Blip
This page details how to configure the Blip binary (blip
) that runs monitors. To configure (customize) metrics collection, see Plans and Metrics.
Config File
Blip uses a single YAML file for configuration. See Configure / Config File for the full list of Blip config variables.
๐ By default, Blip reads blip.yaml
in the current working directory, if it exits.
You can specify a different config file with the --config
command-line option:
$ blip --config FILE
Or, you can specify a config file with the BLIP_CONFIG
environment variable:
$ export BLIP_CONFIG=FILE
$ blip
The command-line option takes precedent over the environment variable. In the following example, Blip uses only FILE_2
:
$ export BLIP_CONFIG=FILE_1
$ blip --config FILE_2
Command Line Options
Run blip --help
to list command line options.
--config FILE
Default: blip.yaml
Env var: BLIP_CONFIG
Specify Blip configuration file. The specified file must exist, else Blip will error on boot.
Blip will boot successfully if the default file, blip.yaml
, does not exist in the current working directly. In this case, Blip tries to auto-detect a local MySQL instance, which is useful for local development.
--debug
Env var: BLIP_CONFIG
Print debug to STDERR.
--help
Print help and exit.
--log
Default: false
Env var: BLIP_LOG
Print all โinfoโ events to STDOUT. By default, Blip prints only errors to STDERR
. See Logging.
--print-config
Print the server config after booting.
This option does not stop Blip from running. Specify --run=false
to print the final server config and exit.
--print-domains
Print domains and collector options, then exit.
--print-monitors
Print the monitors after booting. The monitors are finalized: monitor defaults and interpolation have been applied.
This option does not stop Blip from running. Specify --run=false
to print monitors and exit.
--run
Default: true
Env var: BLIP_RUN
Run Blip and all monitors after successful boot. If --run=false
(or environment varialbe BLIP_RUN=false
), Blip starts and loads everything, but exits before running monitors. See Startup.
--version
Print version and exit.
Startup
Blip has a two-phase startup sequence: boot, then run.
The boot phase loads and validates everything: Blip config, monitors, plans, and so forth. Any error on boot causes Blip to exit with a non-zero exit status.
The run phase runs monitors and the Blip API. Once running, Blip does not exit until it receives a signal. It retires on all errors, including panic.
To boot without running (which is useful to validate everything), specify --run=false
on the commnand line, or environment varialbe BLIP_RUN=false
.
Signals
Blip does a controlled shutdown on SIGTERM
. Other signals are not caught.