Skip to content

[Bug]: resolve_session_name sanitizes gateway_session_key without length truncation, overflows Honcho's 100-char session ID limit #13868

Closed
Closed
[Bug]: resolve_session_name sanitizes gateway_session_key without length truncation, overflows Honcho's 100-char session ID limit#13868
#15381 (+2)
Labels
P2Medium — degraded but workaround existscomp/gatewayGateway runner, session dispatch, deliverycomp/pluginsPlugin system and bundled pluginsplatform/matrixMatrix adapter (E2EE)type/bugSomething isn't working
@pichiu

Description

@pichiu
pichiu
opened on Apr 22, 2026

Bug Description

Matrix adapter session keys passed through HonchoClientConfig.resolve_session_name() (plugins/memory/honcho/client.py) are sanitized but not length-bounded, producing session IDs that overflow Honcho's Pydantic max_length=100 on Session.id. Bug introduced in v2026.4.16 with the new gateway_session_key code path; v2026.4.13 predates this path and does not exhibit the problem (at the cost of losing per-chat memory isolation for Matrix rooms).

Steps to Reproduce

  1. Configure hermes v2026.4.16 to use Honcho as memory provider: set memory.provider: honcho in config.yaml and place a valid honcho.json in $HERMES_HOME pointing at a reachable Honcho v3.0.6 instance.
  2. Connect the Matrix adapter to any homeserver whose sanitized hostname exceeds ~13 characters (common patterns: IP-encoded DNS like sslip.io / nip.io, Tailscale .ts.net, subdomains of public registrars).
  3. Send any message from the room to the bot.

Expected Behavior

Honcho initializes successfully; Matrix session appears in the Honcho workspace (peers / sessions / messages populated, per-chat memory works).

Actual Behavior

plugins.memory.honcho: Honcho init failed on every Matrix-origin message:

{'type': 'string_too_long',
 'loc': ['body', 'id'],
 'msg': 'String should have at most 100 characters',
 'ctx': {'max_length': 100}}

Workspace ends up with 0 peers / 0 sessions — memory is effectively disabled for Matrix chats.

Affected Component

Gateway (Telegram/Discord/Slack/WhatsApp), Agent Core (conversation loop, context compression, memory)

Messaging Platform (if gateway-related)

N/A (CLI only)

Debug Report

--- hermes dump ---
version:          0.10.0 (2026.4.16) [(unknown)]
os:               Linux 6.8.0-100-generic x86_64
python:           3.13.5
openai_sdk:       2.32.0
profile:          default
hermes_home:      /opt/data
model:            us.anthropic.claude-opus-4-6-v1
provider:         custom
terminal:         local

api_keys:
  openrouter           set
  openai               not set
  anthropic            not set
  anthropic_token      not set
  nous                 not set
  glm/zai              not set
  zai                  not set
  kimi                 not set
  minimax              not set
  deepseek             not set
  dashscope            not set
  huggingface          not set
  ai_gateway           not set
  opencode_zen         not set
  opencode_go          not set
  kilocode             not set
  firecrawl            not set
  tavily               not set
  browserbase          not set
  fal                  not set
  elevenlabs           not set
  github               set

features:
  toolsets:           hermes-cli
  mcp_servers:        0
  memory_provider:    honcho
  gateway:            running (docker, pid 1)
  platforms:          none
  cron_jobs:          0
  skills:             79
--- end dump ---


