Current main: applies to the application at
51a835a. See version and availability for newer candidate work.
Command-line reference¶
Server¶
| Flag | Default | Purpose |
|---|---|---|
-yaml |
monitors.yaml |
YAML or JSON monitor manifest. |
-config |
empty | Alias overriding -yaml. |
-allow-empty |
false | Permit a deliberately empty manifest. |
-web |
true | Enable dashboard, API, and metrics. |
-web.addr |
localhost:8060 |
Listener address; a remote bind requires authentication. |
-web.auth-file |
CPRA_AUTH_TOKEN_FILE |
Read the authentication token from a file. |
-web.auth |
CPRA_AUTH_TOKEN |
Set a token directly; prefer a file to command-line exposure. |
-web.cors |
empty | Comma-separated allowed origins; intended for development. |
-debug |
false | Enable debug logging. |
-pprof |
false | Enable profiling; use a loopback address. |
-pprof.addr |
localhost:6060 |
Profiling listener address. |
-ssrf-protect |
false | Block non-public destinations for HTTP(S) paths at connection time. |
-version |
false | Print build information and exit. |
A token file overrides a direct or environment token. An unreadable or empty token file fails startup. CPRA_ENTITY_THRESHOLD optionally influences startup queue selection; there is no live queue migration.
Client commands¶
./bin/cpractl health
./bin/cpractl get overview
./bin/cpractl get monitors --status down
./bin/cpractl get monitors --query example --page 1 --size 50
./bin/cpractl get monitor 2 -o json
./bin/cpractl get incidents
./bin/cpractl get systems
./bin/cpractl get queues
./bin/cpractl get pools
./bin/cpractl get config
./bin/cpractl metrics
Replace example ID 2 with an ID returned by your instance.
| Global option | Default |
|---|---|
--server |
CPRA_SERVER or http://localhost:8060 |
--token-file |
CPRA_AUTH_TOKEN_FILE |
--request-timeout |
10s |
--output / -o |
table; also accepts wide, json, and yaml |
The client also reads CPRA_AUTH_TOKEN. health checks liveness; it is not a readiness or fleet-health command. Failed API requests produce an error and an unsuccessful exit status. An HTTP 503 listing response must not be interpreted as zero incidents.