What is the smallest urgentry deployment?

Tiny mode starts as one binary on one machine with SQLite. It is the fastest way to run the switch proof.

When should I skip Tiny mode and start with self-hosted mode?

Start with self-hosted mode when your team already needs shared services, split roles, or sustained ingest above the Tiny lane.

Does this page claim urgentry beats Sentry everywhere?

No. This page compares the self-hosted shape first: installation, operator burden, memory footprint, and the first proof path.

Sentry self-hosted alternative

If Sentry self-hosted feels too heavy, compare the operator shape first.

Use this page if the question is not product breadth but footprint: install shape, RAM floor, service count, and the first proof path on one machine.

Migration Guide Benchmarks

1 binary Tiny mode starts as one process on one machine.

52.3 MB Peak memory on the current Tiny reference lane.

391.8 MB Peak memory on the current urgentry self-hosted lane.

8191.7 MB Peak memory on the current Sentry self-hosted lane.

Operator shape

Compare install, footprint, and runtime shape.

Keep the comparison on the first operator questions. Do not start with a parity debate.

Question	urgentry Tiny	urgentry self-hosted	Sentry self-hosted
First install shape	One binary and SQLite on one machine.	`api`, `ingest`, `worker`, and `scheduler` with PostgreSQL, MinIO, Valkey, and NATS.	A larger multi-service stack.
Same-host stable throughput	400 eps	2,200 eps	1,000 eps
Same-host query p95	78.66 ms	48.82 ms	1400.81 ms
Same-host peak memory	52.3 MB	391.8 MB	8191.7 MB

Trust receipt

Who maintains this comparison and where to verify it

This page compares operator shape, footprint, and first-proof tradeoffs. Use these receipts to verify the claim and the current public evidence.

Page facts

Last updated: April 16, 2026
Maintained by: Wraxle LLC and the public repo maintainer lane
Support: /docs/support/
Security: /security/

Claim scope

This page is for teams that already use Sentry SDKs and want to compare install shape, RAM floor, service count, and operator work before they commit to a self-hosted migration.

Migration Guide Benchmarks Self-Hosted Mode

Fit check

Use this page if install shape is the main problem.

Test urgentry first

You already use Sentry SDKs and do not want to re-instrument.
You want to compare RAM floor, service count, and operator work.
You want a one-machine proof before you widen the rollout.

Do more checking first

You need full Sentry parity on day one.
You are buying for enterprise governance first.
You are not willing to operate anything yourself.

Next docs

Use the migration guide for the DSN change path. Then use the self-hosted docs and benchmarks if the operator shape is still the main question.

Migration Guide Self-Hosted Mode Benchmarks