Sentry alternatives: where to send your errors, and how to switch without re-instrumenting
Six tools ranked, what each costs at 100,000 and 1 million errors a month, and a seven-step migration plan.
The short answer: GlitchTip if you want open source, hosting in Germany and your Sentry SDKs untouched, from $15 a month for 100,000 events or free on your own server. Bugsink if you want the lightest self-hosted setup: one container. Better Stack if volume is your problem: 100,000 exceptions free each month, then $0.00005 per exception in its Europe region. PostHog if your product team already runs on it. Honeybadger for a solo dev or agency that wants errors, uptime and a status page in one plan. AppSignal if what you really want is a full performance monitoring tool, billed per request.
Two things to know before you move. GlitchTip, Bugsink and Better Stack accept Sentry SDKs: you change the DSN (the address the SDK sends errors to) and your code stays put. Your error history, though, stays in Sentry. We compared the six tools in this guide and went through their official pricing pages in 2026.
Rankings are editorial: a sponsored spot is always labeled on its card, and no affiliate link changes a score.
Why teams leave Sentry, and what keeps them
Sentry is an error tracker: a small library (the SDK) inside your app sends it every crash, with the stack trace that points to the failing line. Plenty of teams joined on the free plan and never left. When they do go, it's rarely about the product. It's about the bill, the pain of self-hosting, and where the data lives. Six reasons keep coming up.
The entry price isn't the bill
Team costs $26 a month billed yearly ($29 monthly) with 50,000 errors, 5 million spans (the steps of a performance trace), 50 replays and 5 GB of logs. Every meter beyond that is billed separately: $0.50 per extra GB of logs, $0.78 per cron monitor, $0.25 per hour of UI profiling. One developer on Hacker News describes a bill in the "$3-5000 range just from the tracing". The price on the page covers errors, not your actual usage.
Run out of quota and errors get dropped
Once your reserved volume and pay-as-you-go budget are used up, Sentry drops further events and doesn't bill them. You're blind until the cycle ends, often right when an error loop needs watching most. The pay-as-you-go budget is shared across categories, first come, first served. And reserved volume you didn't use is gone at month end.
Spike protection drops events too
Spike protection sets a per-project threshold and drops events above it to save your quota. It's off during the trial, so you first meet it in production, and its notifications need turning on separately. Jobs that burst by design (a cron, an Airflow run) get cut. Sentry's own docs suggest disabling it for them.
Self-hosting is heavy
Self-hosted Sentry needs 4 cores, 16 GB of RAM plus 16 GB of swap (32 GB recommended), and runs Kafka, ClickHouse, Postgres, Redis and a dozen more services. Sentry itself describes it as packaged for low-volume deployments and proofs of concept, with no guaranteed support. One developer puts it plainly: "It also doesn't make sense to spin up a 16GB machine (minimum!) to track the errors on those 4-8GB VPS which are running my production services."
The license isn't open source
Sentry ships under the Functional Source License (FSL). You can use, modify and deploy it internally, but you can't resell hosted Sentry or compete with it. Each release converts to Apache 2.0 or MIT after two years. It's not an OSI open source license, which is a blocker for some company policies.
The EU region doesn't cover everything
You pick the region (US or Frankfurt) when you create the organization, and you can't change it later: you'd need a new organization, with nothing moved over. Even in the EU region, accounts, org settings, project metadata, DSN keys, audit logs and SSO data stay in the US. As one Hacker News commenter put it: "I was looking into moving away from managed Sentry for GDPR reasons."
You're leaving because… go to…
The right Sentry alternative depends on what's pushing you out. Usually one or two rows apply. Each card jumps to the full write-up below.
MIT license, German instance, $15 a month for 100,000 events.
Self-hosted Sentry wants a 16 GB serverBugsinkA single container, free on your own server.
Your bill climbs with error volumeBetter Stack100,000 exceptions free, then $0.00005 each in Europe.
Your product team already uses PostHogPostHogErrors, sessions and analytics in one tool.
You're solo or an agency and want something simpleHoneybadgerErrors, uptime and a status page in one plan.
Application performance is your real concernAppSignalFull performance monitoring, unlimited users.
