Sentry is where many teams start error tracking. They look elsewhere for three reasons: an error spike that runs through the monthly quota and either bills or drops events, self-hosted Sentry asking for 16 GB of RAM and its own Kafka, or wanting errors, traces and logs from OpenTelemetry in one place. These six tools answer one or more of those, and three of them keep the Sentry SDKs already in your code.
Maple, the first entry below: a trace on its waterfall view.
We build Maple and list it first because this is our site: that placement is our bias, not a ranking. Next come the three options that keep the Sentry SDKs, then two that need a different SDK. Every price comes from the vendor's own pricing page, linked at the bottom with the month we checked it.
How we compared: we re-read each vendor's pricing page and the documentation cited below in November 2026. Sentry, GlitchTip and PostHog have no comparison page on this site, so their facts link to their own pages below; Better Stack and SigNoz reuse sourced rows from our comparison pages, each with the month it was last checked. We did not benchmark the tools against each other; treat this as a map of trade-offs, not a test result.
Updated
The short answer
Keep the Sentry SDKs, lower the bill
Better Stack or GlitchTip. Both accept the Sentry SDKs. Better Stack includes 100,000 exceptions a month; GlitchTip's hosted plans start at $15 for 100,000 events.
Self-host on a small server
GlitchTip. PostgreSQL and one service, with 512 MB of RAM recommended.
Errors, traces, logs and replay from OpenTelemetry in one app
Maple. Each error issue opens the trace that threw it, and exceptions are billed as part of trace volume, with no separate error quota.
Keep Sentry itself, without the per-event bill
Self-hosted Sentry. The same product and SDKs, if you can run a 16 GB server and support it yourself.
Product team already on PostHog
PostHog. Every issue comes with the user's session replay and events, and 100,000 exceptions a month are free.
At a glance
Tool
Pick it if
Pricing
License
Sentry SDKs
Sentry (for reference)
The tool you are replacing
Free for one user; Team $29/mo, Business $89/mo, each with 50k errors and 5M spans, then per event
You are moving to OpenTelemetry and want exceptions next to traces, hosted or self-hosted.
$49/mo including $49 of usage, then $0.30 per GB of logs and traces
MIT core, SSO and RBAC under a commercial license
No; OpenTelemetry SDKs
Maple
Built by us
Pick it if: You want errors, traces, logs and session replay from OpenTelemetry in one app, priced per GB.
Maple builds error tracking from OpenTelemetry. Exceptions recorded on spans are grouped into issues by type, message and stack, and every issue opens the trace that threw it, with the query and status code at the time. Issues carry an owner, a severity and notes, and a resolved issue that comes back reopens. There is no per-error meter: exceptions are part of trace volume. Session replay, uptime checks, cron heartbeats and on-call are in the same plan.
Strengths
+Every issue linked to the span and trace that threw it
+No per-error meter or error quota
+Session replay with the backend trace behind each click
+Uptime checks, cron heartbeats, on-call schedules and SLOs in one plan
Trade-offs
-No intake for the Sentry SDKs; you instrument with OpenTelemetry
-Mobile is a Swift SDK for iOS
-No free plan; it starts at $39 a month
-Younger, with a smaller ecosystem; SOC 2 and ISO 27001 certification still in progress
→ propose_fix{ issue: "iss_8a91", fix: "scope idempotency_key by user_id" }
← fix attached. opened pr#412.
>
4 tool calls · 4.2sopen mcp protocol
An agent session over Maple's MCP server.
$39/mo with 100 GB each of logs, traces and metrics, then $0.30 per GB; 5,000 browser sessions included · Source-available (FSL-1.1, Apache 2.0 after two years)
Pick it if: You want Sentry-compatible error tracking on a small server, or a low-cost hosted plan.
GlitchTip is open-source error tracking that works with Sentry's client SDKs, so the SDKs already in your code can send to it. It also covers performance, uptime checks and logs. Self-hosted, it needs PostgreSQL and a single service, with 512 MB of RAM recommended and Valkey optional. Its hosted plans count each error and each performance transaction as one event, and ten uptime checks or log lines as one.
Strengths
+Works with Sentry's client SDKs
+Self-hosts on PostgreSQL with 512 MB of RAM recommended
+Hosted plans from $15 a month for 100,000 events
Trade-offs
-Session replay is not on its feature list
-Disk grows with volume: about 30 GB for a million events a month, per its docs
Free to 1,000 events/mo; $15/mo for 100k, $50 for 500k, $250 for 3M · Open source (MIT)
Pick it if: You want to keep Sentry as it is, stop paying per event, and can run one large server.
Sentry's own self-hosted build is, in Sentry's words, the Business plan without software limits and no paid tier. It leaves out billing features, spike protection and closed-source AI features such as Seer. Every service runs on one node, Kafka included. The minimum is 4 CPU cores, 16 GB of RAM plus 16 GB of swap and 20 GB of disk, with 32 GB of RAM recommended. Sentry offers no support or scaling guidance for it.
