Set a service's health check and deploy strategy
Sets the custom health check (null = TCP on the first port) and how deploys roll out: rolling (zero downtime, health gated) or recreate, and how long a new version may take to become healthy. Applies on the next deploy.
Needs services.update on the service (a grant on its project covers it). Available to MCP clients and the assistant.
seylo_session<token>application/json- body
serviceId*stringPanel-issued id.
^[a-z0-9]{16}$healthCheck*|nulldeploy?Success.
application/json- response
healthCheck*|nulldeploy*curl -X POST 'https://example.com/api/v1/services.setHealthCheck' \ -H 'Content-Type: application/json' \ -d '{ "serviceId": "string", "healthCheck": { "probe": { "type": "tcp", "port": 1 }, "intervalSeconds": 1, "timeoutSeconds": 1, "startPeriodSeconds": 0, "failureThreshold": 1 }}'{ "healthCheck": { "probe": { "type": "tcp", "port": 1 }, "intervalSeconds": 1, "timeoutSeconds": 1, "startPeriodSeconds": 0, "failureThreshold": 1 }, "deploy": { "mode": "rolling", "healthTimeoutSeconds": 10 }}Set a service's crash policy POST
Restart on or off after a crash, and with restart on, whether to stop after N crashes in a row (0 = never). The wait between restarts is a fixed 5 seconds. Applies to running containers right away. Needs `services.update` on the service (a grant on its project covers it). Available to MCP clients and the assistant.
Edit a service's notes POST
Replaces the free-text notes of a service (Markdown, shown on its Notes tab). Notes are not secret; keep secrets in the environment. Needs `services.update` on the service (a grant on its project covers it). Available to MCP clients and the assistant.