Gatus is a health dashboard that gives you the ability to monitor your services using HTTP, ICMP, TCP, and even DNS queries as well as evaluate the result of said queries by using a list of conditions on values like the status code, the response time, the certificate expiration, the body and many others. The icing on top is that each of these health checks can be paired with alerting via Slack, PagerDuty, Discord and even Twilio.

I personally deploy it in my Kubernetes cluster and let it monitor the status of my core applications: https://status.twinnation.org/

docker run -p 8080:8080 --name gatus twinproduction/gatus

• Why Gatus?
• Features
• Usage
• Configuration
  • Conditions
    • Placeholders
    • Functions
  • Storage
  • Client configuration
  • Alerting
    • Configuring Discord alerts
    • Configuring Mattermost alerts
    • Configuring Messagebird alerts
    • Configuring PagerDuty alerts
    • Configuring Slack alerts
    • Configuring Teams alerts
    • Configuring Telegram alerts
    • Configuring Twilio alerts
    • Configuring custom alerts
    • Setting a default alert
• Deployment
  • Docker
  • Helm Chart
  • Terraform
• Running the tests
• Using in Production
• FAQ
  • Sending a GraphQL request
  • Recommended interval
  • Default timeouts
  • Monitoring a TCP service
  • Monitoring a service using ICMP
  • Monitoring a service using DNS queries
  • Monitoring a service using STARTTLS
  • Basic authentication
  • disable-monitoring-lock
  • Reloading configuration on the fly
  • Service groups
  • Exposing Gatus on a custom port
  • Badges
    • Uptime
    • Response time
  • API
  • High level design overview
• Sponsors

Before getting into the specifics, I want to address the most common question:

Why would I use Gatus when I can just use Prometheus’ Alertmanager, Cloudwatch or even Splunk?

Neither of these can tell you that there’s a problem if there are no clients actively calling the endpoint. In other words, it's because monitoring metrics mostly rely on existing traffic, which effectively means that unless your clients are already experiencing a problem, you won't be notified.

Gatus, on the other hand, allows you to configure health checks for each of your features, which in turn allows it to monitor these features and potentially alert you before any clients are impacted.

A sign you may want to look into Gatus is by simply asking yourself whether you'd receive an alert if your load balancer was to go down right now. Will any of your existing alerts by triggered? Your metrics won’t report an increase in errors if there’s no traffic that makes it to your applications. This puts you in a situation where your clients are the ones that will notify you about the degradation of your services rather than you reassuring them that you're working on fixing the issue before they even know about it.

The main features of Gatus are:

• Highly flexible health check conditions: While checking the response status may be enough for some use cases, Gatus goes much further and allows you to add conditions on the response time, the response body and even the IP address.
• Ability to use Gatus for user acceptance tests: Thanks to the point above, you can leverage this application to create automated user acceptance tests.
• Very easy to configure: Not only is the configuration designed to be as readable as possible, it's also extremely easy to add a new service or a new endpoint to monitor.
• Alerting: While having a pretty visual dashboard is useful to keep track of the state of your application(s), you probably don't want to stare at it all day. Thus, notifications via Slack, Mattermost, Messagebird, PagerDuty, Twilio and Teams are supported out of the box with the ability to configure a custom alerting provider for any needs you might have, whether it be a different provider or a custom application that manages automated rollbacks.
• Metrics
• Low resource consumption: As with most Go applications, the resource footprint that this application requires is negligibly small.
• Badges:

By default, the configuration file is expected to be at config/config.yaml.

You can specify a custom path by setting the GATUS_CONFIG_FILE environment variable.

