Signal notifier for Prometheus alertmanager

This implements an alertmanager webhook that can connect to signald.

Building

go get github.com/dgl/alertmanager-webhook-signald

Will give you a alertmanager-webhook-signald binary to run, in your Go bin directory ($(go env GOPATH)/bin/alertmanager-webhook-signald).

Setup

Follow the account creation steps at https://signald.org/articles/getting-started/ and get a number registered in your signald instance. Set this number as the 'sender' in the config below.

Example:

go get gitlab.com/signald/signald-go/cmd/signaldctl
signaldctl account register +1555...

Configuration

Save something like the following as config.yaml:

defaults:
  # Phone number of sender, must be registered in this signald per Setup.
  sender: +1555...
  template: '{{ template "signal.message" . }}'
  # Subscribe to responses from signald. May help to keep the connection alive.
  subscribe: true

templates:
  # Copy this file to the same place as the configuration file.
  - "alerts.tmpl"

receivers:
  - name: something
    to:
      - group:xxxx
      - tel:+44...
    # Optional: the sender, template, etc. fields as in defaults above.

See example.yaml for a more complete configuration example.

To get ID to put after group: you can use signaldctl, add the account to the group, then run signaldctl group list -a +1555....

You'll also need a template file for the alert message text, just putting alerts.tmpl in the same directory as the configuration file will work for most cases.

Alertmanager configuration

receivers:
  - name: something
    webhook_configs:
      - url: http://localhost:9716/alert

The receiver name defined in alertmanager configuration will be sent to the receiver with the matching name in the receivers section of the configuration file (i.e. "something" in this example must be the same string in both alertmanager configuration and this webhook's configuration).

Running

alertmanager-webhook-signald -config config.yaml

Monitoring

Prometheus configuration

Use Prometheus to check the health of the webhook itself.

Configure Prometheus to scrape it:

scrape_configs:
  - job_name: alertmanager-signald-webhook
    static_configs:
      - targets: ['localhost:9716']

Configure some rules like the rules in example-rules.yaml to alert you -- ideally via another alert receiver!