--- agent.log (last 5000 lines) ---
  <... 2968 kube-probe /health 200 lines elided ...>
2026-04-22 02:49:58,425 INFO gateway.run: Received SIGTERM/SIGINT — initiating shutdown
2026-04-22 02:49:58,431 INFO gateway.run: Shutdown diagnostic — no other hermes processes found
2026-04-22 02:49:58,431 INFO gateway.run: Stopping gateway...
2026-04-22 02:49:58,433 INFO gateway.platforms.matrix: Matrix: disconnected
2026-04-22 02:49:58,433 INFO gateway.run: ✓ matrix disconnected
2026-04-22 02:49:58,434 INFO gateway.platforms.api_server: [Api_Server] API server stopped
2026-04-22 02:49:58,434 INFO gateway.run: ✓ api_server disconnected
2026-04-22 02:49:58,437 INFO gateway.run: Gateway stopped
2026-04-22 02:49:58,437 INFO gateway.run: Cron ticker stopped
2026-04-22 02:49:58,437 INFO gateway.run: Exiting with code 1 (signal-initiated shutdown without restart request) so systemd Restart=on-failure can revive the gateway.
2026-04-22 02:50:09,464 INFO gateway.run: Starting Hermes Gateway...
2026-04-22 02:50:09,464 INFO gateway.run: Session storage: /opt/data/sessions
2026-04-22 02:50:09,470 INFO gateway.run: Previous gateway exited cleanly — skipping session suspension
2026-04-22 02:50:09,545 INFO gateway.run: Connecting to matrix...
2026-04-22 02:50:09,615 INFO gateway.platforms.matrix: Matrix: using access token for @<bot>:<HOMESERVER> (device <DEVICE_ID>)
2026-04-22 02:50:09,730 INFO gateway.platforms.matrix: Matrix: initial sync complete, joined 3 rooms
2026-04-22 02:50:09,747 INFO gateway.run: ✓ matrix connected
2026-04-22 02:50:09,792 INFO gateway.run: Connecting to api_server...
2026-04-22 02:50:09,794 INFO gateway.platforms.api_server: [Api_Server] API server listening on http://<IP>:8642 (model: hermes-agent)
2026-04-22 02:50:09,795 INFO gateway.run: ✓ api_server connected
2026-04-22 02:50:09,795 INFO gateway.run: 1 hook(s) loaded
2026-04-22 02:50:09,795 INFO gateway.run: Gateway running with 2 platform(s)
2026-04-22 02:50:09,797 INFO gateway.run: Channel directory built: 4 target(s)
2026-04-22 02:50:09,798 INFO gateway.run: Press Ctrl+C to stop
2026-04-22 02:50:09,798 INFO gateway.run: Cron ticker started (interval=60s)
  <... 36 kube-probe /health 200 lines elided ...>
2026-04-22 02:54:05,678 INFO gateway.platforms.matrix: [Matrix] Flushing text batch agent:main:matrix:group:!<ROOM_18>:<HOMESERVER>:$<EVENT_ID_43> (7 chars)
2026-04-22 02:54:06,137 INFO gateway.run: inbound message: platform=matrix user=<user> chat=!<ROOM_18>:<HOMESERVER> msg='<REDACTED_USER_MESSAGE>'
2026-04-22 02:54:06,592 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 02:54:07,033 INFO run_agent: Loaded environment variables from /opt/data/.env
2026-04-22 02:54:07,796 INFO agent.auxiliary_client: Vision auto-detect: using active provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 02:54:07,803 INFO agent.memory_manager: Memory provider 'honcho' registered (4 tools)
2026-04-22 02:54:07,866 INFO plugins.memory.honcho.client: Initializing Honcho client (base_url: http://honcho-13-api.honcho.svc.cluster.local:8000, workspace: hermes-ricer)
2026-04-22 02:54:07,899 INFO plugins.memory.honcho.session: Honcho session 'hermes' created (new)
2026-04-22 02:54:07,908 INFO plugins.memory.honcho.session: Uploaded MEMORY.md to Honcho for hermes (user peer)
2026-04-22 02:54:10,028 INFO agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 02:54:10,039 INFO agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 4 kube-probe /health 200 lines elided ...>
2026-04-22 02:54:30,230 INFO agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 02:54:30,230 INFO agent.auxiliary_client: Auxiliary title_generation: using auto (us.anthropic.claude-opus-4-6-v1) at http://litellm-service.litellm:4000
2026-04-22 02:54:30,242 INFO gateway.run: response ready: platform=matrix chat=!<ROOM_18>:<HOMESERVER> time=24.1s api_calls=1 response=24 chars
2026-04-22 02:54:30,246 INFO gateway.platforms.base: [Matrix] Sending response (24 chars) to !<ROOM_18>:<HOMESERVER>
  <... 1 kube-probe /health 200 lines elided ...>
2026-04-22 02:54:30,765 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 02:54:31,028 INFO gateway.platforms.matrix: Matrix: redacted $<EVENT_ID_43> in !<ROOM_18>:<HOMESERVER>
2026-04-22 02:54:33,457 WARNING plugins.memory.honcho.session: Honcho dialectic query failed: An unexpected error occurred
  <... 5 kube-probe /health 200 lines elided ...>
2026-04-22 02:54:55,756 WARNING plugins.memory.honcho.session: Honcho dialectic query failed: An unexpected error occurred
  <... 268 kube-probe /health 200 lines elided ...>
2026-04-22 03:17:19,712 INFO agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 03:17:19,713 INFO gateway.platforms.base: [Matrix] Sending response (266 chars) to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:17:19,729 INFO [20260422_025406_536aa103] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 03:17:19,734 INFO [20260422_025406_536aa103] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 03:17:19,971 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:17:20,225 INFO gateway.platforms.matrix: Matrix: redacted $<EVENT_ID_43> in !<ROOM_18>:<HOMESERVER>
  <... 3 kube-probe /health 200 lines elided ...>
2026-04-22 03:17:33,761 INFO [20260421_090836_9c0f6f17] gateway.run: Pre-reset memory flush completed for session 20260421_090836_9c0f6f17
  <... 3 kube-probe /health 200 lines elided ...>
2026-04-22 03:17:49,361 INFO gateway.platforms.matrix: [Matrix] Flushing text batch agent:main:matrix:dm:!<ROOM_18>:<HOMESERVER> (14 chars)
2026-04-22 03:17:49,645 INFO gateway.run: inbound message: platform=matrix user=<user> chat=!<ROOM_18>:<HOMESERVER> msg='<REDACTED_USER_MESSAGE>'
2026-04-22 03:17:49,808 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:17:49,839 INFO [20260421_090836_9c0f6f17] agent.auxiliary_client: Vision auto-detect: using active provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 03:17:49,843 INFO [20260421_090836_9c0f6f17] agent.memory_manager: Memory provider 'honcho' registered (4 tools)
2026-04-22 03:17:49,884 INFO [20260421_090836_9c0f6f17] plugins.memory.honcho.session: Honcho session 'hermes' retrieved (3 existing messages)
2026-04-22 03:17:49,885 INFO [20260421_090836_9c0f6f17] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 03:17:49,893 INFO [20260421_090836_9c0f6f17] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 3 kube-probe /health 200 lines elided ...>
2026-04-22 03:18:02,509 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:18:02,820 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
  <... 6 kube-probe /health 200 lines elided ...>
2026-04-22 03:18:32,780 INFO agent.auxiliary_client: Auxiliary title_generation: using auto (us.anthropic.claude-opus-4-6-v1) at http://litellm-service.litellm:4000