Sentry alternatives, tool by tool
For each tool: why move to it, who it's for, its limit, how to migrate from Sentry, and what it costs. Prices come from the official pricing pages in 2026, in US dollars.
Three tools sit outside the ranking. Datadog Error Tracking also accepts Sentry SDKs and costs $25 a month billed yearly under 50,000 errors, which only makes sense if you're already on Datadog (and if Datadog is what you want to leave, see our Datadog alternatives guide). Rollbar and Bugsnag have free plans (5,000 occurrences and 7,500 events), but their paid plans don't show a readable price, and we don't rank a tool without one. And Highlight.io, still recommended in plenty of listicles, has shut down: its site now redirects to LaunchDarkly (see our LaunchDarkly alternatives guide).
GlitchTip: open source that speaks Sentry
Closest to SentryWhy move to GlitchTip. GlitchTip accepts any Sentry-compatible SDK: swap the DSN and your errors show up. The code is MIT-licensed, real open source. Hosted, the Small plan costs $15 a month for 100,000 events and 90 days of history, and a Germany-based instance is available on every plan. Self-hosted, it's free and lean: Postgres 14+, optional Redis or Valkey, 512 MB of RAM recommended, roughly 30 GB of disk per million events a month.
Who it's for. Teams leaving over licensing or GDPR, and developers who want to self-host without running Kafka. One Hacker News user: "Super easy to self-host, and because it uses the existing sentry libs it's really reliable."
Who it's not for. A team that lives in Sentry's session replays and profiling: GlitchTip does less. Check every feature you rely on exists before you cut over.
Its limit. Fewer features, and not everyone found setup easy: "A bit hard to deploy (many moving parts)". Nonprofits and open source projects get plans from $5 a month.
How to migrate. Create the project in GlitchTip, paste its DSN into the environment variable that holds Sentry's, redeploy. Overage is optional ($0.17 per 1,000 events up to 1 million) and can be capped: turn it on if you'd rather pay a bit more than lose errors.
- Free
- 1,000 events a month, or self-hosted
- Paid
- Small $15/mo (100,000)
- Sentry SDKs
- accepted
- Limit
- fewer features than Sentry
Bugsink: one container instead of a factory
Lightest self-hosted optionWhy move to Bugsink. Bugsink runs in a single container and takes errors from the open source Sentry SDKs: you change the DSN. Self-hosted, it's free with unlimited users. The managed version, hosted in the EU, is free for 15,000 events a month and one user, then $16 a month for 75,000, $50 for 600,000 and $158 for 3 million.
Who it's for. A developer or small team that wants to self-host error tracking on a small box next to their services, without giving it someone's week.
Who it's not for. A team that wants performance traces: Bugsink tracks errors and nothing else. Or a company that wants to offer error tracking as a service: the license forbids it.
Its limit. The PolyForm Shield license gives you the source but bars competing with the vendor. One commenter sums it up: "Polyform Shield, basically use-but-don't-compete". And in fairness, the most-read post against self-hosted Sentry was written by Bugsink's author, which a Hacker News reader called out as "reusing Sentry's open-source client SDK's, while spreading FUD".
How to migrate. Start the container, create a project, swap the DSN. If you want support on the self-hosted version, it's $15 per user a month.
- Free
- self-hosted, or 15,000 managed events
- Paid
- $16/mo (75,000), $50 (600,000)
- Sentry SDKs
- accepted
- Limit
- errors only, no traces
Better Stack: your Sentry SDKs, billed per exception
Cheapest at volumeWhy move to Better Stack. Better Stack bills itself as Sentry-compatible and takes your SDKs once you change the DSN. The first 100,000 exceptions each month are free. After that, each one costs $0.00005 in the Europe region, $0.000075 in the US and $0.000175 in Singapore, with 90 days of retention. At a million errors a month in Europe, that's about $45.
Who it's for. Teams whose Sentry bill tracks their traffic, and teams already using this Prague-based vendor for logs and uptime.
Who it's not for. Mobile or native apps: Better Stack doesn't accept debug symbol uploads, so native stack traces arrive unsymbolicated.
Its limit. Errors aren't linked to OpenTelemetry traces by default. And whoever gets paged needs a responder license at $34 a month ($29 billed yearly); read-only members are free.
