Sentry
Publicly available source, self-hostable — but the company itself now says explicitly not to call it "Open Source" anymore. A real terminology shift worth knowing, alongside pricing that layers several separately-billed extras on top.
What is Sentry?
Sentry is a developer-first application monitoring platform built specifically for the moment production breaks: it captures exceptions, stack traces, and breadcrumbs, groups related occurrences into actionable issues, and — genuinely distinctively — surfaces the exact commit that introduced a bug, linking directly back to the source line in GitHub or GitLab through "suspect commits" identification. It's used by more than 4 million developers and 100,000+ organizations, with well-known customers including Disney, Cloudflare, Eventbrite, Slack, Supercell, and Rockstar Games, and supports more than 100 programming languages and frameworks. Beyond error tracking, it offers session replay to reconstruct failed sessions, performance monitoring tracking load times and throughput, and release tracking that connects issues directly to specific deployments. It's worth being precise about scope: Sentry is application-layer, not infrastructure-layer — it's additive to tools like Datadog or New Relic rather than a replacement for full observability stacks, and while it has log-adjacent features like breadcrumbs and extra context, it isn't a dedicated log management platform.
One specific, real, worth-knowing fact: Sentry's source code remains publicly available and self-hosting is genuinely possible, but the company has explicitly moved away from calling this "Open Source." Its own announcement introduces "Fair Source" as the new term and specifically instructs against using the old label — a real, notable licensing and positioning shift, particularly relevant if you're choosing between Sentry and a genuinely open-source alternative like Rollbar's SDKs. On pricing, Sentry uses event-based billing across four separately-metered categories — errors, performance transactions/spans, session replays, and attachments — with error events costing meaningfully more per event than transaction events (roughly 10x more). Published starting prices are Developer/Free ($0, 5,000 error events), Team (~$26-29/month, 50,000 events), and Business (~$80-89/month, 100,000 events), but several features are billed as separate add-ons beyond the base subscription: Seer AI runs $40 per active contributor per month (anyone making 2+ pull requests to a Seer-enabled repo), Continuous Profiling bills pay-as-you-go at $0.0315 per profile hour, and additional cron monitors beyond the one included cost $0.78 each monthly. Real transaction data shows a median contract around $24,151, with a genuinely wide range from roughly $11,300 to $262,800 depending on scale.
The Fair Source terminology shift is drawn directly from a detailed independent 2026 review (CubeAPM) citing Sentry's own announcement. Client list and scale figures are drawn from Capterra's aggregated data. The multi-layered pricing structure (event types, Seer AI, Continuous Profiling) is corroborated across multiple independent 2026 pricing analyses (CheckThat.ai, Last9, sentrypricing.com). Contract-value data is from Vendr's anonymized transaction dataset.
Key features
Suspect commits
Identifies the exact commit that introduced a bug, linking to the source line.
Session Replay
Reconstructs failed sessions to show the full context around an error.
Performance monitoring
Tracks load times, error rates, and throughput across your application.
Release tracking
Connects issues to specific deployments and identifies which commit fixed them.
Spike protection
Automatically throttles ingestion when volume exceeds baseline, protecting your budget.
Seer AI (add-on)
AI-assisted debugging and issue triage, billed separately per active contributor.
Pricing
Developer (Free)
- Functional for small production apps under ~1,000 DAU
- Full error tracking and stack traces included
- Good baseline before scaling to a paid tier
Team / Business
- Reserved capacity available at 20% discount vs pay-as-you-go overage
- Seer AI, Continuous Profiling, extra cron monitors billed separately
- Errors cost ~10x more per event than performance transactions
Enterprise
- Median real-world contract around $24,151/year per Vendr data
- Range spans roughly $11,300 to $262,800 depending on scale
- Worth negotiating actively given the wide real-world variance
Practical cost-reduction tips: use the before_send callback to drop irrelevant errors (cuts volume 30-60%), sample performance transactions at 10-20% rather than capturing all of them (cuts that volume 80-90%), and rely on built-in spike protection to prevent a single bad deploy from blowing your budget. Prices reflect Sentry's published pricing as of July 2026.
Available models
Integrations & platforms
Pros, cons & best for
Pros
- Suspect commits and release tracking genuinely speed up root-cause diagnosis
- Broadest language and framework support in the category
- Trusted at genuine scale by major, well-known engineering organizations
Cons
- No longer describes itself as open source, now using "Fair Source" instead
- Pricing layers several separately-billed add-ons beyond the base subscription
- Costs can scale quickly without active event-volume management
Best for
- Teams wanting the broadest language support and deepest commit-level diagnosis
- Organizations already comfortable actively managing usage-based billing
- Not the pick if genuine open-source licensing or simple flat pricing matter most
Take a look inside
Alternatives
For unlimited users on every tier or genuinely open-source SDKs instead:
Our verdict
Sentry's core diagnostic capability is genuinely excellent — suspect commits and release tracking meaningfully speed up the process of finding not just what broke, but exactly which change caused it, and the breadth of language support plus the roster of major companies trusting it at scale speaks to real, proven reliability. Two honest things worth going in aware of: the company's own shift from "Open Source" to "Fair Source" terminology is a real, specific licensing distinction worth understanding if open-source status matters to your organization's tooling choices, and the pricing structure genuinely layers multiple separately-billed components — event types, Seer AI, Continuous Profiling, extra cron monitors — on top of the base subscription, meaning the sticker price undersells the real bill for teams that don't actively manage usage. For teams wanting the deepest commit-level diagnostics and broadest framework support, and who are prepared to actively manage event volume, Sentry remains one of the strongest, most trusted options in application monitoring.
FAQ
Is Sentry still open source?
Its source remains publicly available and self-hosting is possible, but the company now explicitly uses "Fair Source" terminology and its own announcement specifically instructs against calling it Open Source — a real, notable positioning shift.
How is Sentry actually billed?
By four separately-metered event types — errors, performance transactions, session replays, and attachments — with error events costing roughly 10 times more per event than transaction events, on top of a tiered base subscription.
Are there extra costs beyond Sentry's base subscription?
Yes — Seer AI costs $40 per active contributor per month, Continuous Profiling bills pay-as-you-go at $0.0315 per profile hour, and additional cron monitors beyond the one included cost $0.78 each monthly.
How can I reduce my Sentry bill?
Use the before_send callback to drop irrelevant noise like third-party script errors (cuts volume 30-60%), sample performance transactions at 10-20% instead of capturing everything (cuts that volume 80-90%), and rely on built-in spike protection during incidents.
Is Sentry's free tier usable for a real production app?
Yes, for small projects under roughly 5,000 error events a month (typically apps with under 1,000 daily active users and reasonably good error handling), the free Developer plan is genuinely functional for production.
How does Sentry compare to Rollbar?
Sentry generally offers a broader feature set and deeper commit-level diagnostics, while Rollbar offers unlimited users on every tier (including free) and keeps its SDKs genuinely open source — the right choice depends on which trade-off matters more to your team.