netlens.

Why raw lookup tools are not enough for SRE workflows

2026-04-18 · Yossi Ben Hagai · 2 min read sretoolingdiagnostics

Open any “DNS lookup” site. You paste a domain. You get a table of A records. You get another table of MX records. You get an ad. You close the tab.

This is useless for on-call work. It solves the display problem and ignores the interpretation problem.

What real SRE diagnostic work looks like

A production incident is never “what is the A record of example.com?” It’s always a correlated question:

None of these questions are answerable by a tool that prints a list of records. Answering them requires combining signals across DNS, HTTP, TLS, routing, and email posture - and then explaining what the combination means.

The correlation gap

Consider two real situations:

  1. server: cloudflare + cf-ray header present, but A record on AWS ASN.
  2. SPF passes, DKIM passes, but DMARC is at p=none.

Situation 1 is almost certainly an exposed origin. Situation 2 is almost certainly a spoofing risk. A raw lookup tool prints the inputs and stops. An engineer has to recognize the pattern. That cognitive step is where tools should be helping - not where they should be handing off.

What a useful tool looks like

netlens is an attempt to be that tool. It’s deliberately not another commodity lookup site. If you want one of those, there are fifty. If you want correlated diagnostics you can pipe into your agent, start here.