Lexicon
APM
APM means application performance monitoring: tracking request rates, errors, and latency per service. What it covers and how it relates to observability.
Definition
APM, 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. It is the lens that watches the software itself rather than the infrastructure underneath it.
What it means in observability
APM predates the observability vocabulary and shaped how a generation of teams worked: install the vendor's agent, get automatic visibility into transactions, databases, and errors. The category's core outcomes remain exactly what most engineering teams need day to day, per service and per endpoint views of the golden signals, with the ability to drill from a slow endpoint into the requests behind it. What changed is the foundation: the proprietary agents that defined classic APM are giving way to OpenTelemetry, which delivers the same visibility on open instrumentation.
How it works in practice
Modern APM is assembled from telemetry rather than bolted on by an agent. Traces provide the transaction detail, span derived metrics supply request, error, and duration numbers per service and endpoint, and error tracking groups failures into actionable patterns. Service catalogs and dependency maps come from the same trace data. Dashboards and alerts sit on top, watching latency percentiles and error rates against baselines or thresholds.
Where it gets hard
The classic APM model carried costs beyond the license: proprietary agents to deploy and upgrade everywhere, per host pricing that container platforms turned punitive, and instrumentation that locked telemetry to one vendor. The terminology also confuses buying decisions, since APM, observability, and monitoring overlap heavily and vendors draw the lines wherever it flatters their product.
Where Tsuga fits
Tsuga delivers the APM outcomes, service pages, endpoint and operation analytics, latency percentiles, error patterns, and deployment tracking, on stock OpenTelemetry instrumentation. No agents to run, and pricing is flat per GB rather than per host.
Related terms
- DatadogDatadog is the largest SaaS observability platform, spanning infrastructure monitoring, APM, logs, RUM, security, and dozens of adjacent products, collected largely through its proprietary agent and priced per product.
- Distributed tracingDistributed tracing is the technique of following a request across service boundaries by recording a connected span for each operation it touches.
- DynatraceDynatrace is an enterprise observability platform known for its OneAgent automatic instrumentation, the Davis AI engine for root cause analysis, and the Grail data lakehouse, sold on a consumption based pricing model.
- MTTRMTTR is mean time to resolution, the average elapsed time from a failure occurring to service being restored.
- New RelicNew Relic is one of the original APM vendors, now a broad all in one observability SaaS covering APM, infrastructure, logs, browser monitoring, and more, priced on data ingested plus per user seats.
- ObservabilityObservability is the ability to understand what is happening inside a system from the data it emits: its logs, metrics, and traces.
- TraceA trace is the end to end record of one request or workflow as it moves through a system, composed of all the spans that share a single trace ID.