wrhn

Recent submissions

Comment · 8 months ago
I built a small MCP server that lets any AI agent (Claude Desktop, Cursor, Windsurf) block and wait for a human response before taking irreversible actions.

  Two tools:
  - ask_human(question) – agent blocks until you reply
  - request_approval(action) – agent blocks until you click Approve or Deny

  Notifications via Slack (with interactive buttons), or native macOS
  dialogs. Works with any MCP-compatible client over stdio or SSE.

  The use case that motivated it: I kept coming back to find my agent had done
  something I didn't want while I was away. Now it just asks.

  GitHub: https://github.com/nishantmodak/call-a-human-mcp
Comment · 9 months ago
Problem has never been the storage. Its running those queries to return in milliseconds - if its for a dashboard, an alert of your new AI agent trying to make sense of it.
Comment · 10 months ago
I wrote this after seeing the same failure modes across multiple teams and metrics stacks.

The intent here isn’t to argue that high cardinality is bad but to explain the mechanical reasons unbounded cardinality shows up later as cost surprises, slow queries, and trust erosion.

Curious if others have seen similar query-time or indexing issues they have noticed.

Comment · 1 year ago
Wasn't it pascal who said that all of our problems come from our inability to be bored.
Comment · 2 years ago
I have a different take

>Engineers have to pre-define and send all telemetry data they might need – since it’s so difficult to make changes after the fact – regardless of the percentage chance of the actual need.

YES. Let them send all the data. The best place to solve for it is at Ingestion.

There's typically 5 different stages to this process.

Instrumentation -> Ingestion -> Storage -> Query (Dashboard) -> Query (Alerting)

Instrumentation is the wrong place to solve this.

Ingestion - Build pipelines that allow to process this data and provide for tools like streaming aggregation, cardinality controls that allow to 'process it' or act on anomalous patterns. This atleast makes working on observability data 'dynamic' instead of having to go change instrumentation always. Storage - Provide blaze (2hours), hot(1 month), cold(13 months) of tiered data storage with indipendent read paths.

This, in my opinion has solved for the bulk of cost & re-work challenges associated with telemetry data.

I believe, Observability is the Big Data of today, without the Big Data tools! (Disclosure: I work at Last9.io and we have taken a similar approach to solve for these challenges)

https://last9.io/data-tiering/

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton