{"repo":"Tracer-Cloud/opensre","free":true,"listed":false,"github":"https://github.com/Tracer-Cloud/opensre","clone":"git clone https://github.com/Tracer-Cloud/opensre.git","description":"Build your own AI SRE agents. The open source toolkit for the AI era.","language":"Python","stars":10579,"topics":["ai-sre","alerting","datadog","grafana","incident-management","observability","remediation","root-cause-analysis","site-reliability-engineering","slack","sre"],"license":"Apache-2.0","category":"sre_platform","readme_excerpt":"
\n\n

\n \"OpenSRE\"\n

\n\n

OpenSRE v0.1: Build Your Own AI SRE Agents

\n\n

The open-source framework for AI SRE agents, and the training and evaluation environment they need to improve. Connect the 60+ tools you already run, define your own workflows, and investigate incidents on your own infrastructure.

\n\n

\n \"CI\n\"Project\n \"Apache\n \"Discord\"\n \"Sponsored\n

\n\n

\n \n \n \n

\n\n

\n \n Quickstart ยท\n Docs ยท\n FAQ ยท\n Security\n \n

\n\n
\n\n---\n\n> ๐Ÿšง Public Alpha: Core workflows are usable for early exploration, though not yet fully stable. The project is in active development, and APIs and integrations may evolve\n\n---\n\n## Table of Contents\n\n- [Why OpenSRE?](#why-opensre)\n- [Install](#install)\n- [Quick Start](#quick-start)\n- [Deployment](#deployment)\n- [How OpenSRE Works](#how-opensre-works)\n- [Benchmark](#benchmark)\n- [Capabilities & integrations](#capabilities--integrations)\n- [Contributing & development](#contributing--development)\n- [Security](#security)\n- [Telemetry](#telemetry)\n- [License](#license)\n- [Citations](#citations)\n\n---\n\n## Why OpenSRE?\n\nWhen something breaks in production, the evidence is scattered across logs, metrics, traces, runbooks, and Slack threads. OpenSRE is an open-source framework for AI SRE agents that resolve production incidents, built to run on your own infrastructure.\n\nWe do that because SWE-bench1 gave coding agents scalable training data and clear feedback. Production incident response still lacks an equivalent.\n\nDistributed failures are slower, noisier, and harder to simulate and evaluate than local code tasks, which is why AI SRE, and AI for production debugging more broadly, remains unsolved.\n\nOpenSRE is building _that_ missing layer:\n\n> an open reinforcement learning environment for agentic infrastructure incident response, with end-to-end tests and synthetic incident simulations for realistic production failures\n\nWe do that by:\n\n- building easy-to-deploy, customizable AI SRE agents for production incident investigation and response\n- running scored synthetic RCA suites that check root-cause accuracy, required evidence, and adversarial red herrings [(tests/synthetic)](tests/synthetic/rds_postgres)\n- running real-world end-to-end tests across cloud-backed scenarios including Kubernetes, EC2, CloudWatch, Lambda, ECS Fargate, and Flink [(tests/e2e)](tests/e2e)\n- keeping semantic test-catalog naming so e2e vs synthetic and local vs cloud boundaries stay obvious [(tests/README.md)](tests/README.md)\n\nOur mission is to build AI SRE agents on top of this, scale it to thousands of realistic infrastructure failure scenarios, and establish OpenSRE as the benchmark and training ground for AI SRE.\n\n1 https://arxiv.org/abs/2310.06770\n\n---\n\n## Install\n\nThe root installer URL auto-detects Unix shell vs PowerShell and installs the latest build from `main`. OpenSRE moves quickly, so `main` is the latest stable version for normal installs.\n\nmacOS / Linux:\n\n```bash\ncurl -fsSL https://install.opensre.com | bash\n```\n\nThe macOS/Linux installer does not require sudo. If no writable bin directory is already on `PATH`, it installs to `~/.local/bin` and prints the shell command to apply the PATH update.