[Bug]: Clicking a face in the face library results in error '414 (Request-URI Too Large)'

[tombeynon]

Checklist

• I have updated to the latest available Frigate version.
• I have cleared the cache of my browser.
• I have tried a different browser to see if it is related to my browser.
• I have tried reproducing the issue in incognito mode to rule out problems with any third party extensions or plugins I have installed.
• I have asked the AI at https://docs.frigate.video about my issue.

Describe the problem you are having

Clicking a face in the face library does nothing, and shows the error 414 (Request-URI Too Large) in the browser console. This does not seem to occur with the Home Assistant add-on, but has started to happen (with new and old faces) when I switched to the Docker (compose) installation. I have tried both the authenticated and non-authenticated ports, and I'm connecting to the Frigate IP directly so there is no other nginx instance between the request and Frigate.

The failing request is http://{ip}:{port}/api/event_ids?ids={very long string} and I can see the 414 failure in the Frigate nginx logs.

Is this a bug or something wrong with my recently migrated setup?

Steps to reproduce

I have a feeling this is somewhat specific to my setup as I haven't seen any other mention of it in Issues etc, but steps below.

1. Installed with the included docker-compose which is pretty close to the compose file in the documentation.
2. Navigate to Face Library, click on any face in either the Train or specific face tabs.
3. Nothing happens, see error in browser console

Version

0.16.2-4d58206

In which browser(s) are you experiencing the issue with?

Google Chrome 142.0.7444.60

Frigate config file

mqtt:
  enabled: true
  host: 10.2.3.4
  user: mqtt
  password: XXX
  
tls:
  enabled: False

genai:
  enabled: false
  provider: openai
  api_key: XXX
  model: gpt-4o

detectors:
  ov_1:
    type: openvino
    device: GPU

face_recognition:
  enabled: true
  model_size: large
  min_area: 500

model:
  path: plus://XXX # openvino yolonas base model

objects:
  track:
    - person
    - cat
    - face
  filters:
    face:
      min_score: .65
    person:
      min_score: .65
      threshold: .85
      # min_score: .55
      # threshold: .75
    cat:
      min_score: .65
      threshold: .8
      # min_score: .6
      # threshold: .7
    dog:
      min_score: .7
      threshold: .9
    package:
      min_score: .65
      threshold: .9
    car:
      min_score: .65
      threshold: .85
    license_plate:
      min_score: .6
    amazon:
      min_score: .75
    ups:
      min_score: .75
    fedex:
      min_score: .75

birdseye:
  enabled: true
  restream: true
  mode: motion
  quality: 2
  layout:
    max_cameras: 4

record:
  enabled: true
  retain:
    days: 7
    mode: all
  alerts:
    retain:
      days: 28
  detections:
    retain:
      days: 14
snapshots:
  enabled: true