How to migrate. Check your SDK versions first: at least 7.0.0 for JavaScript, 2.0.0 for Python, 4.0.0 for Ruby, 0.1.0 for Go. The DSN is built from an application token, an ingesting host and an application ID. Pick the Europe region when you create the source.
- Free
- 100,000 exceptions a month
- Paid
- $0.00005 per exception (Europe)
- Sentry SDKs
- accepted (recent versions)
- Limit
- no debug symbols
PostHog: errors next to sessions and funnels
If you already do product analyticsWhy move to PostHog. If your product team already uses PostHog, you add error tracking without another vendor and tie a crash to the user's session and journey. The first 100,000 exceptions a month are free, then $0.00037 each up to 325,000, $0.00014 up to 10 million, and $0.000115 beyond. Data can stay in PostHog Cloud EU.
Who it's for. Startups and product teams that want one tool for analytics, replays and errors.
Who it's not for. A team that wants to swap the DSN and be done in an hour: PostHog uses its own SDK.
Its limit. You re-instrument. And the billing cap, handy as it is, stops capture once reached: the same trap as Sentry's quota if you set it too low.
How to migrate. Install the PostHog SDK alongside Sentry, compare for a few weeks, then remove Sentry. Create the project in the EU region from day one if that matters to you.
- Free
- 100,000 exceptions a month
- Paid
- from $0.00037 per exception
- Sentry SDKs
- no, PostHog SDK
- Limit
- re-instrumentation
Honeybadger: errors, uptime and a status page in one plan
Simplest for small teamsWhy move to Honeybadger. The Team plan, at $26 a month ($286 a year), includes 50,000 errors and 90 days of history, plus logs (Insights), uptime monitoring and a status page. On Sentry, uptime is billed per alert. The Developer plan is free for 5,000 errors, one user and 15 days.
Who it's for. A solo developer or web agency watching several client sites who wants everything in one place, without tuning six meters.
Who it's not for. A team leaving over error pricing: the tiers look a lot like Sentry's (50,000 errors on both). The win is what's included, not the price.
Its limit. Its own SDKs, so you re-instrument. An EU data residency option exists: check which plan it comes with before you sign. The Business plan ($80 a month) extends history to 180 days.
How to migrate. Use the 30-day trial to run Honeybadger next to Sentry, then rebuild your uptime checks there.
- Free
- 5,000 errors, 1 user
- Paid
- Team $26/mo (50,000)
- Sentry SDKs
- no, own SDKs
- Limit
- priced close to Sentry
AppSignal: performance monitoring that doesn't count errors
Performance firstWhy move to AppSignal. AppSignal bills per request, not per error, so a crash loop won't blow up your invoice. The free plan covers 30,000 requests, 1 GB of logs and 5 days of history. The first paid plan is $279 a year (about $23 a month) for 250,000 requests, with unlimited users. There's no overage billing either: AppSignal reviews your usage over 30 days before suggesting the next plan up.
Who it's for. Teams on Ruby, Elixir, Node.js, JavaScript or Python who want errors, performance and logs in one tool without fearing spikes.
Who it's not for. A team that only wants error tracking: you're paying for full performance monitoring. Or a team on PHP, Go or native mobile, outside the supported languages (OpenTelemetry aside).
Its limit. Its own agent, and a bill that follows traffic rather than errors: a busy site pays more than a site that crashes a lot.
How to migrate. Install the AppSignal agent, compare the errors it reports for two weeks, then remove Sentry. Pricing displays in euros or dollars.
