Claus Ibsen created CAMEL-26713:
-----------------------------------
Summary: camel-console - the timeline records the faults of the
chaos rules and the routes they had down
Key: CAMEL-26713
URL: https://issues.apache.org/jira/browse/CAMEL-26713
Project: Camel
Issue Type: Improvement
Components: camel-jbang
Reporter: Claus Ibsen
Fix For: 4.24.0
The timeline dev console (CAMEL-26691) records per second what the routes did,
the errors and the kept exchanges, inside the integration, so the CLI ({{camel
get timeline}}), MCP ({{camel_runtime_timeline}}) and the Timeline screen have
the last 5 minutes from the start.
The chaos part is not recorded there yet. The Timeline screen still takes it
from the chaos state in the status, as it watches the integration:
* the faults per second of each chaos rule (the 🎲 layers): the status has
{{faultsPerSecond}} for the last 60 seconds only;
* the seconds a stall or lifecycle rule had a route down (the shaded seconds):
the status only has the state now.
So a Timeline opened after an outage shows the faults of the last minute, but
not that the shipping route was down for 15 seconds, and {{camel get timeline}}
and the MCP tool have neither.
Proposal: when camel-chaos is in the integration, the timeline console records
per second the faults of each rule (rule id and a short label) and the routes a
rule had down, and returns them as {{chaos}} next to {{errors}} (only present
with chaos). camel-chaos can publish them through a small SPI or context plugin
that the console reads, so camel-console does not depend on camel-chaos. Then:
* the Timeline screen takes the layers and the down seconds from the console,
and no longer samples the chaos state;
* {{camel get timeline}} shows the rule next to the errors and a DOWN column
per route;
* the MCP tool has them, so an AI explaining an outage sees "route shipping
down from 12:03:41 to 12:03:56 by rule r2".
_Claude Code on behalf of Claus Ibsen_
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