Here's a simple example:

metrics: true         # Whether to expose metrics at /metrics
services:
  - name: twinnation  # Name of your service, can be anything
    url: "https://twinnation.org/health"
    interval: 30s     # Duration to wait between every status check (default: 60s)
    conditions:
      - "[STATUS] == 200"         # Status must be 200
      - "[BODY].status == UP"     # The json path "$.status" must be equal to UP
      - "[RESPONSE_TIME] < 300"   # Response time must be under 300ms
  - name: example
    url: "https://example.org/"
    interval: 5m
    conditions:
      - "[STATUS] == 200"

This example would look like this:

Note that you can also use environment variables in the configuration file (e.g. $DOMAIN, ${DOMAIN})

If you want to test it locally, see Docker.

Here are some examples of conditions you can use:

NOTE: Use pat only when you need to. [STATUS] == pat(2*) is a lot more expensive than [STATUS] < 300.

• If storage.type is memory (default) and storage.file is set to a non-blank value. Furthermore, the data is periodically persisted, but everything remains in memory.
• If storage.type is sqlite, storage.file must not be blank:

storage:
  type: sqlite
  file: data.db

See examples/docker-compose-sqlite-storage for an example.

• If storage.type is postgres, storage.file must be the connection URL:

storage:
  type: postgres
  file: "postgres:https://user:[email protected]:5432/gatus?sslmode=disable"

See examples/docker-compose-postgres-storage for an example.

In order to support a wide range of environments, each monitored service has a unique configuration for the client used to send the request.

Note that some of these parameters are ignored based on the type of service. For instance, there's no certificate involved in ICMP requests (ping), therefore, setting client.insecure to true for a service of that type will not do anything.

This default configuration is as follows:

client:
  insecure: false
  ignore-redirect: false
  timeout: 10s

Note that this configuration is only available under services[], alerting.mattermost and alerting.custom.

Here's an example with the client configuration under service[]:

services:
  - name: twinnation
    url: "https://twinnation.org/health"
    client:
      insecure: false
      ignore-redirect: false
      timeout: 10s
    conditions:
      - "[STATUS] == 200"

Gatus supports multiple alerting providers, such as Slack and PagerDuty, and supports different alerts for each individual services with configurable descriptions and thresholds.

Note that if an alerting provider is not properly configured, all alerts configured with the provider's type will be ignored.

alerting:
  discord: 
    webhook-url: "https://discord.com/api/webhooks/**********/**********"

services:
  - name: twinnation
    url: "https://twinnation.org/health"
    interval: 30s
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == UP"
      - "[RESPONSE_TIME] < 300"
    alerts:
      - type: discord
        enabled: true
        description: "healthcheck failed"
        send-on-resolved: true

alerting:
  mattermost: 
    webhook-url: "https://**********/hooks/**********"
    client:
      insecure: true

services:
  - name: twinnation
    url: "https://twinnation.org/health"
    interval: 30s
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == UP"
      - "[RESPONSE_TIME] < 300"
    alerts:
      - type: mattermost
        enabled: true
        description: "healthcheck failed"
        send-on-resolved: true

Here's an example of what the notifications look like:

Example of sending SMS text message alert using Messagebird:

alerting:
  messagebird:
    access-key: "..."
    originator: "31619191918"
    recipients: "31619191919,31619191920"
services:
  - name: twinnation
    interval: 30s
    url: "https://twinnation.org/health"
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == UP"
      - "[RESPONSE_TIME] < 300"
    alerts:
      - type: messagebird
        enabled: true
        failure-threshold: 3
        send-on-resolved: true
        description: "healthcheck failed"

It is highly recommended to set services[].alerts[].send-on-resolved to true for alerts of type pagerduty, because unlike other alerts, the operation resulting from setting said parameter to true will not create another incident, but mark the incident as resolved on PagerDuty instead.

alerting:
  pagerduty: 
    integration-key: "********************************"

services:
  - name: twinnation
    url: "https://twinnation.org/health"
    interval: 30s
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == UP"
      - "[RESPONSE_TIME] < 300"
    alerts:
      - type: pagerduty
        enabled: true
        failure-threshold: 3
        success-threshold: 5
        send-on-resolved: true
        description: "healthcheck failed"

alerting:
  slack: 
    webhook-url: "https://hooks.slack.com/services/**********/**********/**********"

services:
  - name: twinnation
    url: "https://twinnation.org/health"
    interval: 30s
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == UP"
      - "[RESPONSE_TIME] < 300"
    alerts:
      - type: slack
        enabled: true
        description: "healthcheck failed 3 times in a row"
        send-on-resolved: true
      - type: slack
        enabled: true
        failure-threshold: 5
        description: "healthcheck failed 5 times in a row"
        send-on-resolved: true