- Free
- 30,000 requests a month
- Paid
- $279/yr (250,000 requests)
- Sentry SDKs
- no, AppSignal agent
- Limit
- billed on traffic
The 6 alternatives side by side
Our default pick: GlitchTip to keep your SDKs and your data in Europe, Better Stack if volume is the problem. To compare what matters, we worked out one month at 100,000 errors and at 1 million from the official tiers. On Sentry Team, going from 50,000 to 100,000 errors adds roughly $15 to $18 depending on whether you reserve or pay as you go; at 1 million, the error line alone lands around $185 to $230 a month, plan included. Treat these as ballpark figures, excluding logs, traces and replays.
| Sentry SDKs and hosting | Cost at 100,000, then 1 million errors a month | |
|---|---|---|
| Sentry TeamWhere you start | Own SDKs; US or Frankfurt region, DSN and accounts in the US | About $40 to $47, then $185 to $230 |
| GlitchTipOpen source, Europe | Sentry SDKs accepted; German instance or self-hosted | $15 (Small), then $135 (Medium plus overage) or $250 (Large) |
| BugsinkLight self-hosting | Sentry SDKs accepted; EU-hosted or self-hosted | $50 managed, then $158; free on your server |
| Better StackHigh volume | Sentry SDKs accepted (recent versions); Europe region | $0, then about $45 (Europe), plus $29 to $34 per responder license |
| PostHogProduct teams | PostHog SDK; PostHog Cloud EU | $0, then about $178 |
| HoneybadgerSmall teams | Own SDKs; EU residency option | Team $26 for 50,000 errors; beyond that, on its pricing page |
| AppSignalPerformance first | AppSignal agent; Ruby, Elixir, Node.js, JavaScript, Python | Billed per request, from $279 a year for 250,000 requests |
Prices in US dollars excluding tax, from official pricing pages in 2026, subject to change. Costs at 1 million are our own math from the published tiers. Sentry history doesn't transfer to any tool.
Migrating off Sentry in seven steps
Order matters. Changing the DSN takes a minute; the time goes into alerts, integrations and the files that make errors readable. Each step protects something.
List everything Sentry does for you
Day 1Project by project, write down the DSN, alert rules, integrations (Slack, Jira, GitHub), the sourcemap upload in your build pipeline (the files that map minified JavaScript back to readable code), mobile debug symbols, cron monitors and uptime checks.
The inventory also tells you which tool fits: if you lean on replays or profiling, an errors-only tool won't cover you.
- Output
- one row per project
- Risk if rushed
- a forgotten alert
Create the account in the right region
Day 1If you're leaving for GDPR reasons, pick the EU region when you create the account: at PostHog, GlitchTip and Better Stack, that's the moment that counts, and it's rarely changeable later. At Sentry, it never is.
- Decide
- before the first project
- Risk
- data outside Europe
Check your SDK versions
Day 2For GlitchTip, Bugsink and Better Stack, your Sentry SDKs stay in place. Check their versions: Better Stack needs at least 7.0.0 for JavaScript and 2.0.0 for Python. A forgotten legacy project on an old version will send into the void.
For PostHog, Honeybadger and AppSignal, plan to install their SDK project by project.
- Keep the SDK
- GlitchTip, Bugsink, Better Stack
- Re-instrument
- the other three
Move the DSN into an environment variable
Day 2If your DSN is hardcoded, pull it into an environment variable (usually SENTRY_DSN). You can then switch one environment at a time, staging before production, and roll back in a minute.
- Variable
- SENTRY_DSN
- Rollback
- one minute
Send errors to both tools for two to four weeks
Weeks 1 to 4Sentry SDKs can send to two destinations at once with a multiplexed transport, a setup Datadog documents for its own customers. For two to four weeks, compare grouping, noise and alerts.
It's also your only way to lose nothing: events, traces, logs and replays don't leave Sentry.
- Duration
- 2 to 4 weeks
- Doesn't migrate
- history
Rebuild alerts, integrations and sourcemap uploads
Weeks 2 to 3Recreate alert rules and Slack or Jira integrations in the new tool. Replace the sentry-cli step in your pipeline with the new vendor's sourcemap uploader: without it, your JavaScript errors land on one unreadable line.
For mobile apps, check debug symbol uploads. Better Stack doesn't support them.
- Rebuild
- alerts, integrations, sourcemaps
- Trap
- debug symbols
Filter before sending, then turn Sentry off
Week 4Before cutting over, set the SDK's beforeSend filter to drop useless errors (bots, browser extensions) and lower replay sampling. That pays off with every vendor that bills per event.
Then remove the Sentry DSN. Drop to the free Developer plan rather than closing the account: you keep the last 30 days around until you no longer need them.