detect:
  enabled: true

motion:
  mask:
    - 0,0.042,0.34,0.042,0.34,0,0,0

cameras:
  street_cam:
    ffmpeg:
      inputs:
        - path: rtsp://127.0.0.1:8554/street_cam
          input_args: preset-rtsp-restream
          roles:
            - record
            - detect
        # - path: rtsp://127.0.0.1:8554/street_cam_lq
        #   input_args: preset-rtsp-restream
        #   roles:
        #     - detect
    onvif:
      host: 10.2.3.30
      port: 2020
      user: streetcam
      password: XXX
    objects:
      track:
        - person
        - cat
        - face
      filters:
        person:
          # min_score: .65
          # threshold: .85
          min_score: .50
          threshold: .70
        cat:
          # min_score: .65
          # threshold: .8
          min_score: .5
          threshold: .65
    motion:
      mask:
        - 0,0.042,0.34,0.042,0.34,0,0,0
        - 0.793,1,0.787,0.414,0.851,0,1,0,1,0.288,1,1
        - 0.099,0.192,0.471,0.141,0.83,0.169,0.852,0.004,0.098,0
    zones:
      driveway:
        coordinates: 0.786,0.619,0.799,1,0.268,1,0,1,0.33,0.559
        inertia: 3
        loitering_time: 0
    review:
      alerts:
        required_zones: driveway
    live:
      streams:
        HQ Stream: street_cam
        LQ Stream: street_cam_lq
  lounge_cam:
    ffmpeg:
      inputs:
        - path: rtsp://127.0.0.1:8554/lounge_cam
          input_args: preset-rtsp-restream
          roles:
            - record
            - detect
        # - path: rtsp://127.0.0.1:8554/lounge_cam_lq
        #   input_args: preset-rtsp-restream
        #   roles:
        #     - detect
    onvif:
      host: 10.2.3.31
      port: 2020
      user: loungecam
      password: XXX
    live:
      streams:
        HQ Stream: lounge_cam
        LQ Stream: lounge_cam_lq
    motion:
      mask: 0,0.042,0.36,0.042,0.36,0,0,0
  sofa_cam:
    ffmpeg:
      inputs:
        - path: rtsp://127.0.0.1:8554/sofa_cam
          input_args: preset-rtsp-restream
          roles:
            - record
            - detect
        # - path: rtsp://127.0.0.1:8554/sofa_cam_lq
        #   input_args: preset-rtsp-restream
        #   roles:
        #     - detect
    onvif:
      host: 10.2.3.32
      port: 2020
      user: sofacam
      password: XXX
    live:
      streams:
        HQ Stream: sofa_cam
        LQ Stream: sofa_cam_lq
    motion:
      mask: 0,0.042,0.36,0.042,0.36,0,0,0
  spare_room_cam:
    ffmpeg:
      inputs:
        - path: rtsp://127.0.0.1:8554/spare_room_cam
          input_args: preset-rtsp-restream
          roles:
            - record
            - detect
        # - path: rtsp://127.0.0.1:8554/spare_room_cam_lq
        #   input_args: preset-rtsp-restream
        #   roles:
        #     - detect
    onvif:
      host: 10.2.3.33
      port: 2020
      user: spareroomcam
      password: XXX
    live:
      streams:
        HQ Stream: spare_room_cam
        LQ Stream: spare_room_cam_lq
  cat_cam:
    ffmpeg:
      inputs:
        - path: rtsp://127.0.0.1:8554/cat_cam
          input_args: preset-rtsp-restream
          roles:
            - record
            - detect
        # - path: rtsp://127.0.0.1:8554/cat_cam_lq
        #   input_args: preset-rtsp-restream
        #   roles:
        #     - detect
    onvif:
      host: 10.2.3.34
      port: 2020
      user: catcam
      password: XXX
    live:
      streams:
        HQ Stream: cat_cam
        LQ Stream: cat_cam_lq
  kitchen_cam:
    ffmpeg:
      inputs:
        - path: rtsp://127.0.0.1:8554/kitchen_cam
          input_args: preset-rtsp-restream
          roles:
            - record
            - detect
        # - path: rtsp://127.0.0.1:8554/kitchen_cam_lq
        #   input_args: preset-rtsp-restream
        #   roles:
        #     - detect
    onvif:
      host: 10.2.3.35
      port: 2020
      user: kitchencam
      password: XXX
    live:
      streams:
        HQ Stream: kitchen_cam
        LQ Stream: kitchen_cam_lq

go2rtc:
  webrtc:
    candidates:
      - 10.2.3.5:8555
      - stun:8555
  rtsp:
    username: frigate
    password: XXX
  streams:
    street_cam:
      - rtsp://streetcam:[email protected]:554/stream1
    street_cam_lq:
      - rtsp://streetcam:[email protected]:554/stream2
    lounge_cam:
      - rtsp://loungecam:[email protected]:554/stream1
    lounge_cam_lq:
      - rtsp://loungecam:[email protected]:554/stream2
    sofa_cam:
      - rtsp://sofacam:[email protected]:554/stream1
    sofa_cam_lq:
      - rtsp://sofacam:[email protected]:554/stream2
    spare_room_cam:
      - rtsp://spareroomcam:[email protected]:554/stream1
    spare_room_cam_lq:
      - rtsp://spareroomcam:[email protected]:554/stream2
    cat_cam:
      - rtsp://catcam:[email protected]:554/stream1
    cat_cam_lq:
      - rtsp://catcam:[email protected]:554/stream2
    kitchen_cam:
      - rtsp://kitchencam:[email protected]:554/stream1
    kitchen_cam_lq:
      - rtsp://kitchencam:[email protected]:554/stream2

version: 0.16-0
camera_groups:
  Lounge:
    order: 1
    icon: LuSofa
    cameras:
      - lounge_cam
      - sofa_cam
  Kitchen:
    order: 2
    icon: LuChefHat
    cameras: kitchen_cam
  Spare_Room:
    order: 3
    icon: LuCat
    cameras:
      - spare_room_cam
      - cat_cam
  Outside:
    order: 4
    icon: LuTrees
    cameras: street_cam
semantic_search:
  enabled: true
  reindex: false
  model_size: large
lpr:
  enabled: false
classification:
  bird:
    enabled: false

