The API monitor that doesn't just wake you up with an alert — it diagnoses the 500 error from recent Git diffs and opens a GitHub Pull Request to fix it.
flowchart TD
Monitor["⚡ PulseFix Async Prober"] -->|Probes Every 15-30s| TargetAPI["🌐 Production API Endpoint"]
TargetAPI -->|HTTP 500 Internal Error / Exception| IncidentCatcher["🚨 Incident Ingestion & Stack Trace Catcher"]
subgraph Autonomous_SRE_Engine["🧠 Autonomous AI SRE Engine"]
IncidentCatcher --> TraceParser["📑 Traceback & Payload Parser"]
GitAnalyzer["🐙 GitHub Commit Diff Fetcher"] --> ContextEngine
TraceParser --> ContextEngine["🧩 Error-to-Diff Context Engine"]
ContextEngine --> LLM["🤖 Gemini 2.5 Flash SRE Analyzer"]
LLM --> PatchGenerator["🛠️ Unified Diff Patch Generator"]
end
subgraph Auto_Resolution_Hub["🚀 Auto-Resolution & Alert Dispatch"]
PatchGenerator --> GitBranch["🌿 Create Git Branch (pulsefix/incident-id)"]
GitBranch --> GitHubPR["🐙 Open GitHub Pull Request"]
GitHubPR --> TelegramAlert["📱 Rich Telegram Alert with Root Cause & PR Link"]
GitHubPR --> CyberpunkDashboard["💻 Real-Time Cyberpunk SRE Dashboard"]
end
httpx probes measuring TTFB and SSL validity.git clone https://github.com/salomh46-rgb/pulseapi-monitoring-saas.git
cd pulseapi-monitoring-saas
pip install -r server/requirements.txt
TELEGRAM_BOT_TOKEN="your_telegram_bot_token"
GEMINI_API_KEY="your_gemini_api_key" # Optional: uses intelligent heuristic fallback if omitted
GITHUB_TOKEN="your_github_token" # Optional: simulates PR if omitted
python -m uvicorn server.main:app --reload --port 8000
Open client/index.html in your browser. Click "Simulate 500 & Auto-PR" to watch the self-healing workflow execute live!
python -m pytest -v
All 7 core tests cover:
/api/incidents/{id}/self-heal executionTitle: Show HN: PulseFix – An API monitor that diagnoses 500s from Git diffs and opens a fix PR
Body: Hey Hacker News, I'm Javohirbek.
As developers, we've all been woken up at 3 AM by PagerDuty or Sentry: "500 Internal Server Error in /api/checkout". Then begins the tedious process of digging through logs, opening GitHub, finding which commit changed the schema, and writing a patch.
I built PulseFix to turn passive monitoring into an active SRE engineer:
- It probes your endpoints every 15-30s.
- When a 500 occurs, it intercepts the error traceback.
- It correlates the crash against your recent Git commits using Gemini Flash.
- It synthesizes a safe unified diff and opens a GitHub Pull Request automatically.
- It pings your Telegram with: "Here is why it broke, and here is the PR to fix it."
The project is open-source, written in FastAPI + SQLite, and includes a full live demo in the dashboard.
GitHub: https://github.com/salomh46-rgb/pulseapi-monitoring-saas
Would love your feedback and brutal critiques on the architecture!
MIT License © Javohirbek Asqarov (Jasper)
7 commits
HTML
70.4%
Python
29.4%
The API monitor that doesn't just wake you up with an alert — it diagnoses the 500 error from recent Git diffs and opens a GitHub Pull Request to fix it.
flowchart TD
Monitor["⚡ PulseFix Async Prober"] -->|Probes Every 15-30s| TargetAPI["🌐 Production API Endpoint"]
TargetAPI -->|HTTP 500 Internal Error / Exception| IncidentCatcher["🚨 Incident Ingestion & Stack Trace Catcher"]
subgraph Autonomous_SRE_Engine["🧠 Autonomous AI SRE Engine"]
IncidentCatcher --> TraceParser["📑 Traceback & Payload Parser"]
GitAnalyzer["🐙 GitHub Commit Diff Fetcher"] --> ContextEngine
TraceParser --> ContextEngine["🧩 Error-to-Diff Context Engine"]
ContextEngine --> LLM["🤖 Gemini 2.5 Flash SRE Analyzer"]
LLM --> PatchGenerator["🛠️ Unified Diff Patch Generator"]
end
subgraph Auto_Resolution_Hub["🚀 Auto-Resolution & Alert Dispatch"]
PatchGenerator --> GitBranch["🌿 Create Git Branch (pulsefix/incident-id)"]
GitBranch --> GitHubPR["🐙 Open GitHub Pull Request"]
GitHubPR --> TelegramAlert["📱 Rich Telegram Alert with Root Cause & PR Link"]
GitHubPR --> CyberpunkDashboard["💻 Real-Time Cyberpunk SRE Dashboard"]
end
httpx probes measuring TTFB and SSL validity.git clone https://github.com/salomh46-rgb/pulseapi-monitoring-saas.git
cd pulseapi-monitoring-saas
pip install -r server/requirements.txt
TELEGRAM_BOT_TOKEN="your_telegram_bot_token"
GEMINI_API_KEY="your_gemini_api_key" # Optional: uses intelligent heuristic fallback if omitted
GITHUB_TOKEN="your_github_token" # Optional: simulates PR if omitted
python -m uvicorn server.main:app --reload --port 8000
Open client/index.html in your browser. Click "Simulate 500 & Auto-PR" to watch the self-healing workflow execute live!
python -m pytest -v
All 7 core tests cover:
/api/incidents/{id}/self-heal executionTitle: Show HN: PulseFix – An API monitor that diagnoses 500s from Git diffs and opens a fix PR
Body: Hey Hacker News, I'm Javohirbek.
As developers, we've all been woken up at 3 AM by PagerDuty or Sentry: "500 Internal Server Error in /api/checkout". Then begins the tedious process of digging through logs, opening GitHub, finding which commit changed the schema, and writing a patch.
I built PulseFix to turn passive monitoring into an active SRE engineer:
- It probes your endpoints every 15-30s.
- When a 500 occurs, it intercepts the error traceback.
- It correlates the crash against your recent Git commits using Gemini Flash.
- It synthesizes a safe unified diff and opens a GitHub Pull Request automatically.
- It pings your Telegram with: "Here is why it broke, and here is the PR to fix it."
The project is open-source, written in FastAPI + SQLite, and includes a full live demo in the dashboard.
GitHub: https://github.com/salomh46-rgb/pulseapi-monitoring-saas
Would love your feedback and brutal critiques on the architecture!
MIT License © Javohirbek Asqarov (Jasper)
7 commits
HTML
70.4%
Python
29.4%