Tracking application errors with Sentry
We use Sentry to track errors in code running VA.gov applications. Sentry allows teams to discover, triage, and prioritize errors in real-time, reducing overall platform errors.
You’re responsible for configuring and managing your own Sentry teams, managing alerting rules, and monitoring their application’s data. These management rules give you teams more responsibility in relation to Sentry alerting and error ownership.
Current version
The current version of our Sentry instance is 21.6.1.
Before you begin
In order to access Sentry you must have configured the SOCKS proxy. Proxy setup instructions are found in the Internal Tools Documentation.
Creating a new team
Each VFS team will designate a Sentry administrator and provide the name of this person in the #vfs-platform-support Slack channel. A Platform engineer will then grant access and assign permissions to this person. The Sentry administrator will then perform the following tasks to configure their Sentry team:
Create a new team based on the naming and structuring conventions described in the Platform user management process, for example, vfs-[team name here]. If permissions need to be adjusted, ask for help from the Platform in Slack.
Add team members to the Sentry team. This can be done by navigating to (Project) Settings => Teams => Add Member.
Provide a list of ownership rules (paths and urls) that your team wants to monitor for Sentry errors. Examples can be found in the Sentry issue owners documentation. Ownership rules are used for automatic error issue assignment.
Provide the list to a Platform engineer using the Slack support workflow in #vfs-platform-support.
Issue management
Automatic issue assignment is a feature adopted in Sentry 10, and allows for errors that match a given URL path or code path to be automatically to a team or member. This feature allows developers to reduce noise by directing notifications to specific teams based on a path or URL. VFS teams are responsible for providing the Platform with a list of paths and URLs that are of concern to their team. By assigning specific alerting rules, Sentry allows teams to pass issues to the developer who can fix them, faster. By defining paths/URLs of concern to a team, issues are automatically assigned to teams.
Any adjustments to alerting rules must be communicated in the #vfs-platform-support Slack channel since “Issue Owners” is a project-based permission that Platform engineers manage.
Monitoring Slack alerts
Monitoring Sentry alerts via Slack is not currently available but it is a feature we're looking to add.
VFS Sentry team members are responsible to monitor incoming error alerts, but Platform engineers may triage platform-wide errors as needed.
Understanding dashboards and filtering
Issues can be sorted and filtered according to a user’s filtering criteria (for example, last seen or custom tags).
Sentry team filtering
Sentry users can search for issues assigned to their team via custom search. By providing the 'assigned' custom search attribute followed by their team name tag, user can filter by issues assigned to their team.
Example team assignment search: assigned:#benefits-memorial-1
Note: The hashtag needs to be present when searching by team name
Sentry users can obtain more error-related details by drilling down into individual issues for a more comprehensive debugging approach. Unresolved issues will display per project on the issue dashboard. On the project issues dashboard, specific users can be assigned to issues. You can also obtain time-based error data. An overview displaying issue trends can be viewed in the “Overview” tab.
Tagging standards
VFS engineers are responsible for complying with the Platform’s current tagging standards. A list of existing tags can be found in Sentry as a per project based setting. Tags can be utilized as a measure along with issue assignment to filter errors realted to your team.
Help and feedback
Get help from the Platform Support Team in Slack.
Submit a feature idea to the Platform.