DEV Community

aoh
aoh

Posted on

Built an AI Log Analysis Tool — And Its "Analysis History" Feature Is Something I Use Every Day

🔍 I Built an AI Log Analysis Tool — And Its "Analysis History" Feature Is Something I Use Every Day

Author: Pangulab Team
Tool: https://www.pangulab.com

Intro

Last month I shared an AI-powered log analysis tool called Pangu RCA Agent, and got a lot of great feedback from developers. The most common question I received was:

"Can I save the analysis results? Can I view them later?"

Today I want to talk about the Analysis History feature in this tool — honestly, I use it almost every day in my own development work, and it's incredibly handy.

Why Do You Need "Analysis History"?

Let me start with my own use cases.

As a developer, I frequently debug various log issues:

  • Post-incident review: Woken up by an alert at 3 AM, analyzed the logs, found the root cause — but the next morning I can't remember what I found
  • Comparing similar issues: Last week I had a database connection timeout, this week I have another one — I want to compare if they share the same root cause
  • Sharing with teammates: I got the analysis result and want to send it to a colleague

Without history, every time you close the page, the result is gone. You'd have to paste the logs again and re-run the analysis. What a pain.

So I built this feature into the tool, and I made sure it's done right.

Feature Highlights

1. Auto-Save, No Login Required

Every time you complete an analysis, the result is automatically saved in your browser's local storage. No registration, no login — just open the website, use it, and leave.

Each analysis result page shows:

  • Problem type (e.g., "Database Connection Timeout")
  • Confidence score (how confident the AI is about this result)
  • Detailed root cause analysis
  • Key evidence (critical lines extracted from the logs)
  • Fix suggestions

2. Search & Filter

When you have a lot of history records, finding the right one is still easy:

  • Keyword search: Search by problem type or log content — results in seconds
  • Quick filters: Filter by "All", "Favorites", "Today", or "This Week"

3. Favorites & Delete

  • ⭐ Favorite: Star important analysis results for quick access
  • 🗑️ Delete: Remove irrelevant records to keep your list clean

4. Share Analysis Results

This is the feature I find most useful.

Every analysis result has its own unique detail page URL. You can:

  • Share with colleagues: Send the link to your team chat — they can see the full analysis instantly
  • Attach to tickets: Paste the link into Jira, Linear, or any issue tracker for traceability
  • Bookmark for yourself: Save it in your browser for later reference

Real-World Scenarios

Scenario 1: Woken Up by a 2 AM Alert 

2:00 AM — On-call alert: Database connection pool exhausted
Enter fullscreen mode Exit fullscreen mode
  1. Copy the error logs into the tool, click "Analyze"
  2. 5 seconds later: "Database connection pool is too small. Recommended: increase from 10 to 50."
  3. Apply the fix — problem resolved
  4. Star the analysis result for later reference

Next morning: Open the starred record, review the issue, and write the incident postmortem.

Scenario 2: Debugging Multiple Issues in One Service 

Service A reported 3 different errors today
Enter fullscreen mode Exit fullscreen mode
  1. Analyze each log separately — 3 history records created
  2. Search for "Service A" in the history page — all 3 records show up
  3. Compare the results and discover that 2 of the 3 issues share the same root cause

Scenario 3: Team Collaboration 

Colleague: "Have you seen this error before?"
You: "Yes! I analyzed it last week. Here's the link."
Enter fullscreen mode Exit fullscreen mode

Just send the analysis detail page link — your colleague can see the full result without re-pasting logs.

Technical Implementation (Briefly)

This feature is built with localStorage — all data is stored in the browser locally:

  • ✅ No backend storage needed — zero cost
  • ✅ Privacy-first — log data never leaves your browser
  • ✅ Works offline — view results anytime
  • ❌ Clearing browser cache will lose data (cloud sync is on the roadmap)

Each analysis record contains:

  • Original log content
  • AI analysis result (problem type, confidence, detailed cause, key evidence, fix suggestions)
  • Analysis timestamp
  • Favorite status

Final Thoughts

This tool is completely free to use with no usage limits.

If you regularly debug log issues, give it a try:

👉 https://www.pangulab.com

Questions or suggestions? Feel free to reach out: r00txxyyzz@gmail.com

If you found this useful, feel free to share it with your fellow developers 🙌

Top comments (0)