Skip to content
Back to Guides

Sentry pricing in 2026: plans, overage rates and real monthly bills

The Sentry logo next to its pricing options: Developer, Team, Business, Enterprise and pay-as-you-go
LG

Levi Gans

·

Comparisons

Sentry pricing has two parts. A flat monthly fee buys a plan with a fixed amount of each kind of data. Anything above that costs extra, at a rate that depends on the data type, the plan and how you pay. That second part is the hard one to guess.

So what does Sentry cost? Developer is free for one person. Team is $26 to $29 a month and Business $80 to $89, before usage. In our worked example, a web app sending 250,000 errors a month pays $140 to $367 a month depending on plan and billing. The rest of this guide shows the math.

Sentry pricing plans at a glance

PlanBilled monthlyBilled annuallyUsersError retentionBuilt for
Developer$0$0130 daysOne developer on a small project
Team$29 a month$26 a month, $312 a yearUnlimited90 daysMost teams that want Slack alerts and an API
Business$89 a month$80 a month, $960 a yearUnlimited90 daysTeams that need SAML, rate limits or long span history
EnterpriseCustomCustomUnlimitedCustomCompanies that want a contract and an account manager

Sources: Sentry pricing page and data retention docs. Prices exclude sales tax.

Annual billing saves $3 a month on Team and $9 on Business. New accounts get a 14-day Business trial with unlimited events, then drop to Developer unless they upgrade. Sentry also gives new users 5,000 free replays a month for their first three months.

What each plan includes

Here’s what surprises people. Team and Business include exactly the same data. Business buys features, not volume, and charges more for extra errors and spans.

Data typeDeveloperTeam and BusinessExtra usage on Team and Business
Errors5,00050,000Reserved or pay-as-you-go
Spans5M5MReserved or pay-as-you-go
Session replays5050Reserved or pay-as-you-go
Logs5 GB5 GBPay-as-you-go only, $0.50 per GB
Application metrics5 GB5 GBPay-as-you-go only, $0.50 per GB
Attachments1 GB1 GB$0.25 per GB reserved, $0.3125 pay-as-you-go
Cron monitors11Pay-as-you-go only, $0.78 each
Uptime monitors11Pay-as-you-go only, $1.00 each
Size analysis builds100100Enterprise only
Continuous profilingNoneNone includedPay-as-you-go only, $0.0315 per hour
UI profilingNoneNone includedPay-as-you-go only, $0.25 per hour

Team and Business keep errors, replays and attachments for 90 days. Logs, spans, profiles and application metrics get 30 days on every plan, plus 13 months of downsampled span data on Business.

Team adds what a group needs: unlimited users, Slack and other integrations, the API, Google and GitHub sign-in, spend notifications, a spend cap, and access to Seer.

Business adds SAML and SCIM, Code Owners, anomaly detection, unlimited dashboards against Team’s 20, 1,000 metric monitors against 20, per-key rate limits, inbound filters by release and error message, a BAA, and Relay. Enterprise adds a technical account manager, custom quotas, and spend allocation across projects.

How Sentry billing works

Past the included quota, you either prepay reserved volume at a discount or set a pay-as-you-go budget. Unused reserved volume expires at the end of each billing month. The budget is a monthly ceiling, shared across data types first come first served, or split per type.

Pay-as-you-go costs about a quarter more, 24% to 29% above reserved rates in every tier. It’s also the only option for logs, application metrics, monitors and profiling.

What happens when you run through both? Sentry drops new data and doesn’t charge for it, which sounds fine until you read the docs: “you’ll lose monitoring for the remainder of the billing cycle.” SDKs get an HTTP 429 response and discard events until the limit lifts. That’s the part that bites.

The Developer plan has no budget at all, so at its quota it stops accepting that kind of data until the next cycle.

The smaller rules:

  • Annual plans bill the plan and reserved volume up front, and pay-as-you-go usage monthly.
  • Budget increases take effect within minutes, 24 hours at most, and you can’t lower a budget below this cycle’s usage.
  • Downgrades and reserved volume cuts wait for the end of the cycle, or the next billing year on annual plans, with no refunds.
  • Moving from Team to Business mid-cycle reprices that whole cycle’s pay-as-you-go usage at Business rates.

Sentry overage prices

Tiers are graduated, so each rate covers only the units inside its band. Sentry notes that “all per-category prices listed below have been rounded.”