Here's an example of what the notifications look like:

alerting:
  teams:
    webhook-url: "https://********.webhook.office.com/webhookb2/************"

services:
  - name: twinnation
    url: "https://twinnation.org/health"
    interval: 30s
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == UP"
      - "[RESPONSE_TIME] < 300"
    alerts:
      - type: teams
        enabled: true
        description: "healthcheck failed"
        send-on-resolved: true

Here's an example of what the notifications look like:

alerting:
  telegram: 
    token: "123456:ABC-DEF1234ghIkl-zyx57W2v1u123ew11"
    id: "0123456789"

services:
  - name: twinnation
    url: "https://twinnation.org/health"
    interval: 30s
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == UP"
    alerts:
      - type: telegram
        enabled: true
        send-on-resolved: true

Here's an example of what the notifications look like:

alerting:
  twilio:
    sid: "..."
    token: "..."
    from: "+1-234-567-8901"
    to: "+1-234-567-8901"

services:
  - name: twinnation
    interval: 30s
    url: "https://twinnation.org/health"
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == UP"
      - "[RESPONSE_TIME] < 300"
    alerts:
      - type: twilio
        enabled: true
        failure-threshold: 5
        send-on-resolved: true
        description: "healthcheck failed"

While they're called alerts, you can use this feature to call anything.

For instance, you could automate rollbacks by having an application that keeps tracks of new deployments, and by leveraging Gatus, you could have Gatus call that application endpoint when a service starts failing. Your application would then check if the service that started failing was recently deployed, and if it was, then automatically roll it back.

The placeholders [ALERT_DESCRIPTION] and [SERVICE_NAME] are automatically substituted for the alert description and the service name. These placeholders can be used in the body (alerting.custom.body) and in the url (alerting.custom.url).

If you have an alert using the custom provider with send-on-resolved set to true, you can use the [ALERT_TRIGGERED_OR_RESOLVED] placeholder to differentiate the notifications. The aforementioned placeholder will be replaced by TRIGGERED or RESOLVED accordingly, though it can be modified (details at the end of this section).

For all intents and purpose, we'll configure the custom alert with a Slack webhook, but you can call anything you want.

alerting:
  custom:
    url: "https://hooks.slack.com/services/**********/**********/**********"
    method: "POST"
    body: |
      {
        "text": "[ALERT_TRIGGERED_OR_RESOLVED]: [SERVICE_NAME] - [ALERT_DESCRIPTION]"
      }
services:
  - name: twinnation
    url: "https://twinnation.org/health"
    interval: 30s
    conditions:
      - "[STATUS] == 200"
      - "[BODY].status == UP"
      - "[RESPONSE_TIME] < 300"
    alerts:
      - type: custom
        enabled: true
        failure-threshold: 10
        success-threshold: 3
        send-on-resolved: true
        description: "healthcheck failed"

Note that you can customize the resolved values for the [ALERT_TRIGGERED_OR_RESOLVED] placeholder like so:

alerting:
  custom:
    placeholders:
      ALERT_TRIGGERED_OR_RESOLVED:
        TRIGGERED: "partial_outage"
        RESOLVED: "operational"

As a result, the [ALERT_TRIGGERED_OR_RESOLVED] in the body of first example of this section would be replaced by partial_outage when an alert is triggered and operational when an alert is resolved.

While you can specify the alert configuration directly in the service definition, it's tedious and may lead to a very long configuration file.

To avoid such problem, you can use the default-alert parameter present in each provider configuration:

alerting:
  slack: 
    webhook-url: "https://hooks.slack.com/services/**********/**********/**********"
    default-alert:
      enabled: true
      description: "healthcheck failed"
      send-on-resolved: true
      failure-threshold: 5
      success-threshold: 5

As a result, your service configuration looks a lot tidier:

services:
  - name: example
    url: "https://example.org"
    conditions:
      - "[STATUS] == 200"
    alerts:
      - type: slack

  - name: other-example
    url: "https://example.com"
    conditions:
      - "[STATUS] == 200"
    alerts:
      - type: slack

It also allows you to do things like this:

services:
  - name: example
    url: "https://example.org"
    conditions:
      - "[STATUS] == 200"
    alerts:
      - type: slack
        failure-threshold: 5
      - type: slack
        failure-threshold: 10
      - type: slack
        failure-threshold: 15

Of course, you can also mix alert types:

alerting:
  slack:
    webhook-url: "https://hooks.slack.com/services/**********/**********/**********"
    default-alert:
      enabled: true
      failure-threshold: 3
  pagerduty:
    integration-key: "********************************"
    default-alert:
      enabled: true
      failure-threshold: 5

services:
  - name: service-1
    url: "https://example.org"
    conditions:
      - "[STATUS] == 200"
    alerts:
      - type: slack
      - type: pagerduty

  - name: service-2
    url: "https://example.org"
    conditions:
      - "[STATUS] == 200"
    alerts:
      - type: slack
      - type: pagerduty

Many examples can be found in the examples folder, but this section will focus on the most popular ways of deploying Gatus.

To run Gatus locally with Docker:

docker run -p 8080:8080 --name gatus twinproduction/gatus

Other than using one of the examples provided in the examples folder, you can also try it out locally by creating a configuration file, we'll call it config.yaml for this example, and running the following command:

docker run -p 8080:8080 --mount type=bind,source="$(pwd)"/config.yaml,target=/config/config.yaml --name gatus twinproduction/gatus

If you're on Windows, replace "$(pwd)" by the absolute path to your current directory, e.g.:

docker run -p 8080:8080 --mount type=bind,source=C:/Users/Chris/Desktop/config.yaml,target=/config/config.yaml --name gatus twinproduction/gatus

To build the image locally:

docker build . -t twinproduction/gatus

Helm must be installed to use the chart. Please refer to Helm's documentation to get started.

Once Helm is set up properly, add the repository as follows:

helm repo add gatus https://avakarev.github.io/gatus-chart

To get more details, please check chart's configuration and helmfile example

Gatus can be deployed on Terraform by using the following module: terraform-kubernetes-gatus.

go test ./... -mod vendor

See the Deployment section.

By setting services[].graphql to true, the body will automatically be wrapped by the standard GraphQL query parameter.

For instance, the following configuration:

services:
  - name: filter-users-by-gender
    url: https://localhost:8080/playground
    method: POST
    graphql: true
    body: |
      {
        users(gender: "female") {
          id
          name
          gender
          avatar
        }
      }
    conditions:
      - "[STATUS] == 200"
      - "[BODY].data.users[0].gender == female"

will send a POST request to https://localhost:8080/playground with the following body:

{"query":"      {\n        users(gender: \"female\") {\n          id\n          name\n          gender\n          avatar\n        }\n      }"}

NOTE: This does not really apply if disable-monitoring-lock is set to true, as the monitoring lock is what tells Gatus to only evaluate one service at a time.

To ensure that Gatus provides reliable and accurate results (i.e. response time), Gatus only evaluates one service at a time In other words, even if you have multiple services with the exact same interval, they will not execute at the same time.

You can test this yourself by running Gatus with several services configured with a very short, unrealistic interval, such as 1ms. You'll notice that the response time does not fluctuate - that is because while services are evaluated on different goroutines, there's a global lock that prevents multiple services from running at the same time.

