brennerm / uptimerobot-operator

A Kubernetes operator that creates UptimeRobot monitors for your ingresses

uptimerobot-operator

This operator automatically creates uptime monitors at UptimeRobot for your Kubernetes Ingress resources. This allows you to easily integrate uptime monitoring of your services into your Kubernetes deployments.

⚠️ This project is still in an early phase. Use it on your own risk but make sure to create issues for issues you encounter.

❤️ to UptimeRobot for providing a pro account to be able to test the pro plan features!

Usage

After installing the uptimerobot-operator it'll watch for Ingress resources in your Kubernetes cluster. For example after creating the following Ingress

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: my-ingress
  annotations:
    uroperator.brennerm.github.io/monitor.type: HTTPS
    uroperator.brennerm.github.io/monitor.interval: "600"
spec:
  rules:
  - host: brennerm.github.io
    http:
      paths:
      - pathType: Prefix
        path: "/"
        backend:
          service:
            name: service1
            port:
              number: 80

a new monitor for the URL https://brennerm.github.io is automatically being created in your UptimeRobot account.

The same monitor can also be created using the UptimeRobotMonitor resource like so:

apiVersion: uroperator.brennerm.github.io/v1beta1
kind: UptimeRobotMonitor
metadata:
  name: my-custom-monitor
spec:
  url: "https://brennerm.github.io"
  type: HTTPS
  interval: 600

The operator also supports creating public status pages. See below for details.

Installation

Create an UptimeRobot API key

1. Create an or login to your UptimeRobot account (no credit card required and they provide up to 50 monitors for free)
2. Go to "My Settings"
3. Generate and save "Main API Key" (the other API keys do not provide sufficient permissions to create, update and delete monitors)

Deploying to Kubernetes using Helm

1. Add the uptimerobot-operator chart repo helm repo add uptimerobot-operator https://brennerm.github.io/uptimerobot-operator/helm
2. Deploy the Helm chart helm upgrade --install uptimerobot-operator uptimerobot-operator --set uptimeRobotApiKey=$MY_UPTIMEROBOT_API_KEY

Have a look at the values file if you want to customize the deployment.

Running local

ℹ️ The following commands will make the operator work with your currently selected Kubernetes cluster (kubectl config current-context).

1. Install all dependencies pipenv install
2. Set UptimeRobot API key export UPTIMEROBOT_API_KEY=$MY_UPTIMEROBOT_API_KEY
3. Start operator kopf run --standalone ur_operator/handlers.py

Running in self-built Docker

1. Build Docker image docker build -t uptimerobot-operator .
2. Start container docker run -e UPTIMEROBOT_API_KEY=$MY_UPTIMEROBOT_API_KEY -v ~/.kube:/home/ur_operator/.kube uptimerobot-operator

Running in pre-built Docker

1. Start container docker run -e UPTIMEROBOT_API_KEY=$MY_UPTIMEROBOT_API_KEY -v ~/.kube:/home/ur_operator/.kube ghcr.io/brennerm/uptimerobot-operator:latest

Documentation

UptimeRobotMonitor

The UptimeRobotMonitor resource supports all current parameters for monitors that UptimeRobot offers. Below you can find a list that contains all of them.

key	type	description
url (required)	string	URL that will be monitored
type (required)	string	Type of monitor, one of: HTTP,HTTPS,KEYWORD,PING,PORT,HEARTBEAT
friendlyName	string	Friendly name of monitor, defaults to name of UptimeRobotMonitor object
subType	string	Subtype of monitor, one of: HTTP,HTTPS,KEYWORD,PING,PORT,HEARTBEAT
port	integer	Port to monitor when using monitor sub type PORT
keywordType	string	Keyword type when using monitor type KEYWORD, one of: EXISTS,NOT_EXISTS
keywordValue	string	Keyword value when using monitor type KEYWORD
interval	integer	The interval for the monitoring check (300 seconds by default)
httpUsername	string	Used for password protected pages when using monitor type HTTP,HTTP or KEYWORD, deprecated: use httpAuthSecret
httpPassword	string	Used for password protected pages when using monitor type HTTP,HTTP or KEYWORD, deprecated: use httpAuthSecret
httpAuthSecret	string	reference to a Kubernetes secret in the same namespace containing user and password for password protected pages when using monitor type HTTP,HTTPS or KEYWORD
httpAuthType	string	Used for password protected pages when using monitor type HTTP,HTTPS or KEYWORD, one of: BASIC_AUTH,DIGEST
httpMethod	string	The HTTP method to be used, one of: HEAD,GET,POST,PUT,PATCH,DELETE,OPTIONS
postType	string	The format of data to be sent with POST, PUT, PATCH, DELETE, OPTIONS requests
postContentType	string	The Content-Type header to be sent with POST, PUT, PATCH, DELETE, OPTIONS requests, one of: TEXT_HTML,APPLICATION_JSON
postValue	object	The data to be sent with POST, PUT, PATCH, DELETE, OPTIONS requests
customHttpHeaders	object	Custom HTTP headers to be sent along monitor request, formatted as JSON
customHttpStatuses	string	Allows to define HTTP status codes that will be handled as up or down, e.g. 404:0_200:1 to accept 404 as down and 200 as up
ignoreSslErrors	boolean	Flag to ignore SSL certificate related issues
alertContacts	string	Alert contacts to be notified when monitor goes up or down. For syntax check https://uptimerobot.com/api/#newMonitorWrap
mwindows	string	Maintenance window IDs for this monitor

Ingress

For Ingress resources the same parameters are supported. You pass them through annotations attached to your Ingress with the prefix uroperator.brennerm.github.io/monitor.. See below for an example.