2026-04-22 03:18:32,824 INFO gateway.run: response ready: platform=matrix chat=!<ROOM_18>:<HOMESERVER> time=43.2s api_calls=2 response=2206 chars
2026-04-22 03:18:32,828 INFO gateway.platforms.base: [Matrix] Sending response (2206 chars) to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:18:33,231 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:18:33,431 INFO gateway.platforms.matrix: Matrix: redacted $<EVENT_ID_43> in !<ROOM_18>:<HOMESERVER>
  <... 62 kube-probe /health 200 lines elided ...>
2026-04-22 03:23:43,224 INFO aiohttp.access: <IP> [22/Apr/2026:03:23:43 +0000] "GET / HTTP/1.1" 404 174 "-" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/<IP> Safari/537.36 Edg/<IP>"
2026-04-22 03:23:43,436 INFO aiohttp.access: <IP> [22/Apr/2026:03:23:43 +0000] "GET /favicon.ico HTTP/1.1" 404 174 "http://<IP>:30790/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/<IP> Safari/537.36 Edg/<IP>"
  <... 10 kube-probe /health 200 lines elided ...>
2026-04-22 03:24:31,403 INFO aiohttp.access: <IP> [22/Apr/2026:03:24:31 +0000] "GET / HTTP/1.1" 404 174 "-" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/<IP> Safari/537.36 Edg/<IP>"
  <... 441 kube-probe /health 200 lines elided ...>
2026-04-22 04:01:16,141 INFO gateway.run: Session expiry: 5 sessions to flush (matrix:5)
2026-04-22 04:01:19,183 INFO [20260422_031717_2490e5a0] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 04:01:19,193 INFO [20260422_031717_2490e5a0] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 3 kube-probe /health 200 lines elided ...>
2026-04-22 04:01:32,707 INFO [20260422_031717_2490e5a0] gateway.run: Pre-reset memory flush completed for session 20260422_031717_2490e5a0
2026-04-22 04:01:32,732 INFO [20260422_031717_2490e5a0] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 04:01:32,738 INFO [20260422_031717_2490e5a0] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 8 kube-probe /health 200 lines elided ...>
2026-04-22 04:02:12,587 INFO [20260421_083642_16102b11] gateway.run: Pre-reset memory flush completed for session 20260421_083642_16102b11
2026-04-22 04:02:12,607 INFO [20260421_083642_16102b11] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 04:02:12,612 INFO [20260421_083642_16102b11] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 8 kube-probe /health 200 lines elided ...>
2026-04-22 04:02:51,657 INFO [20260421_112241_81926a4f] gateway.run: Pre-reset memory flush completed for session 20260421_112241_81926a4f
2026-04-22 04:02:51,674 INFO [20260421_112241_81926a4f] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 04:02:51,679 INFO [20260421_112241_81926a4f] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 5 kube-probe /health 200 lines elided ...>
2026-04-22 04:03:14,199 INFO [20260421_112614_5b9f7f35] gateway.run: Pre-reset memory flush completed for session 20260421_112614_5b9f7f35
2026-04-22 04:03:14,202 INFO gateway.run: Session expiry done: 5 flushed
  <... 1019 kube-probe /health 200 lines elided ...>
2026-04-22 05:28:13,125 INFO gateway.run: Stopping gateway...
2026-04-22 05:28:13,126 INFO gateway.platforms.matrix: Matrix: disconnected
2026-04-22 05:28:13,127 INFO gateway.run: ✓ matrix disconnected
2026-04-22 05:28:13,128 INFO gateway.platforms.api_server: [Api_Server] API server stopped
2026-04-22 05:28:13,128 INFO gateway.run: ✓ api_server disconnected
2026-04-22 05:28:13,131 INFO gateway.run: Gateway stopped
2026-04-22 05:28:13,131 INFO gateway.run: Cron ticker stopped
2026-04-22 05:28:23,334 INFO gateway.run: Starting Hermes Gateway...
2026-04-22 05:28:23,334 INFO gateway.run: Session storage: /opt/data/sessions
2026-04-22 05:28:23,339 INFO gateway.run: Previous gateway exited cleanly — skipping session suspension
2026-04-22 05:28:23,404 INFO gateway.run: Connecting to matrix...
2026-04-22 05:28:23,480 INFO gateway.platforms.matrix: Matrix: using access token for @<bot>:<HOMESERVER> (device <DEVICE_ID>)
2026-04-22 05:28:23,592 INFO gateway.platforms.matrix: Matrix: initial sync complete, joined 3 rooms
2026-04-22 05:28:23,608 INFO gateway.run: ✓ matrix connected
2026-04-22 05:28:23,663 INFO gateway.run: Connecting to api_server...
2026-04-22 05:28:23,667 INFO gateway.platforms.api_server: [Api_Server] API server listening on http://<IP>:8642 (model: hermes-agent)
2026-04-22 05:28:23,668 INFO gateway.run: ✓ api_server connected
2026-04-22 05:28:23,669 INFO gateway.run: 1 hook(s) loaded
2026-04-22 05:28:23,669 INFO gateway.run: Gateway running with 2 platform(s)
2026-04-22 05:28:23,674 INFO gateway.run: Channel directory built: 5 target(s)
2026-04-22 05:28:23,675 INFO gateway.run: Press Ctrl+C to stop
2026-04-22 05:28:23,676 INFO gateway.run: Cron ticker started (interval=60s)
  <... 17 kube-probe /health 200 lines elided ...>
2026-04-22 05:30:44,902 INFO gateway.platforms.matrix: [Matrix] Flushing text batch agent:main:matrix:group:!<ROOM_18>:<HOMESERVER>:$<EVENT_ID_43> (8 chars)
2026-04-22 05:30:45,525 INFO gateway.run: inbound message: platform=matrix user=<user> chat=!<ROOM_18>:<HOMESERVER> msg='<REDACTED_USER_MESSAGE>'
2026-04-22 05:30:45,962 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 05:30:46,402 INFO run_agent: Loaded environment variables from /opt/data/.env
2026-04-22 05:30:47,218 INFO agent.auxiliary_client: Vision auto-detect: using active provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 05:30:47,230 INFO agent.memory_manager: Memory provider 'honcho' registered (5 tools)
2026-04-22 05:30:47,304 INFO plugins.memory.honcho.client: Initializing Honcho client (base_url: http://honcho-13-api.honcho.svc.cluster.local:8000, workspace: hermes-ricer)
2026-04-22 05:30:47,324 WARNING plugins.memory.honcho: Honcho init failed: [{'type': 'string_too_long', 'loc': ['body', 'id'], 'msg': 'String should have at most 100 characters', 'input': '<SESSION_ID_SANITIZED length=113 chars — reproduces on v2026.4.16>', 'ctx': {'max_length': 100}}]
2026-04-22 05:30:47,437 INFO agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 05:30:47,443 INFO agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 2 kube-probe /health 200 lines elided ...>
2026-04-22 05:30:54,375 INFO agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 05:30:54,375 INFO agent.auxiliary_client: Auxiliary title_generation: using auto (us.anthropic.claude-opus-4-6-v1) at http://litellm-service.litellm:4000
2026-04-22 05:30:54,427 INFO gateway.run: response ready: platform=matrix chat=!<ROOM_18>:<HOMESERVER> time=8.9s api_calls=1 response=41 chars
2026-04-22 05:30:54,432 INFO gateway.platforms.base: [Matrix] Sending response (41 chars) to !<ROOM_18>:<HOMESERVER>
2026-04-22 05:30:54,714 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 05:30:54,939 INFO gateway.platforms.matrix: Matrix: redacted $<EVENT_ID_43> in !<ROOM_18>:<HOMESERVER>
  <... 7 kube-probe /health 200 lines elided ...>

