self-hosted sentry-self-hosted-relay-1 is unhealthy

Self-Hosted Version

25.9.0 (master branch)

CPU Architecture

x86_64

Docker Version

Docker version 28.4.0, build d8eb465

Docker Compose Version

Docker Compose version v2.39.4

Machine Specification

[x] My system meets the minimum system requirements of Sentry

Steps to Reproduce

docker compose up -d --wait after installing.

Also note, this is an airgapped VM.

Expected Result

Relay being healthy or atleast, verbose healthcheck that lets me know where its failing.

Actual Result

The healthcheck that is failing

docker compose exec -it relay /bin/relay --config /work/.relay healthcheck
error: Relay is unhealthy. Status code: 503 Service Unavailable

relay logs:

relay-1  | 2025-10-04T11:27:51.075770Z  INFO relay::setup: launching relay from config folder /work/.relay
relay-1  | 2025-10-04T11:27:51.122966Z  INFO relay::setup:   relay mode: proxy
relay-1  | 2025-10-04T11:27:51.122993Z  INFO relay::setup:   relay id: 94d45cba-46f9-42b8-b309-30dcbe444bfa
relay-1  | 2025-10-04T11:27:51.123003Z  INFO relay::setup:   public key: [relay_public_key_here]
relay-1  | 2025-10-04T11:27:51.123014Z  INFO relay::setup:   log level: trace
relay-1  | 2025-10-04T11:27:51.340404Z  INFO relay_server: relay server starting
relay-1  | 2025-10-04T11:27:51.666802Z  INFO relay_server::services::outcome: Configured to emit outcomes as client reports
relay-1  | 2025-10-04T11:27:51.688968Z  INFO relay_server::services::outcome: OutcomeProducer started.
relay-1  | 2025-10-04T11:27:51.688857Z  INFO relay_server::services::outcome_aggregator: outcome aggregator started
relay-1  | 2025-10-04T11:27:51.718974Z  INFO relay_server::services::global_config: global config service starting
relay-1  | 2025-10-04T11:27:51.728283Z  INFO relay_server::services::projects::source::upstream: project upstream cache started
relay-1  | 2025-10-04T11:27:51.750826Z  INFO relay_server::services::global_config: serving default global configs due to lacking static global config file
relay-1  | 2025-10-04T11:27:51.797168Z  INFO relay_server::service: starting 1 store workers
relay-1  | 2025-10-04T11:27:51.797656Z  INFO relay_server::services::metrics::router: metrics router started
relay-1  | 2025-10-04T11:27:51.798552Z  INFO relay_server::service: starting 6 envelope processing workers
relay-1  | 2025-10-04T11:27:51.825918Z TRACE relay_server::services::buffer::envelope_buffer: PolymorphicEnvelopeBuffer: initializing memory envelope buffer
relay-1  | 2025-10-04T11:27:51.837546Z  INFO relay_server::services::relays: key cache started
relay-1  | 2025-10-04T11:27:51.837654Z  INFO relay_server::services::server: spawning http server
relay-1  | 2025-10-04T11:27:51.837684Z  INFO relay_server::services::server:   listening on http://0.0.0.0:3000/
relay-1  | 2025-10-04T11:27:51.876868Z  INFO relay_server::services::buffer: EnvelopeBufferService 0: starting

Web logs:

