You are not alone. Sentry is designed for enterprise budgets. If you are a startup or small team, there is a simpler option.
Sentry Team is $26/mo for 50k errors. Business is $80/mo for 100k. Volume pricing above that. Self-hosted Sentry requires Kafka, Clickhouse, Postgres, and Redis — a significant infrastructure commitment.
For an enterprise team with a large budget and complex requirements, Sentry is probably worth it. For everyone else, the math does not work.
Seismograph is a self-hosted error tracking tool. Single Go binary, embedded SQLite, running in 30 seconds. Free tier: 5,000 errors/mo. Pro: $0.99/mo/mo.
curl -fsSL https://stockyard.dev/seismograph/install.sh | sh
| Seismograph | Sentry | |
|---|---|---|
| Pricing | $0.99/mo/mo flat | $26/mo |
| Hosting | Self-hosted | Cloud only |
| Free tier | 5,000 errors/mo | Limited |
| Setup time | 30 seconds | Account + config |
| Dependencies | None (single binary) | N/A (managed) |
| Data location | Your server | Their cloud |
Seismograph does not have performance monitoring, distributed tracing, session replay, or profiling. It does not have Sentry's 50+ SDK integrations. If you need full application observability, Sentry is the more complete platform.
Free tier included. No credit card required.
Sentry's Developer plan is free for one user with 5,000 events per month. Most production applications exceed that within the first week. The Pro plan at $26 per month covers 50,000 events — enough for a low-traffic service, but a single uncaught exception in a loop can burn through that quota in hours. The Business plan at $80 per month provides 100,000 events and performance monitoring. The real trap is event volume: Sentry counts every error occurrence individually, and high-traffic applications routinely generate millions of events during an incident.
Seismograph approaches the same problem differently. It captures error events, deduplicates them by signature, and groups related errors together. There is no event volume limit because the data stays on your disk. During an incident, when error volume spikes by 100x, Seismograph does not throttle your events or send you an overage invoice — it just logs them. The SQLite database grows by a few megabytes, and you deal with the errors instead of the billing page.