GlitchTip vs Sentry
Choose Sentry if you want the full product — session replay, profiling and crons, plus its Seer AI debugger on the hosted plan only — either hosted, with a bill that grows with event volume, or self-hosted on the 16 GB of RAM plus 16 GB of swap its docs require. Choose GlitchTip if what you use is errors, stack traces, releases and alerts: it takes the same Sentry SDKs through a DSN change, adds basic performance, uptime and log monitoring, and recommends 512 MB of RAM under an MIT license.
Side by side
Most "GlitchTip vs Sentry" searches are not really about GlitchTip against Sentry's hosted service. They are about a narrower question: Sentry's own server can be self-hosted, so why would anyone run a different one? The answer is the size of the box. Self-hosted Sentry is nearly the full Sentry platform on one machine, and its documented floor is 16 GB of RAM plus 16 GB of swap. GlitchTip accepts events from the same open-source Sentry SDKs and recommends 512 MB. Everything else in this comparison follows from that gap.
Three ways to get Sentry-style error tracking
Sentry (hosted) is the SaaS at sentry.io. You install an SDK, point it at a DSN, and exceptions, stack traces, releases, and alerts show up in Sentry's cloud. The product has grown well past error tracking: its pricing page lists error monitoring, tracing, session replay, profiling, logs, metrics, cron monitoring, and uptime monitoring, plus an AI debugging tool called Seer. There is a free single-user Developer plan. Paid plans include a monthly quota of errors, spans, and replays, and you pay more as you send more.
Sentry (self-hosted) is the same codebase packaged by the
getsentry/self-hosted repository. You clone it and run its bash
install.sh, which sets up a Docker Compose stack of Sentry's services,
including Kafka, ClickHouse, Redis, PostgreSQL, and Snuba. Sentry's own docs
ask for 4 CPU cores, 16 GB of RAM plus 16 GB of swap, and 20 GB of free
disk, and recommend 32 GB of RAM. They are also candid about who it is
for. They call it "a minimal setup that works out-of-the-box for simple use
cases" with "no guarantees or dedicated support", and a blueprint for
"folks willing to maintain larger installations."
GlitchTip is a separate open-source project that speaks the Sentry
event protocol. Its docs tell you to "install the sentry SDK for your
platform" and paste GlitchTip's DSN, so application code doesn't change.
The server is a Django app that needs PostgreSQL 14 or newer and one
service. That service can run web and worker roles together (all_in_one),
or you can split them to scale. Valkey or Redis is optional but
recommended. GlitchTip recommends 512 MB of RAM and lists 256 MB as the
minimum for the all-in-one setup.
Licensing: open source vs. Fair Source
GlitchTip is MIT licensed.
Sentry's server and the self-hosted repository are both under the
Functional Source License 1.1 with an Apache 2.0 future license
(FSL-1.1-Apache-2.0). Sentry calls this "Fair Source". Running it for
your own company's internal use is permitted. What the license forbids is a
"Competing Use": offering the software to others as a commercial product or
service that substitutes for Sentry. Each release becomes Apache 2.0 two
years after it is published. For a team self-hosting error tracking for its
own apps, that restriction is unlikely to matter. It does mean self-hosted
Sentry is not open source in the OSI sense today, and GlitchTip is.
Footprint: 512 MB against 16 GB plus swap
This is the reason the comparison exists. Self-hosted Sentry runs the
pipeline that powers sentry.io — Kafka for ingestion, ClickHouse and Snuba
for queries, and the processing services around them — on one host. Sentry
offers an errors-only Compose profile (from version 24.8.0), set with
COMPOSE_PROFILES=errors-only in .env. It keeps issues, alerts,
integrations, dashboards, and releases, and drops traces, profiles,
replays, crons, user feedback, and application metrics. Its docs describe
it as a lightweight option "with an emphasis on minimizing system
resources" but do not publish a separate hardware floor for it. Plan for
the documented 16 GB.