Errors, beyond the included 50,000:

Monthly volumeTeam reservedTeam pay-as-you-goBusiness reservedBusiness pay-as-you-go
50k to 100k$0.00029$0.0003625$0.00089$0.0011125
100k to 500k$0.000175$0.0002188$0.0005$0.000625
500k to 10M$0.00015$0.0001875$0.0003$0.000375
10M to 20M$0.00013$0.0001625$0.00026$0.000325
Over 20M$0.00012$0.00015$0.00024$0.0003

Spans, beyond the included 5 million:

Monthly volumeTeam reservedTeam pay-as-you-goBusiness reservedBusiness pay-as-you-go
5M to 100M$0.0000016$0.000002$0.0000032$0.000004
Over 100M$0.0000014$0.0000018$0.0000029$0.0000036

Session replays, beyond the included 50, cost the same on Team and Business:

Monthly volumeReservedPay-as-you-go
50 to 5k$0.003$0.00375
5k to 100k$0.00285$0.003563
100k to 900k$0.002565$0.003206
900k to 4.5M$0.002309$0.002886
Over 4.5M$0.001962$0.002453

Source: Sentry pricing docs.

What Sentry costs at three sizes

Rate tables are hard to picture, so here are three made-up but typical profiles. The monthly columns pay as you go. The annual columns use reserved volume and show the monthly equivalent.

  • Side project. 30,000 errors, 2 million spans, 1,000 replays, 2 GB of logs, one cron monitor and one uptime monitor.
  • Startup web app. 250,000 errors, 25 million spans, 10,000 replays, 20 GB of logs, 5 cron monitors and 3 uptime monitors.
  • Growing SaaS. 3 million errors, 200 million spans, 100,000 replays, 150 GB of logs, 20 cron monitors and 10 uptime monitors.
Monthly billTeam, monthlyTeam, annualBusiness, monthlyBusiness, annual
Side project$32.56$28.85$92.56$82.85
Startup web app$168.94$140.47$367.37$305.22
Growing SaaS$1,426.76$1,159.42$2,525.49$2,050.42

Here’s the startup’s Team bill on a monthly plan. The base is $29. Errors cost $50.95, which is 50,000 at $0.0003625 plus 150,000 at $0.0002188. Spans add $40.00, replays $36.38, logs $7.50 for 15 GB over the quota, and monitors $5.12. On Business the same usage costs $367.37, mostly because errors jump to $149.38 and spans to $80.00.

Annual reserved volume saves 10% to 19% here. Business costs 1.8 to 2.9 times Team for the same usage, so pick it for SAML, rate limits or span history, not capacity.

Sentry’s calculator warns that “your actual bill may differ due to rounding,” so treat these as close estimates. Seer isn’t included.

How to keep the Sentry bill down

Some levers are server-side switches, some need a deploy, and three need Business.

  • Spike protection. Per project, Sentry sets an hourly threshold from your quota and the last seven days of traffic, and drops events above it without counting them. It covers errors, spans and attachments, not replays or logs. If high volume lasts, Sentry warns “it’ll become your new baseline.”
  • Rate limits. On Business and Enterprise, you can cap each DSN key, for example at 500 events per minute.
  • Inbound filters. Every plan can drop browser extension errors, localhost events, legacy browsers, web crawlers and chosen IP addresses. Filtering by release or error message needs Business.
  • Delete and discard. On Business and Enterprise, future events for a deleted and discarded issue don’t count. Events for ignored or resolved issues still do, which catches people out.
  • SDK sampling. sampleRate sends a fraction of errors, tracesSampleRate controls spans, and replaysSessionSampleRate and replaysOnErrorSampleRate control replays. Changes need a deploy. New customers can’t get server-side dynamic sampling.
  • SDK filters. beforeSend, beforeSendTransaction and beforeSendLog drop or trim data before it leaves the app.

Pair these with a pay-as-you-go budget you can live with, so a spike doesn’t leave you dark for the rest of the month.

Seer pricing now and after December 8, 2026

Seer is Sentry’s AI debugging agent. It finds a root cause, can open a fix as a pull request, and can hand it to Claude Code, Cursor Cloud Agents or GitHub Copilot. It needs a paid plan, works with GitHub.com and GitLab.com but not self-managed instances, and bills outside your pay-as-you-go budget.

Until December 8, 2026, Seer costs $40 a month per active contributor, meaning anyone with two or more pull requests in a month to a repository with a Seer feature turned on. On GitHub, every repository member counts except accounts marked [bot]. On GitLab, pull requests that Seer opens count too. Six engineers who each do that add $240 a month.

