Skip to content

prom2teams is an HTTP server built with Python that receives alert notifications from a previously configured Prometheus Alertmanager instance and forwards it to Microsoft Teams using defined connectors

License

Notifications You must be signed in to change notification settings

idealista/prom2teams

Repository files navigation

logo

Build Status Quality Gate Status Docker Build Status Docker Hub Pulls

prom2teams: Prometheus Alertmanager/Microsoft Teams integration

Alert example

prom2teams is a service built with Python that receives alert notifications from a previously configured Prometheus Alertmanager instance and forwards it to Microsoft Teams using defined connectors.

It presents grouping of alerts, labels/annotations exclusion and a Teams' alert retry policy among its key features.

Getting Started

Prerequisites

The application has been tested with Prometheus 2.2.1, Python 3.8.0 and pip 9.0.1.

Newer versions of Prometheus/Python/pip should work but could also present issues.

Installing

prom2teams is present on PyPI, so could be installed using pip3:

$ pip3 install prom2teams

Note: Works since v1.1.1

Usage

Important: Config path must be provided with at least one Microsoft Teams Connector. Check the options to know how you can supply it.

# To start the server (enable metrics, config file path , group alerts by, log file path, log level and Jinja2 template path are optional arguments):
$ prom2teams [--enablemetrics] [--configpath <config file path>] [--groupalertsby ("name"|"description"|"instance"|"severity"|"summary")] [--logfilepath <log file path>] [--loglevel (DEBUG|INFO|WARNING|ERROR|CRITICAL)] [--templatepath <Jinja2 template file path>]

# To show the help message:
$ prom2teams --help

Other options to start the service are:

export APP_CONFIG_FILE=<config file path>
$ prom2teams

Note: Grouping alerts works since v2.2.1

Docker image

Every new Prom2teams release, a new Docker image is built in our Dockerhub. We strongly recommend you to use the images with the version tag, though it will be possible to use them without it.

There are two things you need to bear in mind when creating a Prom2teams container:

  • The connector URL must be passed as the environment variable PROM2TEAMS_CONNECTOR
  • In case you want to group alerts, you need to pass the field as the environment variable PROM2TEAMS_GROUP_ALERTS_BY
  • You need to map container's Prom2teams port to one on your host.

So a sample Docker run command would be:

$ docker run -it -d -e PROM2TEAMS_GROUP_ALERTS_BY=FIELD_YOU_WANT_TO_GROUP_BY -e PROM2TEAMS_CONNECTOR="CONNECTOR_URL" -p 8089:8089 idealista/prom2teams:VERSION

Provide custom config file

If you prefer to use your own config file, you just need to provide it as a Docker volume to the container and map it to /opt/prom2teams/config.ini. Sample:

$ docker run -it -d -v pathToTheLocalConfigFile:/opt/prom2teams/config.ini -p 8089:8089 idealista/prom2teams:VERSION

Helm chart

Installing the Chart

To install the chart with the release name my-release run:

$ helm install --name my-release /location/of/prom2teams_ROOT/helm

After a few seconds, Prom2Teams should be running.

Tip: List all releases using helm list, a release is a name used to track a specific deployment

Uninstalling the Chart

To uninstall/delete the my-release deployment:

Helm 2
$ helm delete my-release

Tip: Use helm delete --purge my-release to completely remove the release from Helm internal storage

The command removes all the Kubernetes components associated with the chart and deletes the release.

Helm 3
$ helm uninstall my-release

The command removes all the Kubernetes components associated with the chart and deletes the release.

Configuration

The following table lists the configurable parameters of the Prom2teams chart and their default values.

Parameter Description Default
image.repository The image repository to pull from idealista/prom2teams
image.tag The image tag to pull <empty>
image.pullPolicy The image pull policy IfNotPresent
resources.requests.cpu CPU requested for being run in a node 100m
resources.requests.memory Memory requested for being run in a node 128Mi
resources.limits.cpu CPU limit 200m
resources.limits.memory Memory limit 200Mi
service.type Service Map (NodePort/ClusterIP) ClusterIP
service.port Service Port 8089
prom2teams.host IP to bind to 0.0.0.0
prom2teams.port Port to bind to 8089
prom2teams.connector Connector URL <empty>
prom2teams.connectors A map where the keys are the connector names and the values are the connector webhook urls {}
prom2teams.group_alerts_by Group_alerts_by field <empty>
prom2teams.loglevel Loglevel INFO
prom2teams.templatepath Custom Template path (files/teams.j2) /opt/prom2teams/helmconfig/teams.j2
prom2teams.config Config (specific to Helm) /opt/prom2teams/helmconfig/config.ini
prom2teams.extraEnv Dictionary of arbitrary additional environment variables for deployment (eg. HTTP_PROXY) <empty>

