Login rate limit with several replicas #105

Closed
opened 2026-11-18 05:14:41 +00:00 by jcoffey-dev · 2 comments
Owner

K-W3 · size S · part of inbuxa for Kubernetes (tracking issue in inbuxa-charts)

The per-process limiter multiplies with replicas. Keep it as a first line and rely on the server's own auth limits across replicas; document that.

Done when

  • Behaviour with N replicas documented and tested
**K-W3** · size S · part of inbuxa for Kubernetes (tracking issue in inbuxa-charts) The per-process limiter multiplies with replicas. Keep it as a first line and rely on the server's own auth limits across replicas; document that. **Done when** - Behaviour with N replicas documented and tested
jcoffey-dev added this to the 27.1 milestone 2026-11-18 05:14:41 +00:00
jcoffey-dev added the securityenhancement labels 2026-11-18 05:14:41 +00:00
jcoffey-dev self-assigned this 2026-11-18 06:28:58 +00:00
jcoffey-dev added the
priority
p2
label 2026-11-18 06:28:59 +00:00
Author
Owner

Folded into the checklist on inbuxa/inbuxa-charts#14 to keep the issue list short. The details stay here; reopen this when work on it starts.

Folded into the checklist on inbuxa/inbuxa-charts#14 to keep the issue list short. The details stay here; reopen this when work on it starts.
Author
Owner

Done in #115, inbuxa/inbuxa-server#467, inbuxa/inbuxa-charts#22 and inbuxa/inbuxa-k8s-lab#23.

  • The webmail's limits count per replica and are a first line. The README's "Several replicas" section says so.
  • The server's failed-sign-in counts are shared across its pods (tested: banned at 101 with failures split between two pods). The webmail passes the browser's address on, and the chart has the server believe it, so those limits count people, not the proxy.
  • Fixed on the way: the webmail never trusted an IPv4 proxy, because the image listens on :: and Node reports IPv4 peers as ::ffff:a.b.c.d. That bug is in current releases; the fix ships in 27.1-rc.7.

lab/chart/ratelimit-test.sh checks all of it.

Done in #115, inbuxa/inbuxa-server#467, inbuxa/inbuxa-charts#22 and inbuxa/inbuxa-k8s-lab#23. - The webmail's limits count per replica and are a first line. The README's "Several replicas" section says so. - The server's failed-sign-in counts are shared across its pods (tested: banned at 101 with failures split between two pods). The webmail passes the browser's address on, and the chart has the server believe it, so those limits count people, not the proxy. - Fixed on the way: the webmail never trusted an IPv4 proxy, because the image listens on `::` and Node reports IPv4 peers as `::ffff:a.b.c.d`. That bug is in current releases; the fix ships in 27.1-rc.7. `lab/chart/ratelimit-test.sh` checks all of it.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: inbuxa/inbuxa-webmail#105