docker-compose file or Docker CLI command

services:
  frigate:
    container_name: frigate
    privileged: true # this may not be necessary for all setups
    restart: unless-stopped
    stop_grace_period: 30s # allow enough time to shut down the various services
    image: ghcr.io/blakeblackshear/frigate:stable
    shm_size: "1024mb" # update for your cameras based on calculation above
    devices:
      #- /dev/bus/usb:/dev/bus/usb # Passes the USB Coral, needs to be modified for other versions
      #- /dev/apex_0:/dev/apex_0 # Passes a PCIe Coral, follow driver instructions here https://coral.ai/docs/m2/get-started/#2a-on-linux
      #- /dev/video11:/dev/video11 # For Raspberry Pi 4B
      - /dev/dri/renderD128:/dev/dri/renderD128 # For intel hwaccel, needs to be updated for your hardware
    volumes:
      - /etc/localtime:/etc/localtime:ro
      - /share/Configuration/frigate:/config
      - /share/Media/frigate/clips:/media/frigate/clips
      - /share/Media/frigate/recordings:/media/frigate/recordings
      - /mnt/Backup/Recordings/frigate/exports:/media/frigate/exports
      - type: tmpfs # Optional: 1GB of memory, reduces SSD/SD Card wear
        target: /tmp/cache
        tmpfs:
          size: 1000000000
    ports:
      - "8971:8971"
      - "5000:5000" # Internal unauthenticated access. Expose carefully.
      - "8554:8554" # RTSP feeds
      - "8555:8555/tcp" # WebRTC over tcp
      - "8555:8555/udp" # WebRTC over udp
    environment:
      FRIGATE_RTSP_PASSWORD: "XXXX"
      PLUS_API_KEY: "XXXX"

Relevant Frigate log output

Log in nginx tab is split into two entries, but roughly like this:

Line 1: "GET /api/event_ids?ids=1762130629.869811-60r6dh,1762130618.177495-9fyyei,1762130089.813638-ot6qkx,1762129501.976676-lj884b,1762129490.183854-zy79hm,1762129463.585412-o6st7j,1762128084.977159-........
Line 2: 724-hg99o0,1762096632.56377-j71iv3,1762095802" 414 177 "-" "-" "-"

Relevant go2rtc log output

N/A

Operating system

Debian

Install method

Docker Compose

Network connection

Wired

Camera make and model

Tapo C225

Screenshots of the Frigate UI's System metrics pages

N/A

Any other information that may be helpful

No response

[inkeep-ai-assistant[bot]]

The issue you're experiencing with the 414 (Request-URI Too Large) error when clicking faces in the Face Library is related to how the /api/event_ids endpoint handles large numbers of event IDs in the query string(1).

The /api/event_ids endpoint accepts a query parameter ids that contains a comma-separated list of event IDs(1). When you have a large number of face detections associated with a particular face, the resulting URL becomes too long for nginx to process, triggering the 414 error.

This is a known limitation with HTTP GET requests that pass large amounts of data in the query string. The nginx server has a default limit on the maximum URI size it will accept.

Potential workarounds:

1. Increase nginx URI size limit: You could try increasing the nginx large_client_header_buffers setting in your Frigate configuration, though this would require modifying the internal nginx configuration which isn't directly exposed in the standard Frigate setup.

2. Use the unauthenticated port (5000): Based on the available sources, both ports 8971 (authenticated) and 5000 (unauthenticated) should handle the same API requests(2)(3). Since you mentioned this didn't occur with the Home Assistant add-on, and you've tried both ports without success, this suggests the issue is related to the number of face detections rather than the port itself.

3. Report as a bug: This appears to be a legitimate bug where the UI should either paginate the results or use a POST request instead of GET to avoid URI length limitations. The /api/event_ids endpoint currently only supports GET requests(1), which is the root cause of your issue.

The fact that this worked with the Home Assistant add-on but not with Docker Compose suggests there may be differences in how nginx is configured between the two installations, or you may have accumulated more face detections since migrating.

I recommend reporting this as a bug on the Frigate GitHub repository, as the Face Library UI should handle large numbers of detections gracefully without hitting URI length limits.

📚 Sources:

1. GET /event_ids
2. Authentication
3. Installation

---

Have another question? Just tag @inkeep.

[tombeynon]

FWIW, clearing the face library (deleting clips/faces) fixed this, but I expect it will occur again when the library becomes large enough.

[hawkeye217]

How large is your face library and what is your face_recognition config?

[tombeynon]

Only two people with 30-40 images each, but the train folder had 1,050 files which I suspect is where all those event IDs come from. My full config is above but the face recognition config is essentially default:

face_recognition:
  enabled: true
  model_size: large
  min_area: 500

[NickM-27]

the train folder should only have 100 by default, makes it seem like either an earlier beta version was being run or something else was going on

[tombeynon]

Interesting! I've never run a beta version, only the Home Assistant add-on since 0.14 if I remember correctly. I was migrating to a docker install on a new server, so copied all the existing files. I guess something must have stopped them from being cleaned up correctly previously.

Makes sense though, and I know how to fix it if it happens again in the future. Thanks so much for your help!