Healthchecks is a cron job monitoring service. It listens for HTTP requests and email messages ("pings") from your cron jobs and scheduled tasks ("checks"). When a ping does not arrive on time, Healthchecks sends out alerts.
Healthchecks comes with a web dashboard, API, 25+ integrations for delivering notifications, monthly email reports, WebAuthn 2FA support, team management features: projects, team members, read-only access.
The building blocks are:
- Python 3.8+
- Django 4
- PostgreSQL or MySQL
Healthchecks is licensed under the BSD 3-clause license.
Healthchecks is available as a hosted service at https://healthchecks.io/.
Setting Up for Development
To set up Healthchecks development environment:
Install dependencies (Debian/Ubuntu):
$ sudo apt-get update $ sudo apt-get install -y gcc python3-dev python3-venv libpq-dev
Prepare directory for project code and virtualenv. Feel free to use a different location:
$ mkdir -p ~/webapps $ cd ~/webapps
Prepare virtual environment (with virtualenv you get pip, we'll use it soon to install requirements):
$ python3 -m venv hc-venv $ source hc-venv/bin/activate $ pip3 install wheel # make sure wheel is installed in the venv
Check out project code:
$ git clone https://github.com/healthchecks/healthchecks.git
Install requirements (Django, ...) into virtualenv:
$ pip install -r healthchecks/requirements.txt
Create database tables and a superuser account:
$ cd ~/webapps/healthchecks $ ./manage.py migrate $ ./manage.py createsuperuser
With the default configuration, Healthchecks stores data in a SQLite file
hc.sqlitein the checkout directory (
To use PostgreSQL or MySQL, see the section Database Configuration section below.
$ ./manage.py test
Run development server:
$ ./manage.py runserver
The site should now be running at
To access Django administration site, log in as a superuser, then
Healthchecks reads configuration from environment variables.
Accessing Administration Panel
Healthchecks comes with Django's administration panel where you can manually view and modify user accounts, projects, checks, integrations etc. To access it,
- if you haven't already, create a superuser account:
- log into the site using superuser credentials
- in the top navigation, "Account" dropdown, select "Site Administration"
Healthchecks must be able to send email messages, so it can send out login links and alerts to users. Specify your SMTP credentials using the following environment variables:
Implicit TLS (recommended):
EMAIL_HOST = "your-smtp-server-here.com" EMAIL_PORT = 465 EMAIL_HOST_USER = "smtp-username" EMAIL_HOST_PASSWORD = "smtp-password" EMAIL_USE_TLS = False EMAIL_USE_SSL = True
Port 465 should be the preferred method according to RFC8314 Section 3.3: Implicit TLS for SMTP Submission. Be sure to use a TLS certificate and not an SSL one.
EMAIL_HOST = "your-smtp-server-here.com" EMAIL_PORT = 587 EMAIL_HOST_USER = "smtp-username" EMAIL_HOST_PASSWORD = "smtp-password" EMAIL_USE_TLS = True
For more information, have a look at Django documentation, Sending Email section.
Healthchecks comes with a
smtpd management command, which starts up a
SMTP listener service. With the command running, you can ping your
checks by sending email messages
firstname.lastname@example.org email addresses.
Start the SMTP listener on port 2525:
$ ./manage.py smtpd --port 2525
Send a test email:
$ curl --url 'smtp://127.0.0.1:2525' \ --mail-from 'email@example.com' \ --mail-rcpt 'firstname.lastname@example.org' \ -F '='
Sending Status Notifications
healtchecks comes with a
sendalerts management command, which continuously
polls database for any checks changing state, and sends out notifications as
needed. Within an activated virtualenv, you can manually run
sendalerts command like so:
$ ./manage.py sendalerts
In a production setup, you will want to run this command from a process manager like supervisor or systemd.
Healthchecks deletes old entries from
tables automatically. By default, Healthchecks keeps the 100 most recent
pings for every check. You can set the limit higher to keep a longer history:
go to the Administration Panel, look up user's Profile and modify its
"Ping log limit" field.
For each check, Healthchecks removes notifications that are older than the oldest stored ping for same check.
Healthchecks also provides management commands for cleaning up
Remove user accounts that match either of these conditions:
Account was created more than 6 months ago, and user has never logged in. These can happen when user enters invalid email address when signing up.
