Sentry alternatives when the job is exceptions

You are leaving Sentry Cloud, or you are refusing a 16 GB self-host, and you still want grouped exceptions from the SDK you already ship. This page is the leave path. Install shortlist detail: Best self-hosted. Free quotas: Best free. Bills: Cheapest.

What "alternative" means here

An alternative keeps exception ingest with an official Sentry SDK after a DSN change, or asks you to adopt a different SDK on purpose. It is not a claim of full Sentry protocol parity. Epure stores exception events and drops transaction, replay, and profile items. Say that out loud before you migrate.

When to stay on Sentry

Stay when session replay closes tickets, when traces are the daily debugger, when mobile symbolication is required, or when the team already lives in Sentry's query language. Stay on Sentry self-host if you want that suite on your metal and can fund about 16 GB RAM. Disclose COMPOSE_PROFILES=errors-only when you argue a lighter Sentry footprint; it is still a large fleet next to two containers. Sentry self-hosted RAM.

Small self-host alternatives

GlitchTip (MIT, about four services), Bugsink (Polyform Shield, one container), Epure (Apache 2.0, two containers, PostgreSQL 16), Telebugs ($299 once). All take a DSN swap for errors with the caveats each vendor documents. Pairwise: vs Sentry, vs GlitchTip, vs Bugsink, vs Telebugs. Research matrix: Self-hosted research.

Hosted alternatives with different SDKs

PostHog Errors, Honeybadger, Bugsnag, Rollbar, and others sell hosted error products with their own SDKs or adapters. They can be the right buy when you want a vendor-operated cluster and accept a new client library. They are the wrong comparison when the constraint is "keep @sentry/* and change the DSN."

Lived leave

We dual-sent for seven days. Production still had Sentry with sample rates at 0 for traces and replay. Epure received exceptions on a second DSN. Issues matched on the bugs we cared about. We cut Sentry when the on-call doc pointed at Epure and the pager test passed. Gotcha: one worker retained the old DSN in a secrets manager copy. Gotcha: a release string mismatch left maps behind on the new host.

How to leave with a DSN swap

• Stand up the new host (docker compose up -d for Epure; health {"status":"ok"}).
• Create a project; copy the DSN (numeric project id path).
• Set traces/replay/profile rates to 0.
• Dual-send or cut over; verify HTTP 202 and Issues.
• Remove the old DSN from every process.

Narrative: Migrate from Sentry. Failures: Invalid DSN, Events not showing.

Limits and bias

First-party page. Epure is the Apache two-container row. We do not claim to replace Sentry's suite. Cloud list prices on this site are not a live checkout as of 2026-10-05. Self-host is what ships. If your RFP requires session replay as a product feature, keep Sentry or another suite.

Scoreboard (leave path only)

Story: the secrets manager copy

The migrate looked clean in Git. One ECS task definition still pulled SENTRY_DSN from an old parameter. Exceptions from that worker never hit Epure. We found it when Issues counts disagreed with access logs. Fix: one DSN name in secrets, grep the org for the old host, rotate the old key. Invalid DSN covers revoked keys when you rotate on purpose.

What not to do

Do not publish a second "best self-hosted" matrix under this slug. Do not promise 100% Sentry compatibility. Do not put Cloud list prices CTAs on the leave path. Do not use "exception-only" in the title or primary query; keep job language in the head and honesty in the body.

Activation links (leave path)

• Install: Installation
• Verify: Verify
• Coolify panel path: Coolify
• Cheap VPS path: Cheap VPS install

Run one path end to end before you delete the old DSN. A half-migrated fleet is how "alternatives" posts turn into incident posts.

Takeaway

Pick the alternative that matches the job you actually open during an outage. If that screen is exception, count, and stack, a small self-host with a DSN swap is enough for many teams. If that screen is a replay or a trace waterfall, keep the suite and budget for it.

Pricing honesty (fetched 2026-10-05)

Sentry Team about $26/mo annual default with included error volume then PAYG (sentry.io/pricing). Bugsink hosted from about $16 for 75k on their site. Epure self-host $0 Apache 2.0. Epure Cloud list Pro $29 / Scale $89 are not live checkout. GlitchTip Cloud publishes free sample then paid tiers on glitchtip.com. Re-fetch on decision day. Numbers rot.

Takeaway

Alternatives are a job change, not a logo change. Match the host to the screen you open when production throws. Then migrate with dual-send and a secrets grep.

Write the decision down in the on-call doc before you delete the old DSN.