v0.54.3
caddyserver/caddyv0.54.3Sep 9, 2026by github-actions[bot]
Upgrading?
AI Summary
This maintenance release introduces configuration flexibility for databases and Redis, enhances AI provider integrations like LiteLLM and Gemini, and fixes various workflow and alert bugs.
Key Highlights
- Exposed SQLAlchemy pool settings and Redis DB selection via environment variables.
- LiteLLM integration now returns token usage and cost with the response.
- Improved provider stability with better model defaults for Grok and OpenAI.
- Fixed workflow deduplication and alert URL encoding issues.
New Features
- SQLAlchemy pool_recycle and pool_timeout env vars
- REDIS_DB environment variable
- Litellm token usage and cost reporting
- Grafana datasource query API
- SFTP monitor type with password and private key authentication
Full Release Notes
## What's Changed * feat(db): expose SQLAlchemy pool_recycle and pool_timeout as env vars by @ajuijas in https://github.com/keephq/keep/pull/6637 * fix(db): refresh user after role update to avoid DetachedInstanceError by @mvanhorn in https://github.com/keephq/keep/pull/6638 * ci: pin ruff version to make lint deterministic (#6648) by @shahargl in https://github.com/keephq/keep/pull/6651 * feat: expose REDIS_DB environment variable for Redis database selection (#6643) by @aries1980 in https://github.com/keephq/keep/pull/6644 * fix: avoid double-encoding alert URLs by @aditya-786 in https://github.com/keephq/keep/pull/6653 * ci: read the PR description from env instead of interpolating it into the shell by @kobihikri in https://github.com/keephq/keep/pull/6654 * fix: widen alert table column resize hit target by @ausias-armesto in https://github.com/keephq/keep/pull/6661 * fix(victorialogs): return empty headers for NoAuth instead of None by @artemrootman in https://github.com/keephq/keep/pull/6668 * feat(grafana): query any datasource via /api/ds/query by @artemrootman in https://github.com/keephq/keep/pull/6673 * fix(litellm): fail loudly on null content and accept fenced JSON by @artemrootman in https://github.com/keephq/keep/pull/6671 * feat(litellm): return token usage and cost with the response by @artemrootman in https://github.com/keephq/keep/pull/6678 * fix: correctly parse PUSHER_USE_SSL environment variable (#6582) by @ChickenisLegit in https://github.com/keephq/keep/pull/6680 * fix(gemini): default to a model Google still serves by @chiruu12 in https://github.com/keephq/keep/pull/6697 * fix(presets): don't rewrite search text when translating CEL contains by @dngr2 in https://github.com/keephq/keep/pull/6698 * fix(providers): don't crash on an empty choices list by @chiruu12 in https://github.com/keephq/keep/pull/6705 * fix(providers): default openai to a model that supports json_schema by @chiruu12 in https://github.com/keephq/keep/pull/6713 * fix(providers): default grok to a served model by @chiruu12 in https://github.com/keephq/keep/pull/6711 * fix(providers): don't crash on empty choices in grok by @chiruu12 in https://github.com/keephq/keep/pull/6710 * fix: prevent duplicate workflow runs on rapid manual run clicks by @kate1128 in https://github.com/keephq/keep/pull/6691 * fix(keep-provider): don't mutate the shared fingerprint_fields list by @Prabal864 in https://github.com/keephq/keep/pull/6720 * fix(servicenow): forward sysparm_offset in pagination and fix undefined error in _notify_update by @Veydop in https://github.com/keephq/keep/pull/6725 * fix(topologies): skip deleted incidents when looking up application incidents by @Georgefifth in https://github.com/keephq/keep/pull/6733 * fix(tests): wait for the workflow detail header before asserting on it by @chiruu12 in https://github.com/keephq/keep/pull/6730 * fix: Enable Zabbix ID deduplication by @mvanhorn in https://github.com/keephq/keep/pull/6745 * fix: skip alias rendering when a step's if condition is false by @ausias-armesto in https://github.com/keephq/keep/pull/6744 * fix(functions): read an offsetless timestamp as UTC by @chiruu12 in https://github.com/keephq/keep/pull/6777 * fix: refactored apply custom deduplication rules so that deduplicatio… by @Veydop in https://github.com/keephq/keep/pull/6794 * fix: ingest array payloads posted to provider endpoints by @Veydop in https://github.com/keephq/keep/pull/6798 * chore: Bump version from 0.54.2 to 0.54.3 by @shahargl in https://github.com/keephq/keep/pull/6800 ## New Contributors * @ajuijas made their first contribution in https://github.com/keephq/keep/pull/6637 * @mvanhorn made their first contribution in https://github.com/keephq/keep/pull/6638 * @aries1980 made their first contribution in https://github.com/keephq/keep/pull/6644 * @kobihikri made their first contribution in https://github.com/keephq/keep/pull/6654 * @artemrootman made their first contribution in https://github.com/keephq/keep/pull/6668 * @ChickenisLegit made their first contribution in https://github.com/keephq/keep/pull/6680 * @chiruu12 made their first contribution in https://github.com/keephq/keep/pull/6697 * @dngr2 made their first contribution in https://github.com/keephq/keep/pull/6698 * @kate1128 made their first contribution in https://github.com/keephq/keep/pull/6691 * @Prabal864 made their first contribution in https://github.com/keephq/keep/pull/6720 * @Veydop made their first contribution in https://github.com/keephq/keep/pull/6725 * @Georgefifth made their first contribution in https://github.com/keephq/keep/pull/6733 **Full Changelog**: https://github.com/keephq/keep/compare/v0.54.2...v0.54.3
Related Repositories
ollama/ollama★ 184,240
Get up and running with Kimi, GLM, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.
fatedier/frp★ 110,065
A fast reverse proxy to help you expose a local server behind a NAT or firewall to the internet.
infiniflow/ragflow★ 92,634
RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs
gohugoio/hugo★ 90,197
The world’s fastest framework for building websites.
junegunn/fzf★ 83,714
:cherry_blossom: A command-line fuzzy finder