Last login was more than 6 months ago, and the account has no checks. Assume the user doesn't intend to use the account any more and would probably want it removed.
$ ./manage.py pruneusers
Remove old records from the
api_tokenbuckettable. The TokenBucket model is used for rate-limiting login attempts and similar operations. Any records older than one day can be safely removed.
$ ./manage.py prunetokenbucket
Remove old records from the
api_fliptable. The Flip objects are used to track status changes of checks, and to calculate downtime statistics month by month. Flip objects from more than 3 months ago are not used and can be safely removed.
$ ./manage.py pruneflips
Remove old objects from external object storage. When an user removes a check, removes a project, or closes their account, Healthchecks does not remove the associated objects from the external object storage on the fly. Instead, you should run
pruneobjectsoccasionally (for example, once a month). This command first takes an inventory of all checks in the database, and then iterates over top-level keys in the object storage bucket, and deletes any that don't also exist in the database.
$ ./manage.py pruneobjects
When you first try these commands on your data, it is a good idea to test them on a copy of your database, not on the live database right away. In a production setup, you should also have regular, automated database backups set up.
Healthchecks optionally supports two-factor authentication using the WebAuthn
standard. To enable WebAuthn support, set the
RP_ID (relying party identifier )
setting to a non-null value. Set its value to your site's domain without scheme
and without port. For example, if your site runs on
Note that WebAuthn requires HTTPS, even if running on localhost. To test WebAuthn
locally with a self-signed certificate, you can use the
Healthchecks supports external authentication by means of HTTP headers set by reverse proxies or the WSGI server. This allows you to integrate it into your existing authentication system (e.g., LDAP or OAuth) via an authenticating proxy. When this option is enabled, healtchecks will trust the header's value implicitly, so it is very important to ensure that attackers cannot set the value themselves (and thus impersonate any user). How to do this varies by your chosen proxy, but generally involves configuring it to strip out headers that normalize to the same name as the chosen identity header.
To enable this feature, set the
REMOTE_USER_HEADER value to a header you wish to
authenticate with. HTTP headers will be prefixed with
HTTP_ and have any dashes
converted to underscores. Headers without that prefix can be set by the WSGI server
itself only, which is more secure.
REMOTE_USER_HEADER is set, Healthchecks will:
- assume the header contains user's email address
- look up and automatically log in the user with a matching email address
- automatically create an user account if it does not exist
- disable the default authentication methods (login link to email, password)
External Object Storage
Healthchecks can optionally store large ping bodies in S3-compatible object storage. To enable this feature, you will need to:
- ensure you have the MinIO Python library installed:
pip install minio
- configure the credentials for accessing object storage:
Healthchecks will use external object storage for storing any request bodies that exceed 100 bytes. If the size of a request body is 100 bytes or below, Healthchecks will still store it in the database.
Healthchecks automatically removes old stored ping bodies from object
storage while uploading new data. However, Healthchecks does not automatically
clean up data when you delete checks, projects or entire user accounts.
pruneobjects management command to remove data for checks that don't
exist any more.
To enable the Slack "self-service" integration, you will need to create a "Slack App".
To do so:
- Create a new Slack app on https://api.slack.com/apps/
- Add at least one scope in the permissions section to be able to deploy the app in your workspace (By example
Bot Token Scopeshttps://api.slack.com/apps/APP_ID/oauth?).
- Add a redirect url in the format
SITE_ROOT/integrations/add_slack_btn/. For example, if your SITE_ROOT is
https://my-hc.example.orgthen the redirect URL would be
- Look up your Slack app for the Client ID and Client Secret at https://api.slack.com/apps/APP_ID/general? . Put them
To enable Discord integration, you will need to:
- register a new application on https://discordapp.com/developers/applications/me
- add a redirect URI to your Discord application. The URI format is
SITE_ROOT/integrations/add_discord/. For example, if you are running a development server on
localhost:8000then the redirect URI would be
- Look up your Discord app's Client ID and Client Secret. Put them
Pushover integration works by creating an application on Pushover.net which is then subscribed to by Healthchecks users. The registration workflow is as follows:
- On Healthchecks, the user adds a "Pushover" integration to a project
- Healthchecks redirects user's browser to a Pushover.net subscription page
- User approves adding the Healthchecks subscription to their Pushover account
- Pushover.net HTTP redirects back to Healthchecks with a subscription token
- Healthchecks saves the subscription token and uses it for sending Pushover notifications
To enable the Pushover integration, you will need to:
- Register a new application on Pushover via https://pushover.net/apps/build.
