is the creation of an index over selected labels or tags that describe log context rather than the full log text.
- log pruningLog pruning is the selective removal or reduction of log events before storage or further processing.
- correlation idA trace identifier is the identifier shared by spans belonging to one distributed request trace.
- vertical scalingVertical scaling is the practice of increasing the resources of an existing system instance, such as memory or storage.
- cardinalityMetric cardinality is the number of distinct values or combinations represented by a metric’s dimensions or labels.
- fluentdFluentd is an open-source data collector that unifies the collection and forwarding of logs and other machine data.
- forward input pluginThe Forward input plugin is a Fluent Bit input component that receives log records through the Fluentd Forward protocol.
- logging pipelineA logging pipeline is a sequence of components that collects, routes, processes, stores, and presents log records.
- PixiePixie is a Kubernetes-specific, real-time debugging platform that collects and analyzes telemetry without requiring manual instrumentation or application redeploys.
- AWS FargateAWS Fargate is a managed AWS compute engine that runs containers without requiring users to manage the underlying servers.
- icicle chartAn Icicle Graph is Parca's term for a flame-graph-style hierarchical visualization of profile data.
- OpenTelemetry APIThe OpenTelemetry API is the language-facing interface for creating and accessing OpenTelemetry telemetry instrumentation components.
- resource attributeA resource attribute is metadata describing the entity or infrastructure resource that produced telemetry.
- CortexCortex is a horizontally scalable, highly available metrics storage system compatible with Prometheus.
- journalctljournalctl is the command-line interface for querying and displaying logs stored by the systemd journal.
- JVMThe Java virtual machine is the runtime environment that executes Java bytecode and manages resources such as memory and threads.
- systemdsystemd is a Linux system and service manager that includes a journal for collecting and storing system logs.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.