monitor-log
The state of a company's services, in words anyone understands.
What it is
monitor-log reads the logs an organisation is already producing and translates them: what happened, who it affects and what to do. It collects nothing and keeps not a single log line. It connects read-only to the store the company already uses, such as New Relic, CloudWatch, Loki or Elastic, and shows anyone on the team, whether or not they know infrastructure, whether a service is working.
What the log says
2026-08-20T14:02:11Z ERROR orders-api
[a3f9c2] Queue full (max 50), rejecting
publish to orders.confirmWhat the team reads
Orders are queueing up
Whoever is buying is left waiting for a confirmation that never comes.
What to do Check the consumer of the confirmation queue: it reached its limit and is rejecting.
What it does
- Translates each technical message into its consequence and what to do, with a dictionary written by whoever knows the business and edited on screen.
- Prioritises: out of thousands of lines a day it says which need action today, and why each one is on that list.
- Watches the silence. A process that stops running writes no errors, so each flow learns its pace and alerts when it stops.
- Follows a case end to end: with an identifier, such as an order or an invoice, it shows its path through every service, in order.
- Says what it does not know: which lines it could not explain and which services gave no signal.
How it works
Security and isolation
Each company has its own read credentials and its own rules. The company travels signed inside the session, so there is no parameter to change to see another company's data, and even the server's cache is separated by company. Credentials are declared by variable name and the panel does not store them. The team gets in with their work email or a one-time link.
On the blog
-
October 16, 2026
Observability in business language
Logs tell what is happening in production, but in a language few people read.