Production

For production environments you should prefer using a WSGI server. uWSGI dependency is installed for an easy usage. Some considerations must be taken to use it:

The binary prom2teams_uwsgi launches the app using the uwsgi server. Due to some incompatibilities with wheel you must install prom2teams using sudo pip install --no-binary :all: prom2teams (pypa/wheel#92)

$ prom2teams_uwsgi <path to uwsgi ini config>

And uwsgi would look like:

[uwsgi]
master = true
processes = 5
#socket = 0.0.0.0:8001
#protocol = http
socket = /tmp/prom2teams.sock
chmod-socket = 777
vacuum = true
env = APP_ENVIRONMENT=pro
env = APP_CONFIG_FILE=/etc/default/prom2teams.ini

Consider not provide chdir property neither module property.

Also you can set the module file, by doing a symbolic link: sudo mkdir -p /usr/local/etc/prom2teams/ && sudo ln -sf /usr/local/lib/python3.7/dist-packages/usr/local/etc/prom2teams/wsgi.py /usr/local/etc/prom2teams/wsgi.py (check your dist-packages folder)

Another approach is to provide yourself the module file module example and the bin uwsgi call uwsgi example

Note: default log level is DEBUG. Messages are redirected to stdout. To enable file log, set the env APP_ENVIRONMENT=(pro|pre)

Config file

The config file is an INI file and should have the structure described below:

[Microsoft Teams]
# At least one connector is required here
Connector: <webhook url>
AnotherConnector: <webhook url>   
...

[HTTP Server]
Host: <host ip> # default: localhost
Port: <host port> # default: 8089

[Log]
Level: <loglevel (DEBUG|INFO|WARNING|ERROR|CRITICAL)> # default: DEBUG
Path: <log file path>  # default: /var/log/prom2teams/prom2teams.log

[Template]
Path: <Jinja2 template path> # default: app resources default template (./prom2teams/resources/templates/teams.j2)

[Group Alerts]
Field: <Field to group alerts by> # alerts won't be grouped by default

[Labels]
Excluded: <Comma separated list of labels to ignore>

[Annotations]
Excluded: <Comma separated list of annotations to ignore>

[Teams Client]
RequestTimeout: <Configures the request timeout> # defaults to 30 secs
RetryEnable: <Enables teams client retry policy> # defaults to false
RetryWaitTime: <Wait time between retries> # default: 60 secs
MaxPayload: <Teams client payload limit in bytes> # default: 24KB

Note: Grouping alerts works since v2.2.0

Configuring Prometheus

The webhook receiver in Prometheus allows configuring a prom2teams server.

The url is formed by the host and port defined in the previous step.

Note: In order to keep compatibility with previous versions, v2.0 keep attending the default connector ("Connector") in the endpoint 0.0.0.0:8089. This will be removed in future versions.

// The prom2teams endpoint to send HTTP POST requests to.
url: 0.0.0.0:8089/v2/<Connector1>

Prom2teams Prometheus metrics

Prom2teams uses Flask and, to have the service monitored, we use @rycus66's Prometheus Flask Exporter. This will enable an endpoint in /metrics where you could find interesting metrics to monitor such as number of responses with a certain status. To enable this endpoint, just either:

  • Use the --enablemetrics or -m flag when launching prom2teams.
  • Set the environment variable PROM2TEAMS_PROMETHEUS_METRICS=true.

Templating

prom2teams provides a default template built with Jinja2 to render messages in Microsoft Teams. This template could be overrided using the 'templatepath' argument ('--templatepath ') during the application start.

Some fields are considered mandatory when received from Alert Manager. If such a field is not included a default value of 'unknown' is assigned.

All non-mandatory labels not in excluded list are injected in extra_labels key. All non-mandatory annotations not in excluded list are injected in extra_annotations key.

Alertmanager fingerprints are available in the fingerprint key. Fingerprints are supported by Alertmanager 0.19.0 or greater.

Documentation

Swagger UI

Accessing to <Host>:<Port> (e.g. localhost:8089) in a web browser shows the API v1 documentation.

Swagger UI

Accessing to <Host>:<Port>/v2 (e.g. localhost:8089/v2) in a web browser shows the API v2 documentation.

Swagger UI

Testing

To run the test suite you should type the following:

// After cloning prom2teams :)
$ pip install -r requirements.txt
$ python3 -m unittest discover tests
$ cd tests/e2e
$ ./test.sh

Built With

Python 3.8.0 pip 9.0.1

Versioning

For the versions available, see the tags on this repository.

Additionaly you can see what change in each version in the CHANGELOG.md file.

Authors

See also the list of contributors who participated in this project.

License

Apache 2.0 License

This project is licensed under the Apache 2.0 license - see the LICENSE file for details.

Contributing

Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.