RED Method
RED stands for Rate, Errors, Duration — the three service-level signals every request-driven service should emit. Rate is requests per second; Errors is the fraction that fail; Duration is the latency distribution. A RED dashboard answers 'is this service healthy right now' in under five seconds. TantraDev ships a RED dashboard per service before the first cutover on every cloud engagement.
Concepts that travel with this one.
Architecture rarely lives in isolation — these are the terms that come up in the same conversation.
Observability
Observability is the property of a system that lets you answer questions about its behaviour from its outputs alone, without shipping new code. The three signals are metrics, logs, and traces; the operational test is whether an on-call engineer can root-cause a novel incident from the existing dashboards. Observability is in the data model from week one on every TantraDev engagement, not bolted on after launch.
Grafana
Grafana is the de-facto dashboarding layer for production telemetry — Prometheus metrics, Loki logs, Tempo traces, and a long list of other sources rendered into one queryable surface. TantraDev ships Grafana dashboards as a deliverable on every cloud engagement, with the RED and USE dashboards built before the first cutover and the alert rules wired into the client's existing PagerDuty.
USE Method
USE — Utilisation, Saturation, Errors — is Brendan Gregg's framework for diagnosing resource-level health (CPU, memory, disk, network). Utilisation is the percent of time the resource was busy; Saturation is the queue depth waiting on it; Errors is the count of operations that failed. RED tells you *that* a service is unhealthy; USE tells you *which resource* is to blame.
Golden Signals
The Four Golden Signals, from Google's SRE book, are Latency, Traffic, Errors, and Saturation — the minimum set of signals to monitor on any user-facing service. They overlap with RED and USE but stay user-facing in framing: a latency spike that customers feel matters more than CPU saturation that they don't. TantraDev alert policies are golden-signal-shaped.
Building a system where RED Method is the load-bearing decision?
30 minutes on the phone, one page in your inbox — what to build, what to skip, what it will cost. You keep the audit even if we are not the right fit.