2 pointsby naiks12148 hours ago1 comment
  • naiks12148 hours ago
    Debugging incidents from logs is usually painful, especially when it spans multiple files.

    I made a simple CLI tool that tries to figure things out based on things we already do — like looking for ERROR logs, time windows, and some context.

    You give logs (even from multiple services), and it:

    * builds a timeline * highlights important parts * gives a rough root cause

    Example:

    `sydes incident analyze api.log worker.log gateway.log --query "Why are session lookups failing?"`

    Works best when logs are from multiple services (api + worker etc), where context is spread out.

    Still early. Would like to know:

    * Does this make sense for you? * Would you trust something like this? * Any improvements you would like to suggest?