Grafana OnCall open source guide
Grafana OnCall is a developer-friendly incident response tool that’s available to Grafana open source and Grafana Cloud users. The OSS version of Grafana OnCall provides the same reliable on-call management solution along with the flexibility of a self-managed environment.
This guide describes the necessary installation and configuration steps needed to configure OSS Grafana OnCall.
Install Grafana OnCall OSS
There are three Grafana OnCall OSS environments available:
- Hobby playground environment for local usage: README.md
- Development environment for contributors: development README.md
- Production environment for reliable cloud installation using Helm: Production Environment
We suggest using our official helm chart for the reliable production deployment of Grafana OnCall. It will deploy Grafana OnCall engine and celery workers, along with RabbitMQ cluster, Redis Cluster, and the database.
Note: The Grafana OnCall engine currently supports one instance of the Grafana OnCall plugin at a time.
Check the helm chart for more details.
We’ll always be happy to provide assistance with production deployment in our communities!
Update Grafana OnCall OSS
To update an OSS installation of Grafana OnCall, please see the update docs:
- Hobby playground environment: README.md
- Production Helm environment: Helm update
The Slack integration for Grafana OnCall leverages Slack API features to provide a customizable and useful integration. Refer to the following steps to configure the Slack integration:
Ensure your Grafana OnCall environment is up and running.
Grafana OnCall must be accessible through HTTPS. For development purposes, use localtunnel. For production purposes, consider establishing a proper web server with HTTPS termination. For localtunnel, refer to the following configuration:
# Choose the unique prefix instead of pretty-turkey-83 # Localtunnel will generate an url, e.g. https://pretty-turkey-83.loca.lt # it is referred as <ONCALL_ENGINE_PUBLIC_URL> below lt --port 8080 -s pretty-turkey-83 --print-requests
If using localtunnel, open your external URL and click Continue to allow requests to bypass the warning page.
Create a Slack Workspace for development, or use your company workspace.
Go to https://api.slack.com/apps and click Create an App .
From an app manifestoption and select your workspace.
Replace the text with the following YAML code block . Be sure to replace
<ONCALL_ENGINE_PUBLIC_URL>fields with the appropriate information.
_metadata: major_version: 1 minor_version: 1 display_information: name: <YOUR_BOT_NAME> features: app_home: home_tab_enabled: false messages_tab_enabled: true messages_tab_read_only_enabled: false bot_user: display_name: <YOUR_BOT_NAME> always_online: true shortcuts: - name: Create a new incident type: message callback_id: incident_create description: Creates a new OnCall incident - name: Add to resolution note type: message callback_id: add_resolution_note description: Add this message to resolution note slash_commands: - command: /oncall url: <ONCALL_ENGINE_PUBLIC_URL>/slack/interactive_api_endpoint/ description: Create a manual alert group should_escape: false - command: /escalate url: <ONCALL_ENGINE_PUBLIC_URL>/slack/interactive_api_endpoint/ description: Direct page user(s) or schedule(s) should_escape: false oauth_config: redirect_urls: - <ONCALL_ENGINE_PUBLIC_URL>/api/internal/v1/complete/slack-install-free/ - <ONCALL_ENGINE_PUBLIC_URL>/api/internal/v1/complete/slack-login/ scopes: user: - channels:read - chat:write - identify - users.profile:read bot: - app_mentions:read - channels:history - channels:read - chat:write - chat:write.customize - chat:write.public - commands - files:write - groups:history - groups:read - im:history - im:read - im:write - mpim:history - mpim:read - mpim:write - reactions:write - team:read - usergroups:read - usergroups:write - users.profile:read - users:read - users:read.email - users:write settings: event_subscriptions: request_url: <ONCALL_ENGINE_PUBLIC_URL>/slack/event_api_endpoint/ bot_events: - app_home_opened - app_mention - channel_archive - channel_created - channel_deleted - channel_rename - channel_unarchive - member_joined_channel - message.channels - message.im - subteam_created - subteam_members_changed - subteam_updated - user_profile_changed interactivity: is_enabled: true request_url: <ONCALL_ENGINE_PUBLIC_URL>/slack/interactive_api_endpoint/ org_deploy_enabled: false socket_mode_enabled: false
Set environment variables by navigating to your Grafana OnCall, then click Env Variables and set the following:
SLACK_CLIENT_OAUTH_ID = Basic Information -> App Credentials -> Client ID SLACK_CLIENT_OAUTH_SECRET = Basic Information -> App Credentials -> Client Secret SLACK_SIGNING_SECRET = Basic Information -> App Credentials -> Signing Secret SLACK_INSTALL_RETURN_REDIRECT_HOST = << OnCall external URL >>
In OnCall, navigate to ChatOps, select Slack and click Install Slack integration.
Configure additional Slack settings.
