A minimal monitoring tool
Go to file
Ian Fijolek f6a9199f25
Some checks failed
continuous-integration/drone/push Build is failing
Change exec find
2021-01-08 18:01:58 -05:00
scripts Scripts: Add echoing log lines to helper scripts 2020-11-16 15:52:21 -08:00
test Add a default log alert 2020-02-19 17:38:07 -08:00
.drone.yml Change exec find 2021-01-08 18:01:58 -05:00
.gitignore Move static binaries to dist/ for easier publishing 2020-07-06 20:15:21 -07:00
.pre-commit-config.yaml Remove docker linting for now 2020-07-14 17:29:54 -07:00
alert_test.go Make default log alert conditional 2020-06-19 09:51:42 -07:00
alert.go Update for go 1.15 2020-11-16 15:56:31 -08:00
compress_release.sh Change exec find 2021-01-08 18:01:58 -05:00
config_test.go Make Python compatability a flag 2020-02-19 17:38:07 -08:00
config.go Make Python compatability a flag 2020-02-19 17:38:07 -08:00
Dockerfile Move static binaries to dist/ for easier publishing 2020-07-06 20:15:21 -07:00
Dockerfile.multi-stage Update Dockerfiles to newer (roughly) pinned versions 2019-11-22 14:44:21 -08:00
go.mod Add prometheus metrics exporter 2019-11-15 17:14:20 -08:00
go.sum Add prometheus metrics exporter 2019-11-15 17:14:20 -08:00
LICENSE Add pre-commit to the repo 2019-11-21 15:32:57 -08:00
main_test.go Switch to a single key for command and command shell 2020-02-19 17:38:06 -08:00
main.go Update exported status metric to properly reflect alerting status of a monitor 2020-07-14 17:09:56 -07:00
Makefile Fix Makefile comment 2020-11-16 15:51:41 -08:00
manifest.tmpl Add multi-arch builds 2020-01-10 13:58:17 -08:00
metrics.go Add pre-commit to the repo 2019-11-21 15:32:57 -08:00
monitor_test.go Update exported status metric to properly reflect alerting status of a monitor 2020-07-14 17:09:56 -07:00
monitor.go Update exported status metric to properly reflect alerting status of a monitor 2020-07-14 17:09:56 -07:00
README.md Make default log alert conditional 2020-06-19 09:51:42 -07:00
sample-config.yml Switch to a single key for command and command shell 2020-02-19 17:38:06 -08:00
util_test.go Refactor a bit more for testing, update tests 2019-10-04 14:47:38 -07:00
util.go Fix issue with shell commands containing "<>" and unecessary (and poor) escaping 2020-01-07 10:37:53 -08:00

minitor-go

A minimal monitoring system

What does it do?

Minitor accepts a YAML configuration file with a set of commands to run and a set of alerts to execute when those commands fail. It is designed to be as simple as possible and relies on other command line tools to do checks and issue alerts.

But why?

I'm running a few small services and found Sensu, Consul, Nagios, etc. to all be far too complicated for my usecase.

So how do I use it?

Running

Install and execute with:

go get github.com/iamthefij/minitor-go
minitor

If locally developing you can use:

make run

It will read the contents of config.yml and begin its loop. You could also run it directly and provide a new config file via the -config argument.

Docker

You can pull this repository directly from Docker:

docker pull iamthefij/minitor-go:latest

The Docker image uses a default config.yml that is copied from sample-config.yml. This won't really do anything for you, so when you run the Docker image, you should supply your own config.yml file:

docker run -v $PWD/config.yml:/app/config.yml iamthefij/minitor-go:latest

Images are provided for amd64, arm, and arm64 architechtures.

Configuring

In this repo, you can explore the sample-config.yml file for an example, but the general structure is as follows. It should be noted that environment variable interpolation happens on load of the YAML file.

The global configurations are:

key value
check_interval Maximum frequency to run checks for each monitor
monitors List of all monitors. Detailed description below
alerts List of all alerts. Detailed description below

Monitors

All monitors should be listed under monitors.