Unfortunately, there is a drawback. If you have a lot of services, including some that are very slow or prone to time out (the default timeout is 10s), then it means that for the entire duration of the request, no other services can be evaluated.

This does mean that Gatus will be unable to evaluate the health of other services. The interval does not include the duration of the request itself, which means that if a service has an interval of 30s and the request takes 2s to complete, the timestamp between two evaluations will be 32s, not 30s.

While this does not prevent Gatus' from performing health checks on all other services, it may cause Gatus to be unable to respect the configured interval, for instance:

• Service A has an interval of 5s, and times out after 10s to complete
• Service B has an interval of 5s, and takes 1ms to complete
• Service B will be unable to run every 5s, because service A's health evaluation takes longer than its interval

To sum it up, while Gatus can really handle any interval you throw at it, you're better off having slow requests with higher interval.

As a rule of the thumb, I personally set interval for more complex health checks to 5m (5 minutes) and simple health checks used for alerting (PagerDuty/Twilio) to 30s.

To modify the timeout, see Client configuration.

By prefixing services[].url with tcp:\\, you can monitor TCP services at a very basic level:

services:
  - name: redis
    url: "tcp:https://127.0.0.1:6379"
    interval: 30s
    conditions:
      - "[CONNECTED] == true"

Placeholders [STATUS] and [BODY] as well as the fields services[].body, services[].insecure, services[].headers, services[].method and services[].graphql are not supported for TCP services.

NOTE: [CONNECTED] == true does not guarantee that the service itself is healthy - it only guarantees that there's something at the given address listening to the given port, and that a connection to that address was successfully established.

By prefixing services[].url with icmp:\\, you can monitor services at a very basic level using ICMP, or more commonly known as "ping" or "echo":

services:
  - name: ping-example
    url: "icmp:https://example.com"
    conditions:
      - "[CONNECTED] == true"

Only the placeholders [CONNECTED], [IP] and [RESPONSE_TIME] are supported for services of type ICMP. You can specify a domain prefixed by icmp:https://, or an IP address prefixed by icmp:https://.

Defining a dns configuration in a service will automatically mark that service as a service of type DNS:

services:
  - name: example-dns-query
    url: "8.8.8.8" # Address of the DNS server to use
    interval: 30s
    dns:
      query-name: "example.com"
      query-type: "A"
    conditions:
      - "[BODY] == 93.184.216.34"
      - "[DNS_RCODE] == NOERROR"

There are two placeholders that can be used in the conditions for services of type DNS:

• The placeholder [BODY] resolves to the output of the query. For instance, a query of type A would return an IPv4.
• The placeholder [DNS_RCODE] resolves to the name associated to the response code returned by the query, such as NOERROR, FORMERR, SERVFAIL, NXDOMAIN, etc.

If you have an email server that you want to ensure there are no problems with, monitoring it through STARTTLS will serve as a good initial indicator:

services:
  - name: starttls-smtp-example
    url: "starttls:https://smtp.gmail.com:587"
    interval: 30m
    conditions:
      - "[CONNECTED] == true"
      - "[CERTIFICATE_EXPIRATION] > 48h"

You can require Basic authentication by leveraging the security.basic configuration:

security:
  basic:
    username: "john.doe"
    password-sha512: "6b97ed68d14eb3f1aa959ce5d49c7dc612e1eb1dafd73b1e705847483fd6a6c809f2ceb4e8df6ff9984c6298ff0285cace6614bf8daa9f0070101b6c89899e22"

The example above will require that you authenticate with the username john.doe as well as the password hunter2.

Setting disable-monitoring-lock to true means that multiple services could be monitored at the same time.

While this behavior wouldn't generally be harmful, conditions using the [RESPONSE_TIME] placeholder could be impacted by the evaluation of multiple services at the same time, therefore, the default value for this parameter is false.

