Export sentry project's metrics consistent with the Prometheus exposition formats

• Getting Started
  • Prerequisites
  • Install
  • Run
  • Docker
  • Samples
• Metrics
  • Project Configuration
  • Metrics Configuration
  • Basic Authentication
• Limitations
• Recomendations & Tips
• Documentation
• Contributing
• License
• Show your support
• Author

• python >= 3.7.9
• Sentry API auth token

  Authentication token permissions: project:read org:read project:releases event:read

pip install -r requirements.txt

export SENTRY_BASE_URL="https://sentry.io/api/0/"
export SENTRY_AUTH_TOKEN="[REPLACE_TOKEN]"
export SENTRY_EXPORTER_ORG="[ORGANIZATION_SLUG]"

python exporter.py

Run

docker run --name sentry-exporter -e SENTRY_AUTH_TOKEN=[REPLACE_TOKEN] -e SENTRY_EXPORTER_ORG=[ORGANIZATION_SLUG] italux/sentry-prometheus-exporter

Build local image

docker-compose build

Create a .env file

echo SENTRY_BASE_URL="https://sentry.io/api/0/"
echo SENTRY_AUTH_TOKEN="[REPLACE_TOKEN]"
echo SENTRY_EXPORTER_ORG="[organization_slug]"

Start containers

docker-compose up -d

• sentry_open_issue_events: A Number of open issues (aka is:unresolved) per project in the past 1h
• sentry_issues: Gauge Histogram of open issues split into 3 buckets: 1h, 24h, and 14d
• sentry_events: Total events counts per project
• sentry_rate_limit_events_sec: Rate limit of errors per second accepted for a project.

By default, sentry's API will be polled to retrieve all projects. If you wish for specific projects to be scraped, you can do the following:

export SENTRY_EXPORTER_PROJECTS="project1,project2,project3"

Excepting rate-limit-events metric, by default all metrics are scraped, however, issue or event-related metrics can be disabled by setting the relevant variable to False:

export SENTRY_SCRAPE_ISSUE_METRICS=False
export SENTRY_SCRAPE_EVENT_METRICS=False

Enable rate-limit-events metric by setting the relevant variable to True:

export SENTRY_SCRAPE_RATE_LIMIT_METRICS=True

By default, if SENTRY_SCRAPE_ISSUE_METRICS=True or is unset issue metrics are scraped for 1hour, 24hours and 14days. Any of these can be disabled by setting the relevant variable to False:

export SENTRY_ISSUES_1H=False
export SENTRY_ISSUES_24H=False
export SENTRY_ISSUES_14D=False

As with SENTRY_AUTH_TOKEN, all of these variables can be passed in through the docker run -e VAR_NAME=<> command or via the .env file if using Docker Compose.

Security is always a concern and metrics can contain sensitive information that you'd like to protect and HTTP Basic authentication is the most simple standard that uses fields in the HTTP header to restrict access to a specific URL.

To enable the basic authentication mechanism on Sentry Prometheus Exporter you just need to export the SENTRY_EXPORTER_BASIC_AUTH and configure the credentials as follows

Once enable, the /metrics/ page will prompt username & password window

If you enable the exporter HTTP basic authentication you'l need to configure prometheus scrape to pass the username & password defined on every scrape, please check prometheus <scrape_config> for more information.

basic_auth:
  [ username: <string> ]
  [ password: <secret> ]

Grafana Dashboard Sentry Issues & Events Overview

Prometheus configuration: prometheus.yml

scrape_configs:
  - job_name: 'sentry_exporter'
    static_configs:
    - targets: ['sentry-exporter:9790']
    scrape_interval: 5m
    scrape_timeout: 4m

• Performance: The exporter is serial, if your organization has a high number of issues & events you may experience Context Deadline Exceeded error during a Prometheus scrape

• Sentry API retry calls: The Sentry API limits the rate of requests to 3 per second, so the exporter retries on an HTTP exception.

  You can tweak retry settings with environment variables, though default settings should work:

  Environment variable	Value type	Default value	Purpose
  SENTRY_RETRY_TRIES	Integer	3	How many retries should be made in case of an exception
  SENTRY_RETRY_DELAY	Float	1	How many seconds to wait between retries
  SENTRY_RETRY_MAX_DELAY	Float	10	Max delay to wait between retries
  SENTRY_RETRY_BACKOFF	Float	2	Multiplier applied to delay between attempts
  SENTRY_RETRY_JITTER	Float	0.5	Extra seconds added to delay between attempts

• Use scrape_interval: 5m minimum.

This value will be defined by the number of new issues and events
higher number of events will take more time

• Use a high scrape_timeout for the exporter job

General recomendation is to set scrape_interval - 1 (i.e.: 4m)

• If the scraping of particular metrics are disabled the values above can be reduced depending on your setup.

Sentry Prometheus Exporter documentation

Contributions, issues and feature requests are welcome!

• Feel free to check issues page.

Copyright © 2021 Italo Santos.

This project is GNU General Public License v2.0 licensed.

Give a ⭐️ if this project helped you!

👤 Italo Santos

• Website: http://italosantos.com.br
• Twitter: @italux
• Github: @italux
• LinkedIn: @italosantos

This README was generated with by readme-md-generator