The Telegram integration for Grafana OnCall is designed for collaborative team work and improved incident response. Refer to the following steps to configure the Telegram integration:
- Ensure your Grafana OnCall environment is up and running.
- Create a Telegram bot using BotFather and save the token provided by BotFather. Please make sure to disable Group Privacy for the bot (Bot Settings -> Group Privacy -> Turn off).
- Paste the token provided by BotFather to the
TELEGRAM_TOKENvariable on the Env Variables page of your Grafana OnCall instance.
- Set the
TELEGRAM_WEBHOOK_HOSTvariable to the external address of your Grafana OnCall instance. Please note that
TELEGRAM_WEBHOOK_HOSTmust start with
https://and be publicly available (meaning that it can be reached by Telegram servers). If your host is private or local, consider using a reverse proxy (e.g. ngrok).
- Now you can connect Telegram accounts on the Users page and receive alert groups to Telegram direct messages. Alternatively, in case you want to connect Telegram channels to your Grafana OnCall environment, navigate to the ChatOps tab.
Grafana OSS-Cloud Setup
The benefits of connecting to Grafana Cloud include:
- Cloud OnCall could monitor OSS OnCall uptime using heartbeat
- SMS for user notifications
- Phone calls for user notifications.
To connect to Grafana Cloud, refer to the Cloud page in your OSS Grafana OnCall instance.
Grafana OnCall supports Twilio SMS and phone call notifications delivery. If you prefer to configure SMS and phone call notifications using Twilio, complete the following steps:
GRAFANA_CLOUD_NOTIFICATIONS_ENABLEDas False to ensure the Grafana OSS <-> Cloud connector is disabled.
- From your OnCall environment, select Env Variables and configure all variables starting with
Grafana OnCall is capable of sending emails using SMTP as a user notification step. To setup email notifications, populate the following env variables with your SMTP server credentials:
EMAIL_HOST- SMTP server host
EMAIL_HOST_USER- SMTP server user
EMAIL_HOST_PASSWORD- SMTP server password
587) - SMTP server port
True) - To enable/disable TLS
EMAIL_FROM_ADDRESS(optional) - Email address used to send emails. If not specified,
EMAIL_HOST_USERwill be used.
After enabling the email integration, it will be possible to use the
Notify by email notification step in user settings.
Inbound Email Setup
Grafana OnCall is capable of creating alert groups from Inbound Email integration.
To configure Inbound Email integration for Grafana OnCall OSS populate env variables with your Email Service Provider data:
INBOUND_EMAIL_ESP- Inbound email ESP name. Available options:
INBOUND_EMAIL_DOMAIN- Inbound email domain
INBOUND_EMAIL_WEBHOOK_SECRET- Inbound email webhook secret
You will also need to configure your ESP to forward messages to the following URL:
By default, Grafana OnCall limits email and phone notifications (calls, SMS) to 200 per user per day. The limit can be changed using env variables:
200) - phone notifications per user
200) - emails per user
Mobile application set up
Note: This application is currently in beta
Grafana OnCall OSS users can use the mobile app to receive push notifications from OnCall. Grafana OnCall OSS relies on Grafana Cloud as on relay for push notifications. You must first connect your Grafana OnCall OSS to Grafana Cloud for the mobile app to work.
Refer to Grafana OSS-Cloud Setup in this document before continuing with the mobile app.
For Grafana OnCall OSS, the mobile app QR code includes an authentication token along with a backend URL. Your Grafana OnCall OSS instance should be reachable from the same network as your mobile device, preferably from the internet.
For more information, see Grafana OnCall mobile app
Alert Group Escalation Auditor
Grafana OnCall has a periodic background task, which runs to check that all alert group escalations have finished properly. This feature, if configured, can also ping an OnCall Webhook Integration’s heartbeat URL, so that you can be alerted, in the event that something goes wrong.
Logs originating from the celery worker, for the
task, that reference a
indicate that the auditor periodic task is failing check(s) on one or more alert groups. Logs for this task which
.. passed the audit checks indicate that there were no issues with with the escalation on the audited
To configure this feature as such:
- Create a Webhook, or Formatted Webhook, Integration type.
- Under the “Heartbeat” tab in the Integration modal, copy the unique heartbeat URL that is shown.
- Set the hearbeat’s expected time interval to 15 minutes (see note below regarding
- Configure the integration’s escalation chain as necessary
- Populate the following env variables:
ALERT_GROUP_ESCALATION_AUDITOR_CELERY_TASK_HEARTBEAT_URL- integration’s unique heartbeat URL
ALERT_GROUP_ESCALATION_AUDITOR_CELERY_TASK_HEARTBEAT_INTERVAL- how often the auditor task should run. By default the task runs every 13 minutes so we therefore recommend setting the heartbeat’s expected time interval to 15 minutes. If you would like to modify this, we recommend configuring this env variable to 1 or 2 minutes less than the value set for the integration’s heartbeat expected time interval.