GlitchTip's official Compose file has three services: PostgreSQL, Valkey,
and one GlitchTip container in all_in_one mode. The file's own comments
say that to target 256–512 MB you can drop Valkey (PostgreSQL then handles
the task queue and cache) and switch off log ingestion and uptime
monitoring. That fits on the smallest VPS tiers. Self-hosted Sentry needs a
much larger server.
Features: where Sentry is ahead
GlitchTip's own feature list covers error tracking, performance
monitoring, uptime monitoring, and logs. Its SDK docs support traces for
performance data, source maps and debug symbols through the CLI, and log
ingestion over OpenTelemetry. They also state that "GlitchTip does not
support session tracking", and tell you to set auto_session_tracking to
false. Session replay, profiling and Sentry-style cron check-ins are not on
GlitchTip's feature list (its uptime monitoring does accept scheduled
heartbeat pings); if your team relies on them, Sentry provides them hosted
or self-hosted. Seer, Sentry's AI debugger, is hosted-only — Sentry's docs
list it as unavailable on self-hosted.
The Sentry SDKs are the part both sides share, so trying GlitchTip costs little: change the DSN, and your existing instrumentation reports to the new server. The reverse works too: moving back to Sentry is also a DSN change for your applications.
Ops burden
Hosted Sentry has nothing to operate. Self-hosted Sentry has the most to operate of the three: Sentry says upgrades go through the same install script, and you keep Kafka, ClickHouse, and PostgreSQL healthy on one large host yourself. GlitchTip is in between. It is a PostgreSQL-backed Django app you back up like any other PostgreSQL database, with a standard Compose file and a Helm chart for Kubernetes. GlitchTip also runs its own hosted service if you want its feature set without running a server.
Which to choose
Choose hosted Sentry if you want the full product (replay, profiling, Seer, crons) and don't want to run anything. Accept that the bill grows with event volume.
Choose self-hosted Sentry if you need Sentry's feature set (no Seer) on your own infrastructure, for data-residency or policy reasons. You also need a 16–32 GB server and someone willing to maintain a multi-service stack that Sentry itself says comes with no guarantees.
Choose GlitchTip if what you actually use is errors, stack traces, releases, and alerts, maybe with basic performance and uptime checks. It keeps your existing Sentry SDKs and runs on a small VPS, under an MIT license.
Running GlitchTip on a VPS
The GlitchTip page has the official Compose
install, which we have verified on Ubuntu 26.04. Set GLITCHTIP_DOMAIN and
EMAIL_URL before real use, and put a reverse proxy in front for TLS. If
GlitchTip still feels like more than you need, Bugsink vs.
GlitchTip compares it with an even smaller
single-container tracker. For every self-hosted option we list against
Sentry, see Sentry alternatives.
Common questions
GlitchTip vs Sentry — which should I use?
Use Sentry if you need session replay, profiling or cron check-ins (hosted or self-hosted), or its Seer AI debugger (hosted only). Use GlitchTip if you mainly need errors, stack traces, releases and alerts on a small server: it accepts the same Sentry SDKs and recommends 512 MB of RAM.
Can I self-host Sentry instead of using GlitchTip?
Yes. Sentry publishes getsentry/self-hosted, a Docker Compose stack installed with its install.sh script. Its docs require 4 CPU cores, 16 GB of RAM plus 16 GB of swap and 20 GB of free disk, recommend 32 GB of RAM, and describe it as coming with no guarantees or dedicated support.
Do my existing Sentry SDKs work with GlitchTip?
Yes. GlitchTip's docs tell you to install the Sentry SDK for your platform and point it at your GlitchTip project's DSN. Set auto_session_tracking to false, because GlitchTip does not support session tracking.
Is self-hosted Sentry open source?
Not in the OSI sense. Sentry is under the Functional Source License 1.1 with an Apache 2.0 future license: internal use is allowed, offering it as a competing commercial service is not, and each release becomes Apache 2.0 two years after publication. GlitchTip is MIT licensed.
Paid link — we earn a commission if you shop through it.
Other comparisons with these apps
Errors only in one container, or errors plus uptime and performance on PostgreSQL.
The multi-database web client from DBeaver vs. the PostgreSQL-only admin tool.
One conf.yml with an in-browser editor vs. YAML files and Docker label discovery.