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: my-ingress
  annotations:
    uroperator.brennerm.github.io/monitor.type: HTTPS
    uroperator.brennerm.github.io/monitor.interval: "600"
spec:
  rules:
...

To disable ingress handling completely pass the environment variable URO_DISABLE_INGRESS_HANDLING=1 to the operator.

Public Status Pages

The PublicStatusPage resource supports all current parameters for status pages that UptimeRobot offers. Below you can find a list that contains all of them.

key	type	description
monitors (required)	string	the list of monitor IDs to be displayed in status page (the values are seperated with "-" or 0 for all monitors)
friendlyName	string	Friendly name of public status page, defaults to name of PublicStatusPage object
customDomain	string	the domain or subdomain that the status page will run on
password	string	the password for the status page, deprecated: use passwordSecret
passwordSecret	string	reference to a Kubernetes secret in the same namespace containing the password for the status page
sort	string	the sorting of the monitors on the status page, one of: FRIENDLY_NAME_A_Z,FRIENDLY_NAME_Z_A,STATUS_UP_DOWN_PAUSED,STATUS_DOWN_UP_PAUSED
status	string	the status of the status page, one of: PAUSED,ACTIVE
hideUrlLinks	boolean	Flag to remove the UptimeRobot link from the status page (pro plan feature)

apiVersion: uroperator.brennerm.github.io/v1beta1
kind: PublicStatusPage
metadata:
  name: my-public-status-page
spec:
  monitors: "0" # will include all monitors

Maintenance Windows

The MaintenanceWindow resource supports all current parameters for maintenance windows that UptimeRobot offers. Below you can find a list that contains all of them.

key	type	description
type (required)	string	the type of maintenance window, one of: ONCE,DAILY,WEEKLY,MONTHLY
startTime (required)	string	the start time of the maintenance window, in seconds since epoch for type MaintenanceWindowType.ONCE, in HH:mm format for the other types
duration (required)	number	the number of seconds the maintenance window will be active
friendlyName	string	friendly name of the maintenance window, defaults to name of the MaintenanceWindow object
value	string	allows to specify the maintenance window selection, e.g. 2-4-5 for Tuesday-Thursday-Friday or 10-17-26 for the days of the month, only valid and required for MaintenanceWindowType.WEEKLY and MaintenanceWindowType.MONTHLY

apiVersion: uroperator.brennerm.github.io/v1beta1
kind: MaintenanceWindow
metadata:
  name: my-maintenance-window
spec:
  type: DAILY
  startTime: "10:00"
  duration: 30

Alert Contacts

The AlertContact resource supports all current parameters for alert contacts that UptimeRobot offers. Below you can find a list that contains all of them.

key	type	description
type (required)	string	the type of alert contact, one of: SMS,EMAIL,TWITTER_DM,BOXCAR,WEB_HOOK,PUSHBULLET,ZAPIER,PUSHOVER,HIPCHAT,SLACK
value (required)	string	the alert contact's mail address / phone number / URL / connection string
friendlyName	string	friendly name of the alert contact, defaults to name of the AlertContact object

apiVersion: uroperator.brennerm.github.io/v1beta1
kind: AlertContact
metadata:
  name: my-alert-contact
spec:
  type: EMAIL
  value: [email protected]

Planned features

• provide a Helm chart to ease deployment ✔️
• support all configuration parameters for Monitors that UptimeRobot offers ✔️
• add support for creating Uptime Robot ✔️
  • alert contacts, ✔️
  • maintenance windows ✔️
  • public status pages using Kubernetes resources ✔️
• implement automatic detection of HTTP path of Ingress resources
• add an integration for external-dns to support creating monitors for Service resources

uptimerobot-operator's Issues

Unable to create monitor with Basic Auth

Hi there,

I've been trying to create a new Uptime monitor using the following YAML but I'm getting a validation error stating that the httpAuthSecret field is unknown.

My YAML:

apiVersion: uroperator.brennerm.github.io/v1beta1
kind: UptimeRobotMonitor
metadata:
  name: my-application-monitor
  namespace: public
spec:
  url: "https://exampledomain.com"
  type: HTTPS
  interval: 300
  friendlyName: "My Application Monitor"
  alertContacts: "1234_0_0"
  httpAuthType: BASIC_AUTH
  httpAuthSecret: auth-secret

My create command and error:

$ k create -f uptime-monitor-crd.yaml
error: error validating "uptime-monitor-crd.yaml": error validating data: ValidationError(UptimeRobotMonitor.spec): unknown field "httpAuthSecret" in io.github.brennerm.uroperator.v1beta1.UptimeRobotMonitor.spec; if you choose to ignore these errors, turn validation off with --validate=false

I've installed the operator using Helm by cloning the repository and doing it locally, as the steps listed in the installation steps were failing (I've added that snippet below but that's a different issue).