\n\nEquivalent explicit main-channel form:\n\n```bash\ncurl -fsSL https://install.opensre.com | bash -s -- --main\n```\n\nHomebrew:\n\n```bash\nbrew tap tracer-cloud/tap\nbrew install tracer-cloud/tap/opensre\n```\n\nWindows (PowerShell):\n\n```powershell\nirm https://install.opensre.com | iex\n```\n\n\n\n---\n\n## Quick Start\n\nContributors: start at [`main.py`](main.py) for the process entrypoint map.\n\nConfigure once, then pick how you want to run investigations:\n\n```bash\nopensre onboard\n```\n\n**Interactive shell** โ€” with no subcommand, `opensre` starts a REPL (TTY required). Describe incidents in plain language, stream investigations, and use slash commands for session control (`/help`, `/status`, `/cost`, `/sessions`, `/resume`, `/compact`, `/new`, `/exit`), integrations (`/integrations list`, `/integrations verify`), local agent fleet monitoring (`/agents`), and reasoning depth (`/effort` for **OpenAI** and **Codex** โ€” `low` through `max`). Ctrl+C cancels an in-flight investigation without losing session state. See **[interactive shell commands](https://www.opensre.com/docs/interactive-shell-commands)** for the full reference.\n\n```bash\nopensre\n```\n\n**One-shot investigation** โ€” run the agent once against an alert file:\n\n```bash\nopensre investigate -i tests/e2e/kubernetes/fixtures/datadog_k8s_alert.json\n```\n\n**Remote runtime investigation** โ€” investigate a deployed service by name (live health, logs, and deployment status):\n\n```bash\nopensre investigate --service api-backend\n```\n\n**Hermes log watch** โ€” tail a Hermes `errors.log`, classify incidents, and optionally alert on Telegram:\n\n```bash\nopensre hermes watch\n```\n\n**From Python** โ€” drive the agent in-process from your own code (source checkout required):\n\n```python\nfrom core.agent_harness import AgentSession\n\nsession = AgentSession.start()\nresult = session.chat(\"why is checkout-api slow?\")\nif result.answered:\n print(result.primary_response_text)\n```\n\nSee **[Python API](https://www.opensre.com/docs/python-api)** for sessions, conversations, and custom output sinks.\n\n**For your team's daily loop:** embed OpenSRE in the Python services and automations your teammates already use.\nStart with the in-repo [Python API guide](docs/python-api.mdx), then use it every day to make incident response repeatable.\n\nOther useful commands:\n\n```bash\nopensre integrations setup\nopensre agents scan\nopensre update\nopensre uninstall # remove opensre and all local data\n```\n\n---\n\n## Deployment\n\nTwo primary AWS EC2 paths and a general hosted option:\n\n- **Gateway (AMI + systemd):** `make build-gateway-image` then `make deploy-gateway` โ€” Telegram gateway only, no Docker; the gateway is installed into a server image that new servers start from.\n- **Hosted (Railway / ECS / Vercel):** deploy with the repo `Dockerfile`; set `LLM_PROVIDER` and the matching API key (see [`.env.example`](.env.example)), plus `DATABASE_URI` and `REDIS_URI` if persistence is needed.\n\n**[Full deployment steps and prerequisites โ†’ DEPLOYMENT.md](DEPLOYMENT.md)**\n\n---\n\n## How OpenSRE Works\n\n\n\nWhen an alert fires, OpenSRE automatically:\n\n1. **Fetches** the alert context and correlated logs, metrics, traces, and recent deploys\n2. **Masks** sensitive identifiers (optional) before external LLM calls\n3. **Reasons** across your connected systems to test hypotheses in a tool-calling loop\n4. **Generates** a structured investigation report with probable root cause and linked evidence\n5. **Suggests** next steps and, optionally, executes remediation actions\n6. **Posts** a summary directly to Slack, PagerDuty, or Telegram โ€” no context switching needed\n\nFor the current code-level agent architecture after removing the old graph and chain\nframework layers, see [AGENTS.md](AGENTS.md).