Self-Hosted Observability Directory

Observability TCO Calculator

Estimate SaaS observability spend against self-hosted open-source stack costs, migration effort, and break-even timing.

Use this as a planning worksheet before building a shortlist. The defaults are intentionally editable estimates; replace them with your contract, traffic, and staffing numbers before making a migration decision.

Current SaaS spend
Workload shape
Self-hosted plan

Recommended stacks for this scenario

Stack assumptions

Grafana OSS stack

Grafana with Prometheus or Mimir, Loki, Tempo, and Alertmanager. Strong default for teams that want composable pieces and familiar dashboards.

Default ops: 8h/week; migration: 140h; confidence: high.

metricslogstracesapmuptimeprofiling

SigNoz all-in-one

OpenTelemetry-native APM, traces, metrics, logs, alerts, and exceptions in one product backed by ClickHouse.

Default ops: 6h/week; migration: 120h; confidence: high.

metricslogstracesapmerrors

OpenObserve object-storage stack

Rust-based logs, metrics, traces, RUM, and dashboards designed around inexpensive object storage.

Default ops: 5h/week; migration: 110h; confidence: medium.

metricslogstracesapm

Logs-first search stack

OpenSearch Dashboards or Quickwit with Fluent Bit or Vector for teams whose bill is dominated by indexed log search.

Default ops: 7h/week; migration: 130h; confidence: medium.

logs

Errors and uptime stack

GlitchTip or Sentry self-hosted paired with Healthchecks or Uptime Kuma for smaller teams replacing Sentry plus uptime checks.

Default ops: 3h/week; migration: 60h; confidence: medium.

errorsapmuptime

How to read the estimate

  • Infrastructure costs are planning estimates: base stack cost plus host, log, trace, retention, and contingency factors.
  • Operator time is included because self-hosting is never just an infrastructure bill.
  • Break-even only appears when monthly savings are positive after infrastructure and people costs.
  • Use the copied scenario link in review docs, then replace assumptions with measured usage during a pilot.