--- errors.log (last 100 lines) ---
2026-04-21 07:34:43,430 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:34:43,430 WARNING gateway.run: ✗ api_server failed to connect
2026-04-21 07:35:13,465 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:36:13,520 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:36:52,516 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:36:52,516 WARNING gateway.run: ✗ api_server failed to connect
2026-04-21 07:37:22,548 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:38:22,600 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:39:03,528 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:39:03,528 WARNING gateway.run: ✗ api_server failed to connect
2026-04-21 07:39:33,565 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:40:33,616 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:41:12,974 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:41:12,974 WARNING gateway.run: ✗ api_server failed to connect
2026-04-21 07:41:43,007 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 07:42:43,056 ERROR gateway.platforms.api_server: [Api_Server] Refusing to start: binding to <IP> requires API_SERVER_KEY. Set API_SERVER_KEY or use the default <IP>.
2026-04-21 08:03:40,233 WARNING plugins.memory.honcho: Honcho init failed: Connection error: [Errno -2] Name or service not known
2026-04-21 08:33:17,525 WARNING plugins.memory.honcho: Honcho init failed: Connection error: [Errno -2] Name or service not known
2026-04-21 08:35:59,530 ERROR gateway.platforms.matrix: Matrix: failed to send to !<ROOM_18>: User @<bot>:<HOMESERVER> not in room !<ROOM_18>
2026-04-21 08:36:46,774 WARNING plugins.memory.honcho: Honcho init failed: [{'type': 'string_too_long', 'loc': ['body', 'id'], 'msg': 'String should have at most 100 characters', 'input': '<SESSION_ID_SANITIZED length=113 chars — reproduces on v2026.4.16>', 'ctx': {'max_length': 100}}]
2026-04-21 09:09:09,696 WARNING plugins.memory.honcho.session: Honcho dialectic query failed: An unexpected error occurred
2026-04-21 09:09:09,725 WARNING plugins.memory.honcho.session: Honcho dialectic query failed: An unexpected error occurred
2026-04-21 09:09:38,374 WARNING plugins.memory.honcho.session: Honcho dialectic query failed: An unexpected error occurred
2026-04-21 11:16:48,174 WARNING plugins.memory.honcho: Honcho init failed: [{'type': 'string_too_long', 'loc': ['body', 'id'], 'msg': 'String should have at most 100 characters', 'input': '<SESSION_ID_SANITIZED length=113 chars — reproduces on v2026.4.16>', 'ctx': {'max_length': 100}}]
2026-04-21 11:22:41,272 WARNING [20260421_083642_16102b11] plugins.memory.honcho: Honcho init failed: [{'type': 'string_too_long', 'loc': ['body', 'id'], 'msg': 'String should have at most 100 characters', 'input': '<SESSION_ID_SANITIZED length=113 chars — reproduces on v2026.4.16>', 'ctx': {'max_length': 100}}]
2026-04-21 11:26:15,253 WARNING [20260421_112241_81926a4f] plugins.memory.honcho: Honcho init failed: [{'type': 'string_too_long', 'loc': ['body', 'id'], 'msg': 'String should have at most 100 characters', 'input': '<SESSION_ID_SANITIZED length=113 chars — reproduces on v2026.4.16>', 'ctx': {'max_length': 100}}]
2026-04-22 02:54:33,457 WARNING plugins.memory.honcho.session: Honcho dialectic query failed: An unexpected error occurred
2026-04-22 02:54:55,756 WARNING plugins.memory.honcho.session: Honcho dialectic query failed: An unexpected error occurred
2026-04-22 05:30:47,324 WARNING plugins.memory.honcho: Honcho init failed: [{'type': 'string_too_long', 'loc': ['body', 'id'], 'msg': 'String should have at most 100 characters', 'input': '<SESSION_ID_SANITIZED length=113 chars — reproduces on v2026.4.16>', 'ctx': {'max_length': 100}}]