web-1  | Updating certificates in /etc/ssl/certs...
web-1  | 0 added, 0 removed; done.
web-1  | Running hooks in /etc/ca-certificates/update.d...
web-1  | done.
web-1  | sentry/requirements.txt is deprecated, use sentry/enhance-image.sh - see https://develop.sentry.dev/self-hosted/#enhance-sentry-image
web-1  | error : Unknown IO error
web-1  | *** Starting uWSGI 2.0.28 (64bit) on [Sat Oct  4 11:26:42 2025] ***
web-1  | compiled with version: 10.2.1 20210130 (Red Hat 10.2.1-11) on 31 October 2024 18:21:41
web-1  | os: Linux-6.8.0-85-generic #85-Ubuntu SMP PREEMPT_DYNAMIC Thu Sep 18 15:26:59 UTC 2025
web-1  | nodename: acc7ebf9cfb8
web-1  | machine: x86_64
web-1  | clock source: unix
web-1  | pcre jit disabled
web-1  | detected number of CPU cores: 8
web-1  | current working directory: /usr/src/sentry
web-1  | detected binary path: /.venv/bin/python3
web-1  | your memory page size is 4096 bytes
web-1  | detected max file descriptor number: 4096
web-1  | lock engine: pthread robust mutexes
web-1  | thunder lock: enabled
web-1  | uWSGI http bound on 0.0.0.0:9000 fd 4
web-1  | uwsgi socket 0 bound to TCP address 127.0.0.1:41635 (port auto-assigned) fd 3
web-1  | Python version: 3.13.1 (main, Feb  4 2025, 04:56:26) [GCC 12.2.0]
web-1  | --- Python VM already initialized ---
web-1  | Python main interpreter initialized at 0x78288427ca70
web-1  | python threads support enabled
web-1  | your server socket listen backlog is limited to 100 connections
web-1  | your mercy for graceful operations on workers is 60 seconds
web-1  | setting request body buffering size to 65536 bytes
web-1  | mapped 1924224 bytes (1879 KB) for 12 cores
web-1  | *** Operational MODE: preforking+threaded ***
web-1  | spawned uWSGI master process (pid: 739)
web-1  | spawned uWSGI worker 1 (pid: 741, cores: 4)
web-1  | spawned uWSGI worker 2 (pid: 742, cores: 4)
web-1  | spawned uWSGI worker 3 (pid: 743, cores: 4)
web-1  | spawned uWSGI http 1 (pid: 744)
web-1  | error : Unknown IO error
web-1  | error : Unknown IO error
web-1  | error : Unknown IO error
web-1  | WSGI app 0 (mountpoint='') ready in 51 seconds on interpreter 0x78288427ca70 pid: 743 (default app)
web-1  | WSGI app 0 (mountpoint='') ready in 51 seconds on interpreter 0x78288427ca70 pid: 741 (default app)
web-1  | WSGI app 0 (mountpoint='') ready in 51 seconds on interpreter 0x78288427ca70 pid: 742 (default app)

The config file (I updated it to change mode to proxy and disabled the processing)

relay:
  upstream: "http://web:9000/"
  host: 0.0.0.0
  port: 3000
  mode: proxy
logging:
  level: TRACE
processing:
  enabled: false
  kafka_config:
    - {name: "bootstrap.servers", value: "kafka:9092"}
    - {name: "message.max.bytes", value: 50000000} # 50MB
  redis: redis://redis:6379
  geoip_path: "/geoip/GeoLite2-City.mmdb"

Is there a way I can make the healthcheck verbose to see what exactly is going wrong? The relay logs dont seem to be showing anything even at trace. I dont't think(?) this is an resource constraint issue as the logs don't show that either.

Event ID

No response

Oct 05 '25 03:10 kilowatt0922-lang

Also note: the [relay_public_key_here] in the relay logs doesn't mean thats how its stored in my credentials.json, I've just manually redacted it out of my logs.

Oct 05 '25 03:10 kilowatt0922-lang

Both your web and relay container logs looks fine to me. Do you have any logs of relay other than that?

Is there a way I can make the healthcheck verbose to see what exactly is going wrong? The relay logs dont seem to be showing anything even at trace. I dont't think(?) this is an resource constraint issue as the logs don't show that either.

The healthcheck is just a simple curl http://localhost:3000/api/relay/healthcheck/ready/. You can execute it manually by executing it from the web container:

docker compose exec web curl -v http://relay:3000/api/relay/healthcheck/ready/

Oct 14 '25 03:10 aldy505

@kilowatt0922-lang how did you solve your problem? I have the same problem here with 25.10.0. relay don't come up and web is not reachable.

Oct 21 '25 08:10 gizmo84

In my case what help: create sentry/config.yaml, because previously I don't have it.

Why Sentry don't copy sentry/config.example.yaml to sentry/config.yaml while installing?

Docs

Nov 16 '25 12:11 LocalHeroPro

In my case what help: create sentry/config.yaml, because previously I don't have it.

Why Sentry don't copy sentry/config.example.yaml to sentry/config.yaml while installing?

Docs

It should automatically: https://github.com/getsentry/self-hosted/blob/fe477b41d937b838146f83a48bbbe705ec959f49/install/ensure-files-from-examples.sh#L3-L4

https://github.com/getsentry/self-hosted/blob/fe477b41d937b838146f83a48bbbe705ec959f49/install/_lib.sh#L28-L46

Nov 16 '25 12:11 aldy505

This issue has gone three weeks without activity. In another week, I will close it.

But! If you comment or otherwise update it, I will reset the clock, and if you remove the label Waiting for: Community, I will leave it alone ... forever!

"A weed is but an unloved flower." ― Ella Wheeler Wilcox 🥀

Dec 08 '25 08:12 getsantry[bot]