- Within the Pushover 'application' configuration, enable subscriptions.
Make sure the subscription type is set to "URL". Also make sure the redirect
URL is configured to point back to the root of the Healthchecks instance
- Put the Pushover application API Token and the Pushover subscription URL in
PUSHOVER_SUBSCRIPTION_URLenvironment variables. The Pushover subscription URL should look similar to
Healthchecks uses signal-cli to send Signal notifications. Healthcecks interacts with signal-cli over UNIX socket (requires signal-cli 0.10.0 or later).
To enable the Signal integration:
- Set up and configure signal-cli to expose JSON RPC on an UNIX socket
signal-cli -a +xxxxxx daemon --socket /tmp/signal-cli-socket
- Put the socket's location in the
Create a Telegram bot by talking to the BotFather. Set the bot's name, description, user picture, and add a "/start" command. To avoid user confusion, please do not use the Healthchecks.io logo as your bot's user picture, use your own logo.
After creating the bot you will have the bot's name and token. Put them in
settelegramwebhookmanagement command. This command tells Telegram where to forward channel messages by invoking Telegram's setWebhook API call:
$ ./manage.py settelegramwebhook Done, Telegram's webhook set to: https://my-monitoring-project.com/integrations/telegram/bot/
For this to work, your
SITE_ROOT must be correct and must use the "https://"
To enable Apprise integration, you will need to:
- ensure you have apprise installed in your local environment:
pip install apprise
- enable the apprise functionality by setting the
The "Shell Commands" integration runs user-defined local shell commands when checks
go up or down. This integration is disabled by default, and can be enabled by setting
SHELL_ENABLED environment variable to
Note: be careful when using "Shell Commands" integration, and only enable it when
you fully trust the users of your Healthchecks instance. The commands will be executed
manage.py sendalerts process, and will run with the same system permissions as
To enable the Matrix integration you will need to:
- Register a bot user (for posting notifications) in your preferred homeserver.
- Use the Login API call to retrieve bot user's access token. You can run it as shown in the documentation, using curl in command shell.
- Set the
MATRIX_environment variables. Example:
MATRIX_HOMESERVER=https://matrix.org MATRIX_USER_ID=@mychecks:matrix.org MATRIX_ACCESS_TOKEN=[a long string of characters returned by the login call]
PagerDuty Simple Install Flow
To enable PagerDuty Simple Install Flow,
- Register a PagerDuty app at PagerDuty › Developer Mode › My Apps
- In the newly created app, add the "Events Integration" functionality
- Specify a Redirect URL:
- Copy the displayed app_id value (PXXXXX) and put it in the
Running in Production
Here is a non-exhaustive list of pointers and things to check before launching a Healthchecks instance in production.
- Environment variables, settings.py and local_settings.py.
- DEBUG. Make sure it is
- ALLOWED_HOSTS. Make sure it contains the correct domain name you want to use.
- Server Errors. When DEBUG=False, Django will not show detailed error pages, and will not print exception tracebacks to standard output. To receive exception tracebacks in email, review and edit the ADMINS and SERVER_EMAIL settings. Consider setting up exception logging with Sentry.
- DEBUG. Make sure it is set to
- Management commands that need to be run during each deployment.
manage.py compress– creates combined JS and CSS bundles and places them in the
manage.py collectstatic– collects static files in the
manage.py migrate– applies any pending database schema changes and data migrations.
- Processes that need to be running constantly.
manage.py runserveris intended for development only. Do not use it in production, instead consider using uWSGI or gunicorn.
manage.py sendalertsis the process that monitors checks and sends out monitoring alerts. It must be always running, it must be started on reboot, and it must be restarted if it itself crashes. On modern linux systems, a good option is to define a systemd service for it.
- Static files. Healthchecks serves static files on its own, no configuration required. It uses the Whitenoise library for this.
- Make sure the database is secured well and is getting backed up regularly
- Make sure the TLS certificates are secured well and are getting refreshed regularly
- Have monitoring in place to be sure the Healthchecks instance itself is operational (is accepting pings, is sending out alerts, is not running out of resources).