--- gateway.log (last 100 lines) ---
(file not found)



============================================================
FULL agent.log
============================================================

--- hermes dump ---
version:          0.10.0 (2026.4.16) [(unknown)]
os:               Linux 6.8.0-100-generic x86_64
python:           3.13.5
openai_sdk:       2.32.0
profile:          default
hermes_home:      /opt/data
model:            us.anthropic.claude-opus-4-6-v1
provider:         custom
terminal:         local

api_keys:
  openrouter           set
  openai               not set
  anthropic            not set
  anthropic_token      not set
  nous                 not set
  glm/zai              not set
  zai                  not set
  kimi                 not set
  minimax              not set
  deepseek             not set
  dashscope            not set
  huggingface          not set
  ai_gateway           not set
  opencode_zen         not set
  opencode_go          not set
  kilocode             not set
  firecrawl            not set
  tavily               not set
  browserbase          not set
  fal                  not set
  elevenlabs           not set
  github               set

features:
  toolsets:           hermes-cli
  mcp_servers:        0
  memory_provider:    honcho
  gateway:            running (docker, pid 1)
  platforms:          none
  cron_jobs:          0
  skills:             79
--- end dump ---


--- full agent.log ---
[... truncated — showing last ~500KB ...]
  <... 1704 kube-probe /health 200 lines elided ...>
2026-04-22 02:49:58,425 INFO gateway.run: Received SIGTERM/SIGINT — initiating shutdown
2026-04-22 02:49:58,431 INFO gateway.run: Shutdown diagnostic — no other hermes processes found
2026-04-22 02:49:58,431 INFO gateway.run: Stopping gateway...
2026-04-22 02:49:58,433 INFO gateway.platforms.matrix: Matrix: disconnected
2026-04-22 02:49:58,433 INFO gateway.run: ✓ matrix disconnected
2026-04-22 02:49:58,434 INFO gateway.platforms.api_server: [Api_Server] API server stopped
2026-04-22 02:49:58,434 INFO gateway.run: ✓ api_server disconnected
2026-04-22 02:49:58,437 INFO gateway.run: Gateway stopped
2026-04-22 02:49:58,437 INFO gateway.run: Cron ticker stopped
2026-04-22 02:49:58,437 INFO gateway.run: Exiting with code 1 (signal-initiated shutdown without restart request) so systemd Restart=on-failure can revive the gateway.
2026-04-22 02:50:09,464 INFO gateway.run: Starting Hermes Gateway...
2026-04-22 02:50:09,464 INFO gateway.run: Session storage: /opt/data/sessions
2026-04-22 02:50:09,470 INFO gateway.run: Previous gateway exited cleanly — skipping session suspension
2026-04-22 02:50:09,545 INFO gateway.run: Connecting to matrix...
2026-04-22 02:50:09,615 INFO gateway.platforms.matrix: Matrix: using access token for @<bot>:<HOMESERVER> (device <DEVICE_ID>)
2026-04-22 02:50:09,730 INFO gateway.platforms.matrix: Matrix: initial sync complete, joined 3 rooms
2026-04-22 02:50:09,747 INFO gateway.run: ✓ matrix connected
2026-04-22 02:50:09,792 INFO gateway.run: Connecting to api_server...
2026-04-22 02:50:09,794 INFO gateway.platforms.api_server: [Api_Server] API server listening on http://<IP>:8642 (model: hermes-agent)
2026-04-22 02:50:09,795 INFO gateway.run: ✓ api_server connected
2026-04-22 02:50:09,795 INFO gateway.run: 1 hook(s) loaded
2026-04-22 02:50:09,795 INFO gateway.run: Gateway running with 2 platform(s)
2026-04-22 02:50:09,797 INFO gateway.run: Channel directory built: 4 target(s)
2026-04-22 02:50:09,798 INFO gateway.run: Press Ctrl+C to stop
2026-04-22 02:50:09,798 INFO gateway.run: Cron ticker started (interval=60s)
  <... 36 kube-probe /health 200 lines elided ...>
2026-04-22 02:54:05,678 INFO gateway.platforms.matrix: [Matrix] Flushing text batch agent:main:matrix:group:!<ROOM_18>:<HOMESERVER>:$<EVENT_ID_43> (7 chars)
2026-04-22 02:54:06,137 INFO gateway.run: inbound message: platform=matrix user=<user> chat=!<ROOM_18>:<HOMESERVER> msg='<REDACTED_USER_MESSAGE>'
2026-04-22 02:54:06,592 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 02:54:07,033 INFO run_agent: Loaded environment variables from /opt/data/.env
2026-04-22 02:54:07,796 INFO agent.auxiliary_client: Vision auto-detect: using active provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 02:54:07,803 INFO agent.memory_manager: Memory provider 'honcho' registered (4 tools)
2026-04-22 02:54:07,866 INFO plugins.memory.honcho.client: Initializing Honcho client (base_url: http://honcho-13-api.honcho.svc.cluster.local:8000, workspace: hermes-ricer)
2026-04-22 02:54:07,899 INFO plugins.memory.honcho.session: Honcho session 'hermes' created (new)
2026-04-22 02:54:07,908 INFO plugins.memory.honcho.session: Uploaded MEMORY.md to Honcho for hermes (user peer)
2026-04-22 02:54:10,028 INFO agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 02:54:10,039 INFO agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 4 kube-probe /health 200 lines elided ...>
2026-04-22 02:54:30,230 INFO agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 02:54:30,230 INFO agent.auxiliary_client: Auxiliary title_generation: using auto (us.anthropic.claude-opus-4-6-v1) at http://litellm-service.litellm:4000