\n\n---\n\n## Benchmark\n\nRegenerate numbers with **`make benchmark`**; refresh this table from cached results via **`make benchmark-update-readme`**. See **[docs/DEVELOPMENT.md](docs/DEVELOPMENT.md#benchmark)** for details.\n\n\n\n_No benchmark results yet._\n\n\n\n---\n\n## Capabilities & integrations\n\n| | |\n| ---------------------------------------- | --------------------------------------------------------------------------","default_branch":"main","files":4944,"tree":[".claude/settings.json",".cursor/settings.json",".cursor/worktrees.json",".devcontainer/Dockerfile",".devcontainer/devcontainer.json",".dockerignore",".editorconfig",".env.deploy.example",".env.example",".gitattributes",".github/ISSUE_TEMPLATE/bug_report.yml",".github/ISSUE_TEMPLATE/config.yml",".github/ISSUE_TEMPLATE/feature_request.yml",".github/ISSUE_TEMPLATE/improvement.yml",".github/PULL_REQUEST_TEMPLATE.md",".github/assets/celebrations/celebrate.gif",".github/assets/celebrations/fireworks.gif",".github/assets/celebrations/merge-celebrate-1.gif",".github/assets/celebrations/merge-celebrate-2.gif",".github/assets/celebrations/merge-celebrate-3.gif",".github/assets/celebrations/office-celebrate.gif",".github/assets/celebrations/party.gif",".github/assets/celebrations/ship.gif",".github/assets/celebrations/shipped.gif",".github/assets/celebrations/woohoo.gif",".github/assets/opensre-discord-release-icon.webp",".github/ci/check_direct_imports.py",".github/ci/check_import_cycles.py",".github/ci/check_imports.py",".github/ci/run_live_turn_shards.py",".github/ci/run_test_scope.py",".github/ci/test_scope_rules.py",".github/codeql/codeql-config.yml",".github/dependabot.yml",".github/scripts/automerge_pr.py",".github/scripts/build-discord-payload.py",".github/scripts/good_first_issue_assign.py",".github/scripts/merged_pr_celebration_message.py",".github/scripts/sync-homebrew-tap-formula.sh",".github/workflows/README.md",".github/workflows/automerge.yml",".github/workflows/benchmark-image.yml",".github/workflows/benchmark-promote-image.yml",".github/workflows/benchmark-readme.yml",".github/workflows/benchmark-run.yml",".github/workflows/benchmark-seed-secret.yml",".github/workflows/celebrate-merged-pr.yml",".github/workflows/ci-labels-windows.yml",".github/workflows/ci.yml",".github/workflows/closed-loop-weekly.yml",".github/workflows/codeql.yml",".github/workflows/e2e-openclaw.yml",".github/workflows/good-first-issue-assign.yml",".github/workflows/greptile-pr-reminder.yml",".github/workflows/hermes-tests.yml",".github/workflows/interactive-shell-live.yml",".github/workflows/release.yml",".github/workflows/synthetic-deterministic.yml",".github/workflows/terraform-bench.yml",".github/workflows/tracer-demo-prefect-ecs.yml",".github/workflows/trigger-k8s-alert.yml",".gitignore",".importlinter",".importlinter.strict",".pre-commit-config.yaml",".tool-versions","AGENTS.md","CI.md","CLAUDE.md","CODE_OF_CONDUCT.md","CONTRIBUTING.md","DEPLOYMENT.md","Dockerfile","LICENSE","Makefile","README.md","SECURITY.md","SETUP.md","bootstrap/__init__.py","bootstrap/adapters.py","bootstrap/embedded.py","bootstrap/process.py","config/__init__.py","config/config.py","config/constants/__init__.py","config/constants/alertmanager.py","config/constants/aws.py","config/constants/azure.py","config/constants/azure_sql.py","config/constants/betterstack.py","config/constants/billing.py","config/constants/buzz.py","config/constants/coralogix.py","config/constants/dagster.py","config/constants/datadog.py","config/constants/discord.py","config/constants/filestorage.py","config/constants/gateway.py","config/constants/git.py","config/constants/github.py","config/constants/gitlab.py","config/constants/grafana.py","config/constants/groundcover.py","config/constants/helm.py","config/constants/honeycomb.py","config/constants/incident_io.py","config/constants/integration