Nico
07457475b6
<!-- Add the related story/sub-task/bug number, like Resolves #123, or remove if NA --> **Related issue:** Resolves #43544 This blueprint has drifted from the deployed `fleet-mcp` service in two ways. `FLEET_LIVE_QUERY_REST_PERIOD=25s` is too short for a hosted service aimed at a whole fleet: enough targeted hosts are asleep or offline that `run_live_query` returns a near-empty result before they check in, reading as "no data" rather than "still waiting." And `plan: starter` disagrees with the Standard instance the service actually runs on, so a blueprint sync risked silently downgrading it. Both fields carry literal values, making them blueprint-managed, so a dashboard override would be reverted on the next sync. <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **Configuration** * Updated the hosted Fleet deployment to use the standard plan. * Increased the live-query interval to 15 minutes. * Clarified configuration guidance for Fleet server alignment and hosted Fleet behavior. <!-- end of auto-generated comment: release notes by coderabbit.ai -->
26 lines
826 B
YAML
26 lines
826 B
YAML
services:
|
|
- type: web
|
|
name: fleet-mcp
|
|
runtime: go
|
|
plan: standard
|
|
rootDir: cmd/fleet-mcp
|
|
buildCommand: go build -o fleet-mcp .
|
|
startCommand: ./fleet-mcp
|
|
healthCheckPath: /healthz
|
|
envVars:
|
|
- key: FLEET_BASE_URL
|
|
sync: false
|
|
- key: FLEET_API_KEY
|
|
sync: false
|
|
- key: MCP_AUTH_TOKEN
|
|
sync: false
|
|
- key: LOG_LEVEL
|
|
value: info
|
|
# Upper bound run_live_query waits for hosts to report (Go duration).
|
|
# Keep >= the Fleet server's FLEET_LIVE_QUERY_REST_PERIOD. The code default
|
|
# when unset stays 25s; a hosted deployment targets whole fleets, where
|
|
# enough hosts are asleep or offline that 25s returns a misleading
|
|
# near-empty result before they check in.
|
|
- key: FLEET_LIVE_QUERY_REST_PERIOD
|
|
value: 900s
|