2026-04-22 02:54:30,242 INFO gateway.run: response ready: platform=matrix chat=!<ROOM_18>:<HOMESERVER> time=24.1s api_calls=1 response=24 chars
2026-04-22 02:54:30,246 INFO gateway.platforms.base: [Matrix] Sending response (24 chars) to !<ROOM_18>:<HOMESERVER>
  <... 1 kube-probe /health 200 lines elided ...>
2026-04-22 02:54:30,765 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 02:54:31,028 INFO gateway.platforms.matrix: Matrix: redacted $<EVENT_ID_43> in !<ROOM_18>:<HOMESERVER>
2026-04-22 02:54:33,457 WARNING plugins.memory.honcho.session: Honcho dialectic query failed: An unexpected error occurred
  <... 5 kube-probe /health 200 lines elided ...>
2026-04-22 02:54:55,756 WARNING plugins.memory.honcho.session: Honcho dialectic query failed: An unexpected error occurred
  <... 268 kube-probe /health 200 lines elided ...>
2026-04-22 03:17:19,712 INFO agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 03:17:19,713 INFO gateway.platforms.base: [Matrix] Sending response (266 chars) to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:17:19,729 INFO [20260422_025406_536aa103] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 03:17:19,734 INFO [20260422_025406_536aa103] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 03:17:19,971 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:17:20,225 INFO gateway.platforms.matrix: Matrix: redacted $<EVENT_ID_43> in !<ROOM_18>:<HOMESERVER>
  <... 3 kube-probe /health 200 lines elided ...>
2026-04-22 03:17:33,761 INFO [20260421_090836_9c0f6f17] gateway.run: Pre-reset memory flush completed for session 20260421_090836_9c0f6f17
  <... 3 kube-probe /health 200 lines elided ...>
2026-04-22 03:17:49,361 INFO gateway.platforms.matrix: [Matrix] Flushing text batch agent:main:matrix:dm:!<ROOM_18>:<HOMESERVER> (14 chars)
2026-04-22 03:17:49,645 INFO gateway.run: inbound message: platform=matrix user=<user> chat=!<ROOM_18>:<HOMESERVER> msg='<REDACTED_USER_MESSAGE>'
2026-04-22 03:17:49,808 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:17:49,839 INFO [20260421_090836_9c0f6f17] agent.auxiliary_client: Vision auto-detect: using active provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 03:17:49,843 INFO [20260421_090836_9c0f6f17] agent.memory_manager: Memory provider 'honcho' registered (4 tools)
2026-04-22 03:17:49,884 INFO [20260421_090836_9c0f6f17] plugins.memory.honcho.session: Honcho session 'hermes' retrieved (3 existing messages)
2026-04-22 03:17:49,885 INFO [20260421_090836_9c0f6f17] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 03:17:49,893 INFO [20260421_090836_9c0f6f17] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 3 kube-probe /health 200 lines elided ...>
2026-04-22 03:18:02,509 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:18:02,820 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
  <... 6 kube-probe /health 200 lines elided ...>
2026-04-22 03:18:32,780 INFO agent.auxiliary_client: Auxiliary title_generation: using auto (us.anthropic.claude-opus-4-6-v1) at http://litellm-service.litellm:4000
2026-04-22 03:18:32,824 INFO gateway.run: response ready: platform=matrix chat=!<ROOM_18>:<HOMESERVER> time=43.2s api_calls=2 response=2206 chars
2026-04-22 03:18:32,828 INFO gateway.platforms.base: [Matrix] Sending response (2206 chars) to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:18:33,231 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 03:18:33,431 INFO gateway.platforms.matrix: Matrix: redacted $<EVENT_ID_43> in !<ROOM_18>:<HOMESERVER>
  <... 62 kube-probe /health 200 lines elided ...>
2026-04-22 03:23:43,224 INFO aiohttp.access: <IP> [22/Apr/2026:03:23:43 +0000] "GET / HTTP/1.1" 404 174 "-" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/<IP> Safari/537.36 Edg/<IP>"
2026-04-22 03:23:43,436 INFO aiohttp.access: <IP> [22/Apr/2026:03:23:43 +0000] "GET /favicon.ico HTTP/1.1" 404 174 "http://<IP>:30790/" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/<IP> Safari/537.36 Edg/<IP>"
  <... 10 kube-probe /health 200 lines elided ...>
2026-04-22 03:24:31,403 INFO aiohttp.access: <IP> [22/Apr/2026:03:24:31 +0000] "GET / HTTP/1.1" 404 174 "-" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/<IP> Safari/537.36 Edg/<IP>"
  <... 441 kube-probe /health 200 lines elided ...>
2026-04-22 04:01:16,141 INFO gateway.run: Session expiry: 5 sessions to flush (matrix:5)
2026-04-22 04:01:19,183 INFO [20260422_031717_2490e5a0] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 04:01:19,193 INFO [20260422_031717_2490e5a0] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 3 kube-probe /health 200 lines elided ...>
2026-04-22 04:01:32,707 INFO [20260422_031717_2490e5a0] gateway.run: Pre-reset memory flush completed for session 20260422_031717_2490e5a0
2026-04-22 04:01:32,732 INFO [20260422_031717_2490e5a0] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 04:01:32,738 INFO [20260422_031717_2490e5a0] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 8 kube-probe /health 200 lines elided ...>
2026-04-22 04:02:12,587 INFO [20260421_083642_16102b11] gateway.run: Pre-reset memory flush completed for session 20260421_083642_16102b11