There are three main reasons why you might want to disable the monitoring lock:

• You're using Gatus for load testing (each services are periodically evaluated on a different goroutine, so technically, if you create 100 services with a 1 seconds interval, Gatus will send 100 requests per second)
• You have a lot of services to monitor
• You want to test multiple services at very short interval (< 5s)

For the sake on convenience, Gatus automatically reloads the configuration on the fly if the loaded configuration file is updated while Gatus is running.

By default, the application will exit if the updating configuration is invalid, but you can configure Gatus to continue running if the configuration file is updated with an invalid configuration by setting skip-invalid-config-update to true.

Keep in mind that it is in your best interest to ensure the validity of the configuration file after each update you apply to the configuration file while Gatus is running by looking at the log and making sure that you do not see the following message:

The configuration file was updated, but it is not valid. The old configuration will continue being used.

Failure to do so may result in Gatus being unable to start if the application is restarted for whatever reason.

I recommend not setting skip-invalid-config-update to true to avoid a situation like this, but the choice is yours to make.

If you are not using a file storage, updating the configuration while Gatus is running is effectively the same as restarting the application.

NOTE: Updates may not be detected if the config file is bound instead of the config folder. See #151.

Service groups are used for grouping multiple services together on the dashboard.

services:
  - name: frontend
    group: core
    url: "https://example.org/"
    interval: 5m
    conditions:
      - "[STATUS] == 200"

  - name: backend
    group: core
    url: "https://example.org/"
    interval: 5m
    conditions:
      - "[STATUS] == 200"

  - name: monitoring
    group: internal
    url: "https://example.org/"
    interval: 5m
    conditions:
      - "[STATUS] == 200"

  - name: nas
    group: internal
    url: "https://example.org/"
    interval: 5m
    conditions:
      - "[STATUS] == 200"

  - name: random service that isn't part of a group
    url: "https://example.org/"
    interval: 5m
    conditions:
      - "[STATUS] == 200"

The configuration above will result in a dashboard that looks like this:

By default, Gatus is exposed on port 8080, but you may specify a different port by setting the web.port parameter:

web:
  port: 8081

If you're using a PaaS like Heroku that doesn't let you set a custom port and exposes it through an environment variable instead, you can use that environment variable directly in the configuration file:

web:
  port: ${PORT}

Gatus can automatically generate a SVG badge for one of your monitored services. This allows you to put badges in your individual services' README or even create your own status page, if you desire.

The endpoint to generate a badge is the following:

/api/v1/services/{key}/uptimes/{duration}/badge.svg

Where:

• {duration} is 7d, 24h or 1h
• {key} has the pattern <GROUP_NAME>_<SERVICE_NAME> in which both variables have , /, _, , and . replaced by -.

For instance, if you want the uptime during the last 24 hours from the service frontend in the group core, the URL would look like this:

https://example.com/api/v1/services/core_frontend/uptimes/7d/badge.svg

If you want to display a service that is not part of a group, you must leave the group value empty:

https://example.com/api/v1/services/_frontend/uptimes/7d/badge.svg

Example:

![Uptime 24h](https://status.twinnation.org/api/v1/services/core_website-external/uptimes/24h/badge.svg)

If you'd like to see a visual example of each badges available, you can simply navigate to the service's detail page.

The endpoint to generate a badge is the following:

/api/v1/services/{key}/response-times/{duration}/badge.svg

Where:

• {duration} is 7d, 24h or 1h
• {key} has the pattern <GROUP_NAME>_<SERVICE_NAME> in which both variables have , /, _, , and . replaced by -.

Gatus provides a simple read-only API which can be queried in order to programmatically determine service status and history.

All services are available via a GET request to the following endpoint:

/api/v1/services/statuses

Example: https://status.twinnation.org/api/v1/services/statuses

Specific services can also be queried by using the following pattern:

/api/v1/services/{group}_{service}/statuses

Example: https://status.twinnation.org/api/v1/services/core_website-home/statuses

Gzip compression will be used if the Accept-Encoding HTTP header contains gzip.

The API will return a JSON payload with the Content-Type response header set to application/json. No such header is required to query the API.

You can find the full list of sponsors here.