Each monitor allows the following configuration:

key value
name Name of the monitor running. This will show up in messages and logs.
command Specifies the command that should be executed, either in exec or shell form. This command's exit value will determine whether the check is successful
alert_down A list of Alerts to be triggered when the monitor is in a "down" state
alert_up A list of Alerts to be triggered when the monitor moves to an "up" state
check_interval The interval at which this monitor should be checked. This must be greater than the global check_interval value
alert_after Allows specifying the number of failed checks before an alert should be triggered
alert_every Allows specifying how often an alert should be retriggered. There are a few magic numbers here. Defaults to -1 for an exponential backoff. Setting to 0 disables re-alerting. Positive values will allow retriggering after the specified number of checks

Alerts

Alerts exist as objects keyed under alerts. Their key should be the name of the Alert. This is used in your monitor setup in alert_down and alert_up.

Eachy alert allows the following configuration:

key value
command Specifies the command that should be executed, either in exec or shell form. This is the command that will be run when the alert is executed. This can be templated with environment variables or the variables shown in the table below

Also, when alerts are executed, they will be passed through Go's format function with arguments for some attributes of the Monitor. The following monitor specific variables can be referenced using Go formatting syntax:

token value
{{.AlertCount}} Number of times this monitor has alerted
{{.FailureCount}} The total number of sequential failed checks for this monitor
{{.LastCheckOutput}} The last returned value from the check command to either stderr or stdout
{{.LastSuccess}} The ISO datetime of the last successful check
{{.MonitorName}} The name of the monitor that failed and triggered the alert
{{.IsUp}} Indicates if the monitor that is alerting is up or not. Can be used in a conditional message template

Metrics

Minitor supports exporting metrics for Prometheus. Prometheus is an open source tool for reading and querying metrics from different sources. Combined with another tool, Grafana, it allows building of charts and dashboards. You could also opt to just use Minitor to log check results, and instead do your alerting with Grafana.

It is also possible to use the metrics endpoint for monitoring Minitor itself! This allows setting up multiple instances of Minitor on different servers and have them monitor each-other so that you can detect a minitor outage.

To run minitor with metrics, use the -metrics flag. The metrics will be served on port 8080 by default, though it can be overriden using -metrics-port. They will be accessible on the path /metrics. Eg. localhost:8080/metrics.

minitor -metrics
# or
minitor -metrics -metrics-port 3000

Contributing

Whether you're looking to submit a patch or just tell me I broke something, you can contribute through the Github mirror and I can merge PRs back to the source repository.

Primary Repo: https://git.iamthefij.com/iamthefij/minitor.git

Github Mirror: https://github.com/IamTheFij/minitor.git

Original Minitor

This is a reimplementation of Minitor in Go

Minitor is already a minimal monitoring tool. Python 3 was a quick way to get something live, but Python itself comes with a large footprint. Thus Go feels like a better fit for the project, longer term.

Initial target is meant to be roughly compatible requiring only minor changes to configuration. Future iterations may diverge to take advantage of Go specific features.

Differences from Python version

Templating for Alert messages has been updated. In the Python version, str.format(...) was used with certain keys passed in that could be used to format messages. In the Go version, we use a struct, AlertNotice defined in alert.go and the built in Go templating format. Eg.

minitor-py:

alerts:
  log:
    command: 'echo {monitor_name}'

minitor-go:

alerts:
  log:
    command: 'echo {{.MonitorName}}'

For the time being, legacy configs for the Python version of Minitor should be compatible if you apply the -py-compat flag when running Minitor. Eventually, this flag will go away when later breaking changes are introduced.

Future

Future, potentially breaking changes

  • Implement leveled logging (maybe glog or logrus)
  • Consider value of templating vs injecting values into Env variables
  • Async checking
  • Revisit metrics and see if they all make sense
  • Consider dropping alert_up and alert_down in favor of using Go templates that offer more control of messaging (Breaking)
  • Use durations rather than seconds checked in event loop (Potentially breaking)