2026-04-22 04:02:12,607 INFO [20260421_083642_16102b11] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 04:02:12,612 INFO [20260421_083642_16102b11] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 8 kube-probe /health 200 lines elided ...>
2026-04-22 04:02:51,657 INFO [20260421_112241_81926a4f] gateway.run: Pre-reset memory flush completed for session 20260421_112241_81926a4f
2026-04-22 04:02:51,674 INFO [20260421_112241_81926a4f] agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 04:02:51,679 INFO [20260421_112241_81926a4f] agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 5 kube-probe /health 200 lines elided ...>
2026-04-22 04:03:14,199 INFO [20260421_112614_5b9f7f35] gateway.run: Pre-reset memory flush completed for session 20260421_112614_5b9f7f35
2026-04-22 04:03:14,202 INFO gateway.run: Session expiry done: 5 flushed
  <... 1019 kube-probe /health 200 lines elided ...>
2026-04-22 05:28:13,125 INFO gateway.run: Stopping gateway...
2026-04-22 05:28:13,126 INFO gateway.platforms.matrix: Matrix: disconnected
2026-04-22 05:28:13,127 INFO gateway.run: ✓ matrix disconnected
2026-04-22 05:28:13,128 INFO gateway.platforms.api_server: [Api_Server] API server stopped
2026-04-22 05:28:13,128 INFO gateway.run: ✓ api_server disconnected
2026-04-22 05:28:13,131 INFO gateway.run: Gateway stopped
2026-04-22 05:28:13,131 INFO gateway.run: Cron ticker stopped
2026-04-22 05:28:23,334 INFO gateway.run: Starting Hermes Gateway...
2026-04-22 05:28:23,334 INFO gateway.run: Session storage: /opt/data/sessions
2026-04-22 05:28:23,339 INFO gateway.run: Previous gateway exited cleanly — skipping session suspension
2026-04-22 05:28:23,404 INFO gateway.run: Connecting to matrix...
2026-04-22 05:28:23,480 INFO gateway.platforms.matrix: Matrix: using access token for @<bot>:<HOMESERVER> (device <DEVICE_ID>)
2026-04-22 05:28:23,592 INFO gateway.platforms.matrix: Matrix: initial sync complete, joined 3 rooms
2026-04-22 05:28:23,608 INFO gateway.run: ✓ matrix connected
2026-04-22 05:28:23,663 INFO gateway.run: Connecting to api_server...
2026-04-22 05:28:23,667 INFO gateway.platforms.api_server: [Api_Server] API server listening on http://<IP>:8642 (model: hermes-agent)
2026-04-22 05:28:23,668 INFO gateway.run: ✓ api_server connected
2026-04-22 05:28:23,669 INFO gateway.run: 1 hook(s) loaded
2026-04-22 05:28:23,669 INFO gateway.run: Gateway running with 2 platform(s)
2026-04-22 05:28:23,674 INFO gateway.run: Channel directory built: 5 target(s)
2026-04-22 05:28:23,675 INFO gateway.run: Press Ctrl+C to stop
2026-04-22 05:28:23,676 INFO gateway.run: Cron ticker started (interval=60s)
  <... 17 kube-probe /health 200 lines elided ...>
2026-04-22 05:30:44,902 INFO gateway.platforms.matrix: [Matrix] Flushing text batch agent:main:matrix:group:!<ROOM_18>:<HOMESERVER>:$<EVENT_ID_43> (8 chars)
2026-04-22 05:30:45,525 INFO gateway.run: inbound message: platform=matrix user=<user> chat=!<ROOM_18>:<HOMESERVER> msg='<REDACTED_USER_MESSAGE>'
2026-04-22 05:30:45,962 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 05:30:46,402 INFO run_agent: Loaded environment variables from /opt/data/.env
2026-04-22 05:30:47,218 INFO agent.auxiliary_client: Vision auto-detect: using active provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 05:30:47,230 INFO agent.memory_manager: Memory provider 'honcho' registered (5 tools)
2026-04-22 05:30:47,304 INFO plugins.memory.honcho.client: Initializing Honcho client (base_url: http://honcho-13-api.honcho.svc.cluster.local:8000, workspace: hermes-ricer)
2026-04-22 05:30:47,324 WARNING plugins.memory.honcho: Honcho init failed: [{'type': 'string_too_long', 'loc': ['body', 'id'], 'msg': 'String should have at most 100 characters', 'input': '<SESSION_ID_SANITIZED length=113 chars — reproduces on v2026.4.16>', 'ctx': {'max_length': 100}}]
2026-04-22 05:30:47,437 INFO agent.model_metadata: Could not detect context length for model 'us.anthropic.claude-opus-4-6-v1' at http://litellm-service.litellm:4000 — defaulting to 128,000 tokens (probe-down). Set model.context_length in config.yaml to override.
2026-04-22 05:30:47,443 INFO agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
  <... 2 kube-probe /health 200 lines elided ...>
2026-04-22 05:30:54,375 INFO agent.auxiliary_client: Auxiliary auto-detect: using main provider custom (us.anthropic.claude-opus-4-6-v1)
2026-04-22 05:30:54,375 INFO agent.auxiliary_client: Auxiliary title_generation: using auto (us.anthropic.claude-opus-4-6-v1) at http://litellm-service.litellm:4000
2026-04-22 05:30:54,427 INFO gateway.run: response ready: platform=matrix chat=!<ROOM_18>:<HOMESERVER> time=8.9s api_calls=1 response=41 chars
2026-04-22 05:30:54,432 INFO gateway.platforms.base: [Matrix] Sending response (41 chars) to !<ROOM_18>:<HOMESERVER>
2026-04-22 05:30:54,714 INFO gateway.platforms.matrix: Matrix: sent event $<EVENT_ID_43> to !<ROOM_18>:<HOMESERVER>
2026-04-22 05:30:54,939 INFO gateway.platforms.matrix: Matrix: redacted $<EVENT_ID_43> in !<ROOM_18>:<HOMESERVER>
  <... 7 kube-probe /health 200 lines elided ...>