_secrets.py","config/constants/investigation.py","config/constants/jenkins.py","config/constants/kubernetes.py","config/constants/llm.py","config/constants/mariadb.py","config/constants/memory.py","config/constants/mongodb.py","config/constants/mongodb_atlas.py","config/constants/mysql.py","config/constants/new_relic.py","config/constants/openclaw.py","config/constants/opensearch.py","config/constants/operations_log.py","config/constants/organization.py","config/constants/pagerduty.py","config/constants/paths.py","config/constants/platform.py","config/constants/postgresql.py","config/constants/posthog.py","config/constants/posthog_mcp.py","config/constants/product.py","config/constants/prompts.py","config/constants/railway.py","config/constants/rds.py","config/constants/redis.py","config/constants/reporting.py","config/constants/rocketchat.py","config/constants/runtime_metadata.py","config/constants/secrets.py","config/constants/sentry.py","config/constants/sentry_mcp.py","config/constants/servicenow.py","config/constants/signoz.py","config/constants/slack.py","config/constants/slash_commands.py","config/constants/smtp.py","config/constants/telegram.py","config/constants/tempo.py","config/constants/temporal.py","config/constants/tenancy.py","config/constants/tracer.py","config/constants/twilio.py","config/constants/vercel.py","config/constants/work_items.py","config/constants/x_mcp.py","config/constants/yandex_cloud.py","config/env_file.py","config/env_key_sensitivity.py","config/grafana_cloud.py","config/llm_auth/__init__.py","config/llm_auth/auth_method.py","config/llm_auth/credentials.py","config/llm_auth/provider_catalog.py","config/llm_auth/records.py","config/llm_credentials.py","config/llm_models.py","config/llm_reasoning_effort.py","config/local_env.py","config/local_settings.py","config/platform_bootstrap.py","config/principal.py","config/remote_store.py","config/repl_config.py","config/runtime_metadata/__init__.py","config/runtime_metadata/assembly.py","config/runtime_metadata/build_info.py","config/runtime_metadata/contract.py","config/runtime_metadata/probes.py","config/scope_context.py","config/secrets/__init__.py","config/secrets/backend.py","config/secrets/guidance.py","config/secrets/keychain_import.py","config/secrets/local_file.py","config/secrets/os_keyring.py","config/secrets/records.py","config/secrets/store.py","config/strict_config.py","config/version.py","core/__init__.py","core/agent/__init__.py","core/agent/agent.py","core/agent/cancel.py","core/agent/goals.py","core/agent/handle_conclusion.py","core/agent/loop_host.py","core/agent/mixins.py","core/agent/provider_hooks.py","core/agent/react_loop.py","core/agent/run_io.py","core/agent/textual_tool_call.py","core/agent_harness/AGENTS.md","core/agent_harness/__init__.py","core/agent_harness/accounting/__init__.py","core/agent_harness/accounting/run_record.py","core/agent_harness/accounting/token_accounting.py","core/agent_harness/accounting/token_usage.py","core/agent_harness/accounting/turn_accounting.py","core/agent_harness/agent_builder.py","core/agent_harness/closed_llm_verdict.py","core/agent_harness/error_reporting.py","core/agent_harness/grounding/__init__.py","core/agent_harness/grounding/_cache.py","core/agent_harness/grounding/agents_md_reference.py","core/agent_harness/grounding/context.py","core/agent_harness/grounding/diagnostics.py","core/agent_harness/grounding/docs_reference.py","core/agent_harness/grounding/investigation_flow_reference.py","core/agent_harness/grounding/models.py","core/agent_harness/harness.py","core/agent_harness/investigation_api.py","core/agent_harness/llm_resolution.py","core/agent_harness/ports.py","core/agent_harness/prompts/AGENTS.md","core/agent_harness/prompts/__init__.py","core/agent_harness/prompts/action/__init__.py","core/agent_harness/prompts/action/assemble.py","core/agent_harness/prompts/action/multi_step_policy.py","core/agent_harness/prompts/action/text.py","core/agent_harness/prompts/assistant/__init__.py","core/agent_harness/prompts/assistant/assemble.py","core/agent_harness/prompts/assistant/blocks.py","core/agent_harness/prompts/assistant/environment.py","core/agent_harness/prompts/assistant/observation.py","core/agent_harness/prompts/assistant/parts.py","core/agent_harness/prompts/assistant/text.py","core/agent_harness/prompts/assistant/turn.py","core/agent_harness/prompts/gather/__init__.py","core/agent_harness/prompts/gather/assemble.py","core/agent_harness/prompts/grounding/__init__.py","core/agent_harness/prompts/grounding/provider.py","core/agent_harness/prompts/kernel/__init__.py","core/agent_harness/prompts/kernel/envelope.py","core/agent_harness/prompts/kernel/surfaces.py","core/agent_harness/prompts/memory/__init__.py","core/agent_harness/prompts/memory/conversation.py","core/agent_harness/prompts/memory/prior_investigation.py","core/agent_harness/prompts/rules.py","core/agent_harness/prompts/runtime_facts/__init__.py","core/agent_harness/prompts/runtime_facts/render.py","core/agent_harness/prompts/skills/__init__.py","core/agent_harness/prompts/skills/_template/SKILL_TEMPLATE.md","core/agent_harness/prompts/skills/architecture_audit/SKILL.md","core/agent_harness/prompts/skills/architecture_audit/architecture_audit_report.md","core/agent_harness/prompts/skills/github_ci_fix/SKILL.md","core/agent_harness/prompts/skills/github_ci_fix_onboarding/SKILL.md","core/agent_harness/prompts/skills/github_cli/SKILL.md","core/agent_harness/prompts/skills/github_security_fix/SKILL.md","core/agent_harness/prompts/skills/loader.py","core/agent_harness/prompts/skills/morning_report/SKILL.md","core/agent_harness/prompts/synthetic_failure.py","core/agent_harness/session/__init__.py","core/agent_harness/session/history_entry.py","core/agent_harness/session/integration_resolution.py","core/agent_harness/session/lifecycle.py","core/agent_harness/session/memory_extraction.py","core/agent_harness/session/pending_choice.py","core/agent_harness/session/pending_offer.py","core/agent_harness/session/persistence/__init__.py","core/agent_harness/session/persistence/jsonl_repo.py","core/agent_harness/session/persistence/jsonl_store.py","core/agent_harness/session/persistence/memory.py","core/agent_harness/session/persistence/paths.py","core/agent_harness/session/persistence/ports.py","core/agent_harness/session/persistence/wal_recovery.py","core/agent_harness/session/session_core.py","core/agent_harness/session/terminal_access.py","core/agent_harness/session/want_me_to.py","core/agent_harness/session_goal/AGENTS.md","core/agent_harness/session_goal/__init__.py","core/agent_harness/session_goal/confirm.py","core/agent_harness/session_goal/continuation.py","core/agent_harness/session_goal/evaluate.py","core/agent_harness/session_goal/goal.py","core/agent_harness/session_goal/persist.py","core/agent_harness/session_goal/progress.py","core/agent_harness/session_goal/run_until.py","core/agent_harness/tools/__init__.py","core/agent_harness/tools/action_tools.py","core/agent_harness/tools/tool_context.py","core/agent_harness/tools/tool_provider.py","core/agent_harness/turns/__init__.py","core/agent_harness/turns/action_driver.py","core/agent_harness/turns/answer_finalize.py","core/agent_harness/turns/assistant_handoff.py","core/agent_harness/turns/chat_api.py","core/agent_harness/turns/cohort_identity.py","core/agent_harness/turns/conversation_recording.py","core/agent_harness/turns/default_headless_agent.py","core/agent_harness/turns/default_reasoning_client.py","core/agent_harness/turns/evidence_driver.py","core/agent_harness/turns/evidence_kind.py","core/agent_harness/turns/evidence_need.py","core/agent_harness/