$ helm repo add uptimerobot-operator https://brennerm.github.io/uptimerobot-operator/helm
$ helm upgrade --install uptimerobot-operator uptimerobot-operator --set uptimeRobotApiKey=1234
Release "uptimerobot-operator" does not exist. Installing it now.
Error: failed to download "uptimerobot-operator" (hint: running `helm repo update` may help)

Am I doing something wrong perhaps? If there's additional information I can provide to you please let me know.

allow to reference MaintenanceWindow resources from UptimeRobotMonitor

Multiple architecture Docker image

I was reviewing your Github workflow and I see you are not building multi-arch images. It's pretty easy to do using the following Github actions:

• https://github.com/docker/setup-qemu-action
• https://github.com/docker/setup-buildx-action

I can try to open a PR if you want :)

sub_type is not supported

Hi

Thanks a lot for this useful tool!

I tried to add a monitor for a Port:

---
apiVersion: uroperator.brennerm.github.io/v1beta1
kind: UptimeRobotMonitor
metadata:
  name: MYNAME
  namespace: MYNAME
spec:
  friendlyName: "K8S MYNAME"
  url: "MYHOSTNAME"
  type: PORT
  sub_type: "Custom Port"
  port: 22
  interval: 300

This gives me the following error (I'm using Flux v2):

error validating data: ValidationError(UptimeRobotMonitor.spec): unknown field "sub_type" in io.github.brennerm.uroperator.v
1beta1.UptimeRobotMonitor.spec;

If I remove sub_type it passes validation but gets rejected by the UptimeRobot API:

[2021-06-03 09:42:24,986] kopf.objects         [ERROR   ] [MYNAME/MYNAME] Handler 'on_create' failed permanently: failed to create monitor: {'type': 'missing_parameter', 'parameter_name': 'sub_type', 'message': 'sub_type parameter │
│  is missing.'}

According to the API Documentation, sub_type is normally optional, but required for port monitoring.

I use version 0.3.0.

allow to reference K8s secret for PSP password authentication

Post CRDs in YAML manifest to repository

👋🏼 It would be cool if you could post the CRDs as YAML to this repo. I am assuming that they are created by kopf when the operator is deployed.

This is kind of an issue when you want to bootstrap a cluster and the CRDs are not present before Flux gets to installed uptimerobot-operator and your manifests for monitoring. It's kind of a chicken and egg situation. Flux will try to install the custom uroperator manifests but not see any CRDs because uptimerobot-operator is not installed quite yet.

I would like to apply the CRDs in a way that forces them to exist before the uptimerobot-operator and my manifests are installed.

Thanks!

allow to reference AlertContact resources from UptimeRobotMonitor

add resource that adds support for public status pages

add flag to enable/disable watching Ingress resources

Some users may not want to have monitors created for their Ingress resources but only create monitors using UptimeRobotMonitor resources. That's why there should be a flag to disable the monitoring of ingresses.

This looks really cool!

This looks cool :) that is all.

I am keen to use something like this in production and I love that you can manage all this via CRDs.

add resource that adds support for maintenance windows

Docs: More examples / better explanation of fields in CRDs

For example in the UptimeRobotMonitor CRD the field alertContacts is a string... what should go here? I ran the curl request to get my monitors and it returns a JSON object, do I use id, friendlyName or the full object?

curl -X POST \
  -H "Cache-Control: no-cache" \
  -H "Content-Type: application/x-www-form-urlencoded" \
  -d 'api_key=<redacted>&format=json' \
"https://api.uptimerobot.com/v2/getAlertContacts"

{
  "stat": "ok",
  "offset": 0,
  "limit": 50,
  "total": 3,
  "alert_contacts": [
    {
      "id": "<redacted>",
      "friendly_name": "<redacted>@gmail.com",
      "type": 2,
      "status": 1,
      "value": "<redacted>@gmail.com"
    },
    {
      "id": "<redacted>",
      "friendly_name": "Discord",
      "type": 11,
      "status": 2,
      "value": "https://discord.com/api/webhooks/<redacted>/slack"
    },
    {
      "id": "<redacted>",
      "friendly_name": "Phone",
      "type": 13,
      "status": 2
    }
  ]
}

This is pretty vague for the other fields too, not just alertContacts.

enable health check of operator and make use of it in Helm chart

add support for all UptimeRobotMonitor fields to Ingress resources using annotations

add resource that adds support for alert contacts

report events for UpimeRobotMonitor objects

To be able retrieve the state of a single UpimeRobotMonitor resource we should send Kubernetes events that indicate whether the creation/update of a monitor at UptimeRobot was successful or failed.

allow to reference UptimeRobotMonitor resources from PublicStatusPage

allow to reference K8s secret for monitor HTTP basic/digest authentication

support remaining parameters in UptimeRobotMonitor resource

There are a lot of parameters when creating a monitor that are currently not supported.

https://uptimerobot.com/#newMonitorWrap