Strengths
+The same product and SDKs, with no per-event bill
+Business plan features without software limits
Trade-offs
-4 cores and 16 GB of RAM plus swap at minimum; 32 GB recommended
-Single node for every service, with no support or scaling guidance
-No Seer and no spike protection
Free software; you pay for the server · FSL (Apache 2.0 after two years)
Pick it if: You want to keep the Sentry SDKs, pay less per exception, and get uptime and on-call alongside.
Better Stack's error tracking accepts the Sentry SDKs: you keep the code and change where it sends. It includes 100,000 exceptions a month with 90 days of retention, then charges $0.05 per 1,000 in Europe and $0.075 in the US. Errors can be resolved, snoozed and tracked by release, and they sit next to logs, traces, uptime checks and on-call with phone and SMS.
Strengths
+Accepts the Sentry SDKs
+100,000 exceptions a month included, with 90 days of retention
+Uptime, on-call, logs and traces in the same product
Trade-offs
-Closed source, with no self-hosting
-Anyone on call needs a $34/mo Responder license
100,000 exceptions/mo included, then $0.05 per 1,000 (EU) or $0.075 (US) · Closed source; SaaS, custom VPC on Enterprise
Pick it if: You already use PostHog for product analytics, or want each error with the user's session replay.
PostHog added error tracking to its product analytics platform. Every issue comes with the affected user's session replay, events and properties. It takes source maps and symbol sets so minified stack traces point back to your code, and it tags exceptions with the release and commit that introduced them. The first 100,000 exceptions a month are free.
Strengths
+Session replay and product events on every issue
+100,000 exceptions a month free
+Source maps, and release and commit tagging
Trade-offs
-Its own SDKs, not Sentry's
-Self-hosting is a hobby deploy that scales to about 100,000 events a month, without support
First 100,000 exceptions/mo free, then $0.37 per 1,000 · MIT, except the ee directory
Pick it if: You are moving to OpenTelemetry and want exceptions next to traces, hosted or self-hosted.
SigNoz is an OpenTelemetry APM with logs, metrics and alerts. It lists exceptions grouped by service, type and message, each linked to its trace. A group has no state, owner or severity, so triage happens elsewhere. It suits teams whose main goal is tracing, with exceptions as one view among several.
Strengths
+Every exception linked to its trace
+Self-hosted under an MIT core
+No user-based pricing on Cloud
Trade-offs
-No state, owner or severity on an exception group
-15 days for logs and traces by default
$49/mo including $49 of usage, then $0.30 per GB of logs and traces · MIT core, SSO and RBAC under a commercial license
8 services · 8 edges · 60s windowpricing-svc · p99 +18% vs 1h ago
Before you switch
When to stay on Sentry
You use what Sentry adds on top of grouping: release health, Seer for AI debugging, and anomaly detection and Code Owners on Business.
Your team is growing but your error volume is modest. Team includes unlimited users and 50,000 errors for $29 a month.
You work alone. The Developer plan is free, with 5,000 errors, 5 million spans and a 30-day lookback.
You prefer a hard cap. Past your reserved volume and pay-as-you-go budget, Sentry drops events instead of billing them.
FAQ
よくある質問
What is the best Sentry alternative?
It depends on what you want to keep. To keep your Sentry SDKs and pay less, look at Better Stack or GlitchTip. To self-host on a small server, GlitchTip; to self-host Sentry itself, plan for a 16 GB machine. To get errors, traces, logs and replay from OpenTelemetry in one app, Maple. If your product team already lives in PostHog, its error tracking. We build Maple, so weigh our view accordingly.
Is there an open-source alternative to Sentry?
Yes. GlitchTip is MIT and accepts the Sentry SDKs. PostHog is MIT except its ee directory. SigNoz has an MIT core with SSO and RBAC under a commercial license. Maple is source-available under FSL-1.1. Sentry itself is under the FSL too, becomes Apache 2.0 after two years, and can be self-hosted for free.
Can I keep using the Sentry SDKs?
With GlitchTip, Better Stack and self-hosted Sentry, yes: point the SDK at the new endpoint. Maple and SigNoz take OpenTelemetry, so you swap the Sentry SDK for the OpenTelemetry SDK for your language. PostHog uses its own SDKs.
What does self-hosted Sentry need?
At least 4 CPU cores, 16 GB of RAM plus 16 GB of swap, and 20 GB of free disk; Sentry recommends 32 GB of RAM. Every service, Kafka included, runs on one node, and Sentry gives no scaling guidance or support. It excludes spike protection and Seer. GlitchTip, for comparison, recommends 512 MB of RAM.
Why does my Sentry bill spike?
Errors are billed per event above the 50,000 included. Pay-as-you-go starts at $0.3625 per 1,000 errors on Team and $1.1125 on Business, so a release that throws in a loop adds up fast. Once the reserved volume and budget are used, Sentry drops new events for the rest of the billing cycle. Our Sentry pricing guide works through two monthly bills.
出典
Where the vendor facts come from
Prices link to each vendor's own pricing page, with the month we last checked it. Every other claim
repeats a row on the matching Maple vs page, where it carries its own source, or comes from a vendor
page listed here. If something is out of date, tell us and we will fix it.