Operating System

Debian GNU/Linux 13 (trixie)

Python Version

3.13.5

Hermes Version

v0.10.0 (2026.4.16)

Additional Logs / Traceback (optional)



Root Cause Analysis (optional)


plugins/memory/honcho/client.py :: HonchoClientConfig.resolve_session_name returns the incoming gateway_session_key after sanitizing but without any length bound:


if gateway_session_key:
   sanitized = re.sub(r'[^a-zA-Z0-9_-]+', '-', gateway_session_key).strip('-')
   if sanitized:
       return sanitized


For Matrix, the adapter builds the key as agent:main:matrix:group:!<room>:<homeserver>:<event-id>. After sanitization:


len = 26 + R + HS + EV
      ^    ^   ^    ^ event ID hash length (Matrix spec, room v4+: 43 chars)
      |    |   homeserver hostname length (dots → hyphens, length unchanged)
      |    room localpart length (Synapse/Dendrite/Conduit: 18 chars)
      "agent-main-matrix-group" (23) + 3 separators



Two of the three variables are effectively frozen by Matrix protocol / reference implementation:


    

  • EV ≈ 43: Matrix spec (room version ≥ 4, mandatory since 2019) defines event ID as $ + base64url_unpadded(sha256(canonical_json(event))) → always 44 chars including $, 43 after sanitize. Cryptographic, cannot be shortened without losing uniqueness / tamper-evidence.
    • 

  • R ≈ 18: Synapse (and Dendrite, Conduit) generate an 18-char random opaque room identifier via random_string(18). No homeserver-level setting exposes this.
    • 



That leaves the homeserver hostname as the only variable. Honcho Session.id has Pydantic max_length=100, so solving 26 + 18 + HS + 43 ≤ 100 gives:


HS ≤ 13 chars



13 chars is a tight budget — short single-label domains (e.g. m.co) still fit, but many common patterns (IP-encoded DNS like sslip / nip, Tailscale .ts.net, subdomains of public registrars) exceed it, so a sizeable share of real deployments hits this.


Proposed Fix (optional)


No response


Are you willing to submit a PR for this?


    

  •  I'd like to fix this myself and submit a PR
    • 

Metadata
Metadata
Assignees
No one assigned
    Labels
    P2Medium — degraded but workaround existscomp/gatewayGateway runner, session dispatch, deliverycomp/pluginsPlugin system and bundled pluginsplatform/matrixMatrix adapter (E2EE)type/bugSomething isn't working
    Type
    No type
    Fields
    No fields configured for issues without a type.
    Projects
    No projects
    Milestone
    No milestone
    Relationships
    None yet
    Development
    No branches or pull requests
    Issue actions