Skip to main content
A log event isn’t a log line. It’s a pattern that represents thousands or millions of logs that mean the same thing. Tero discovers these patterns from your integrations and turns them into something you can reason about.

​
Example

Here’s a raw log: 
body: "Payment declined: insufficient funds"
severity: ERROR
attributes:
  order_id: "ORD-12345"
  error_code: "card_declined"
  decline_reason: "insufficient_funds"
Tero turns it into a log event: 
name: card_declined_insufficient_funds
service: checkout
description: A card payment was declined due to insufficient funds.

matchers:
  - field_type: log_attribute
    field_name: error_code
    match_type: exact
    match_value: "card_declined"
  - field_type: log_attribute
    field_name: decline_reason
    match_type: exact
    match_value: "insufficient_funds"
These matchers identify 47,000 logs. Each log has different order IDs, timestamps, amounts. But they all match the same attributes: error_code: card_declined and decline_reason: insufficient_funds. That’s the event.

​
Exploring log events

Open any log event to see what Tero learned: the description, matchers, sample logs, classification reasoning. You can browse by service or search across all events. Drill into the matchers to understand exactly which logs this event captures. See sample instances to verify the pattern makes sense.

​
Using in chat

Reference a log event with @ to focus your questions: 
@payment_declined_insufficient_funds how often does this happen?
@database_connection_timeout what services are affected?
@order_created what's the typical latency?
Tero pulls in the event’s context: which service owns it, what attributes it has, how it relates to other events. Your question gets answered with that full picture.

​
Improving context

Tero classifies events automatically, but you know your systems better:
  • Reclassify events that Tero misjudged. That “debug” log might actually matter.
  • Edit descriptions to add context Tero couldn’t infer.
  • Adjust matchers if an event is capturing the wrong logs.
Every correction improves future classifications.