turns/gather_discovery_budget.py","core/agent_harness/turns/gather_observation.py","core/agent_harness/turns/gather_ports.py","core/agent_harness/turns/gather_unreachable.py","core/agent_harness/turns/goal_review.py","core/agent_harness/turns/handoff_contents.py","core/agent_harness/turns/handoff_keys.py","core/agent_harness/turns/handoff_policy.py","core/agent_harness/turns/handoff_tag_parse.py","core/agent_harness/turns/headless_adapters.py","core/agent_harness/turns/headless_dispatch.py","core/agent_harness/turns/host_cancel.py","core/agent_harness/turns/metric_query_floor.py","core/agent_harness/turns/orchestrator.py","core/agent_harness/turns/runtime_fact_check.py","core/agent_harness/turns/source_circuit_breaker.py","core/agent_harness/turns/transcript_compaction.py","core/agent_harness/turns/turn_plan.py","core/agent_harness/turns/turn_results.py","core/agent_harness/turns/turn_route.py","core/agent_harness/turns/turn_snapshot.py","core/agent_harness/turns/wal_recorder.py","core/context_budget.py","core/domain/__init__.py","core/domain/alerts/__init__.py","core/domain/alerts/alert_source.py","core/domain/alerts/extraction.py","core/domain/alerts/fields.py","core/domain/alerts/inbox.py","core/domain/alerts/normalization.py","core/domain/alerts/tool_planning.py","core/domain/correlation/__init__.py","core/domain/correlation/confidence.py","core/domain/correlation/scoring.py","core/domain/diagnosis/__init__.py","core/domain/diagnosis/alignment.py","core/domain/diagnosis/result.py","core/domain/diagnosis/taxonomy_registry.py","core/domain/feedback/__init__.py","core/domain/feedback/misses/__init__.py","core/domain/feedback/misses/export.py","core/domain/feedback/misses/store.py","core/domain/feedback/misses/taxonomy.py","core/domain/memory/__init__.py","core/domain/memory/files.py","core/domain/memory/frontmatter.py","core/domain/memory/index.py","core/domain/memory/models.py","core/domain/memory/safety.py","core/domain/memory/settings.py","core/domain/memory/slugs.py","core/domain/memory/store.py","core/domain/pipeline_spans.py","core/domain/stream.py","core/domain/types/__init__.py","core/domain/types/config.py","core/domain/types/evidence.py","core/domain/types/incident_anchors.py","core/domain/types/incident_window.py","core/domain/types/planning.py","core/domain/types/retrieval.py","core/domain/types/root_cause_categories.py","core/domain/types/tools.py","core/domain/types/upstream.py","core/domain/work_items/__init__.py","core/domain/work_items/files.py","core/domain/work_items/models.py","core/domain/work_items/reminders.py","core/domain/work_items/schedule.py","core/domain/work_items/scoring.py","core/domain/work_items/store.py","core/events.py","core/execution.py","core/llm/AGENTS.md","core/llm/__init__.py","core/llm/client_builders.py","core/llm/factory.py","core/llm/failure_classification.py","core/llm/image_description.py","core/llm/internal/__init__.py","core/llm/internal/client_cache.py","core/llm/internal/client_cache_key.py","core/llm/internal/preload.py","core/llm/parsers/__init__.py","core/llm/parsers/root_cause.py","core/llm/providers/__init__.py","core/llm/providers/azure_openai.py","core/llm/providers/bedrock_model_ids.py","core/llm/providers/custom_endpoints.py","core/llm/providers/openai_compat_providers.py","core/llm/providers/provider_credentials.py","core/llm/providers/provider_registry.py","core/llm/providers/vertex_ai.py","core/llm/shared/__init__.py","core/llm/shared/llm_retry.py","core/llm/shared/openai_chat_completions.py","core/llm/shared/openai_responses.py","core/llm/shared/structured_output.py","core/llm/shared/tool_schema_normalize.py","core/llm/shared/usage.py"],"storefront":"/r/Tracer-Cloud","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/Tracer-Cloud/opensre/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}