Lexicon
Observability
Observability is the ability to understand a system's internal state from its outputs. How it differs from monitoring, and how the practice works day to day.
Definition
Observability is the ability to understand what is happening inside a system from the data it emits: its logs, metrics, and traces. The term names both the property, how understandable a system is from the outside, and the engineering practice of instrumenting and analyzing software to achieve it.
What it means in observability
The classic contrast is with monitoring. Monitoring watches for known failure modes: you decide in advance what can go wrong and set alerts on it. Observability prepares for unknown ones: it captures rich enough telemetry that questions nobody predicted can still be answered after the fact. The distinction earned its keep when architectures went distributed. In a system of hundreds of services, most incidents are novel combinations that no predefined dashboard anticipated, and the ability to interrogate fresh questions against detailed data became the difference between minutes and days of downtime.
How it works in practice
The practice is a loop. Instrumentation makes systems emit telemetry, increasingly through OpenTelemetry. Pipelines collect, shape, and route it. A backend stores each signal for its access pattern and correlates them through shared context. Engineers then work the loop from both ends: alerting on symptoms that matter, and exploring interactively when something looks wrong, pivoting from a metric anomaly to the traces behind it to the logs that explain it.
Where it gets hard
The modern problems are economic and structural more than technical. Telemetry volume compounds faster than budgets, pushing teams into sampling and retention cuts that shrink the very coverage observability depends on. Tooling sprawls across signals and vendors, leaving correlation gaps that engineers bridge by hand at 3am. And the data itself, among a company's most sensitive, typically leaves the building for a vendor's cloud.
Where Tsuga fits
Tsuga was built for these failure modes: one OpenTelemetry native platform for all three signals, flat per GB pricing that removes the pressure to sample, and a BYOC architecture that keeps telemetry inside your own cloud account.
Related terms
- AIOpsAIOps is the application of machine learning and artificial intelligence to IT operations: detecting anomalies, correlating and grouping alerts, suggesting root causes, and increasingly, powering assistants that investigate telemetry conversationally.
- APMAPM, application performance monitoring, is the practice and product category focused on the health of applications in production: request rates, error rates, latency, and the transaction level detail needed to explain them.
- Data residencyData residency is the question of where data physically lives: which country or region stores it, and where it is processed along the way.
- Data sovereigntyData sovereignty is the question of whose laws govern your data and who can be compelled to grant access to it.
- Distributed tracingDistributed tracing is the technique of following a request across service boundaries by recording a connected span for each operation it touches.
- Elastic ObservabilityElastic offers observability built on the Elasticsearch stack: log search at its core, with metrics, traces, and APM layered on, available self managed or as the hosted Elastic Cloud.
- Error budgetAn error budget is the amount of unreliability an SLO permits: one hundred percent minus the target.
- GrafanaGrafana is the open source standard for dashboards, and Grafana Labs builds an observability stack around it: Loki for logs, Tempo for traces, Mimir for metrics, with Grafana Cloud as the managed offering.
- Infrastructure monitoringInfrastructure monitoring is the practice of tracking the health, performance, and capacity of the layer applications run on: hosts, virtual machines, containers, Kubernetes clusters, and managed cloud services.
- Log managementLog management is the practice of collecting, processing, storing, and searching log data at scale: every event record your applications and infrastructure emit, made findable when someone needs it.
- MTTRMTTR is mean time to resolution, the average elapsed time from a failure occurring to service being restored.
- OpenTelemetryOpenTelemetry is an open source framework for generating, collecting, and exporting telemetry: the logs, metrics, and traces that describe how software behaves in production.
- TelemetryTelemetry is the data software and infrastructure emit about their own behavior.