- Setting
- beforeSend and sampling
- Afterwards
- free Developer plan
Is your tool missing from this page? Email us: hello@hackceleration.com
What makes a Sentry migration go wrong
The same mistakes show up in forum threads again and again. None of them is about the DSN: they come from everything around the SDK.
Assuming your history comes along
Sentry only lets you relocate accounts, projects, teams, dashboards and alerts, and only to a self-hosted instance. Errors, traces, logs and replays stay behind. Dual sending is your only bridge.
Forgetting sourcemaps
The new tool gets errors from minute one, but without sourcemaps they point at minified code. The team decides the tool is worse, when the missing piece is a build step.
Picking a tool that has shut down
Highlight.io still shows up in comparison posts, but its site now redirects to LaunchDarkly. Open a tool's pricing page to confirm it still exists before you migrate.
Ignoring ad blockers
A developer on Hacker News points out that uBlock blocks browser errors from reaching Sentry, even through a subdomain. That holds for every tool: a share of browser-side errors never arrives, whichever one you pick.
Shipping sensitive data
A stack trace can carry an email, a token or health data. One healthcare developer says they're "terrified of PHI being caught in a stack trace". Scrub those fields in beforeSend before sending, whatever the vendor.
Setting the cap too low
PostHog's billing cap stops capture, Rollbar stops at the plan limit by default, and Sentry drops events once quota runs out. A cap protects your bill, not your monitoring: set an alert well before it.
- Projects, DSNs, alerts and integrations inventoried
- EU region chosen at account creation if needed
- SDK versions checked for every project
- DSN moved out of the code into an environment variable
- Dual sending running for two to four weeks
- Sourcemaps and debug symbols uploaded to the new tool
- Sensitive data scrubbed in beforeSend
- Sentry account dropped to the free plan, not closed
Sentry alternatives: frequently asked questions
What is the cheapest Sentry alternative?
At low volume, Better Stack and PostHog: the first 100,000 exceptions each month are free. At high volume, Better Stack in its Europe region, at roughly $45 a month for 1 million errors. And if you have a server, self-hosted GlitchTip or Bugsink cost only that server.
Is Sentry open source?
No. Sentry uses the Functional Source License: the code is readable and you can host it for your own use, but not resell it as a service. Each release becomes Apache 2.0 or MIT after two years. For an open source license today, GlitchTip is MIT.
Is self-hosted Sentry free?
The license is; the server isn't. You need at least 16 GB of RAM and maintenance on every upgrade. One developer's take: "If you just want the team sentry plan it's going to be cheaper to use hosted." At the other extreme, a very high-volume team says it pays about $300 a month for a server where the managed plan would run into the tens of thousands.
Is there a free open source alternative to Sentry?
Yes: GlitchTip, under the MIT license, free to self-host and compatible with Sentry SDKs. It needs Postgres and about 512 MB of RAM, a far cry from Sentry's 16 GB. Bugsink is also free to self-host, but its PolyForm Shield license isn't open source.
Can I keep my Sentry SDKs if I switch?
Yes with GlitchTip, Bugsink, Better Stack and Datadog Error Tracking. No with PostHog, Honeybadger, AppSignal, Rollbar, Bugsnag and Raygun, which ship their own SDKs. Check the minimum version: Better Stack wants at least 7.0.0 for JavaScript.
How much does Sentry charge past the error quota?
On Team pay-as-you-go, each error between 50,000 and 100,000 costs $0.0003625, then $0.0002188 up to 500,000. Going from 50,000 to 150,000 errors a month adds about $29. Reserving volume upfront is cheaper ($0.00029 on the first tier), but unused volume is lost.
How can I cut my Sentry bill without leaving?
Three levers. Pay yearly ($26 instead of $29 on Team). Reserve error volume rather than paying as you go. And filter before sending:
beforeSenddrops noise from bots and extensions, and lower sampling trims replays and traces.What's the best Sentry alternative for mobile apps?
One that accepts your debug symbols, or native crashes arrive unreadable. Better Stack doesn't support them. Before choosing, check the tool's docs for iOS and Android symbol uploads, and keep dual sending on while you compare.
Get the next migration guide in your inbox
One tool to leave, one destination, one migration plan. Nothing else.