From December 8, Seer moves to usage-based credits in a Seer Wallet. Per Sentry’s help center:

  • Team and Business get 500 credits a month, which reset each billing cycle and don’t roll over.
  • Credits you buy carry over and expire 12 months after purchase. You can set a daily spend limit.
  • Autofix runs and Seer Agent chats use credits. Failed AI calls don’t.
  • Reading data through Sentry’s MCP server, CLI or API is free unless the call starts a Seer action.
  • Autoload top-ups are off by default, so an empty wallet stops Seer until you add credits or the next billing cycle starts.

What’s missing is the price. Sentry hasn’t published what a credit costs in dollars or how many credits an Autofix run takes, at least not anywhere we looked, and its calculator says “Seer estimates are temporarily unavailable.” You can’t predict a Seer bill from public numbers yet.

Seer Code Review ends on December 8 too, including automatic reviews and @sentry review. Remove it from CI before then, but keep the GitHub or GitLab app installed, since Autofix still needs it. If reviews are why you pay for Seer, see our Sentry Seer alternatives.

Self-hosted Sentry

You can run Sentry yourself for free. Its Functional Source License allows internal use and bars offering it as a competing product, and each release converts to Apache 2.0 two years after it’s published. Sentry calls self-hosted “the Business plan without any software limitations and no paid tier.”

The server isn’t small, though. Sentry’s minimum is 4 CPU cores, 16 GB of RAM plus 16 GB of swap, and 20 GB of free disk, with 32 GB of RAM recommended. On AWS on-demand pricing in US East, a 4 vCPU, 16 GiB m7i.xlarge costs about $147 a month and a 32 GiB r7i.xlarge about $193, before disk and backups. An errors-only mode needs at least 2 CPUs and about 7 GB of RAM.

Then someone has to install upgrades, watch disk I/O and fix it when it breaks. Sentry’s support stops at making sure it runs. The rest goes to the community Discord.

You also lose Seer and the other closed-source AI and ML features, spike protection and spend allocation, Apple’s system symbols for iOS stack traces, and some PlayStation and Switch support.

If your Sentry cloud bill is under $200 a month, the server alone costs about as much, before anyone’s time. Self-host when data has to stay on your own servers.

Sentry’s free tier compared with PostHog’s

Sentry’s Developer plan gives one user 5,000 errors and 50 replays a month. PostHog’s free plan gives unlimited team members 100,000 exceptions, 5,000 web recordings, 2,500 mobile recordings and 10 GB of logs. Both stop at the limit instead of charging you.

On free tiers alone, PostHog is far more generous. Sentry is the stronger error tool, though, and PostHog’s pricing page admits it for mobile: “Even our team thinks Sentry is better if you need mobile support.” PostHog also has no uptime monitoring. PostHog vs Sentry has a worked bill, and Sentry alternatives covers more options.

Where Interfere fits

We make Interfere, so weigh this section accordingly.

A lot of teams now work like software factories, with coding agents opening pull requests all day. Shipping stops being the hard part. Knowing what broke, who it hit and whether the fix held is the hard part now.

That’s why we built Interfere. It groups related errors into one problem with the traces, logs and session replays around it, and writes up the likely cause, tied to the release and the code. Your coding agent picks it up over MCP and writes the fix. Interfere watches production until the problem stops, and reopens it if it comes back.

If you merge more pull requests than anyone can read line by line, that loop is how you keep your sanity.

It’s also narrower than Sentry. It has SDKs for Next.js, React with Vite and TanStack Start, and takes OpenTelemetry data over OTLP/HTTP from everything else. There are no mobile SDKs, profiling, crons, uptime checks or self-hosting, and no single-region data yet. If you need those, Sentry is the better fit. Interfere vs Sentry goes deeper.

Which Sentry plan to pick

  • Solo, under 5,000 errors a month. Developer. It’s free.
  • Any team. Team, plus a modest pay-as-you-go budget.
  • SAML, SCIM, per-key rate limits or 13 months of sampled spans. Business, at about twice Team’s bill or more.
  • Steady volume. An annual plan with reserved volume.
  • Data can’t leave your servers. Self-hosted, if someone will own it.

Whichever you pick, know where your limits are. The bill to worry about isn’t the big one. It’s the month Sentry stops accepting errors and nobody notices.