Error Monitoring

Installation

Before you can start using Honeybadger to squash bugs, you will need to install our library into your app. Installation usually boils down to:

  1. Installing the Honeybadger library
  2. Setting the API key
  3. Enabling error reporting

The details vary a little depending on language and platform. But you can always find the correct installation instructions for YOUR app on the "Project Settings" page.

In case you are wondering, we officially support:

Ruby, client-side JavaScript, Vue, Elixir, Go, NodeJS, Java, Python, PHP, Clojure, and Cocoa

Check out any of these pages to see how to install Honeybadger for your app. Once our client library is installed and configured, errors thrown by your application are automatically sent to our API.

Error Grouping

Honeybadger groups similar exceptions together using rules which we've found to work the best in most cases. The default information we use to group errors is:

  1. The file name, method name, and line number of the error's location
  2. The class name of the error
  3. The component/controller name

We use this information to construct a "fingerprint" of the exception. Exceptions with the same fingerprint are treated as the same error in Honeybadger.

You can customize the grouping for each exception by changing the error class name, component, or stack trace—or by sending a custom fingerprint.

Anatomy of an Error

When your app reports an error to us, we make the details available on a web page that looks like this:

Error Detail Page Overview

Yeah, it looks a little complicated. But once you understand what all the pieces are, you will see it is quite simple.

Let's inspect this page, piece by piece.

Honeybadger groups identical errors together. Using our navigation bar, you can still navigate between each individual occurrence (or notice, as we also call them) of the error.


Resolve / Unresolve

When an error is marked "unresolved" we will not send you any additional notifications when it re-occurs. Errors are auto-resolved on deploy by default, but that is configurable.

Resolved and Unresolved

Error Actions

Here you can perform many of the actions available to your error. Also, as you set up integrations for the project, any that can link back to the error (say, an issue tracker like GitHub) can be managed here.

Error actions

Assign an Error

Using this button, you can assign the error currently viewed to any user in the project. This can later be used to search by assignment, or, with assignment notifications enabled, to inform a user that a new error has been assigned to them.

Pause or Ignore an Error

One of the more common actions used is to ignore the occurrences of an error.

Pausing

You can choose to "Pause" for a time duration or occurrence count. We will still record the occurrences, we just will not notify you.

Ignoring

When you "ignore" an error we will stop recording altogether, so it will not go against your quota.


Resolve on Deploy

By default, Honeybadger resolves all the errors in your project when you report a deployment, causing new alerts to be sent for any of the errors that re-occur. You can turn off this behavior by unchecking the "Resolve on deploy" checkbox in Project Settings. When turned off, there is a new action in the actions area of the error detail page, allowing you to resolve individual errors on the next deploy.

Error actions resolve on deploy

Unsubscribe

Unsubscribing from an error disables it from sending notifications - but only for the user currently viewing the error. To silence it for all users, use the pause action.

Export an Error

Clicking the Export button will queue a process that will export the data for all the occurrences of the error. You will receive an email within minutes that will allow you to download a JSON file containing the data. This file is a newline-delimited JSON file, which means each line is a valid JSON object. Each of those JSON objects has the data (parameters, context, etc.) for a single occurrence of the error.

Merge Errors

Using the merge action, you can combine one error with another error. This takes all the existing notices from the "donor" error and adds them to the "receiver" error's notice history.

Share URL

Sharing the URL of the error will give you a link to a webpage displaying the error. This webpage does not allow for any actions, comments, or the ability to resolve the error.

Delete

This removes the error and all of the notice history for that error. This is especially useful if your error information happens to accidentally include sensitive information. New instances of the error will appear again on a new error page.

Action Integrations

GitHub and other project management integrations will add another action button - in GitHub's case the button is labeled "Create issue". Other integrations will behave similarly.

Anatomy of an Error, Continued

Each error page contains a substantial amount of detail related to the error itself and a history of all the actions done to the error. You can see how many times it was resolved or merged and who performed each action.

From the top of the error page, you can see all of the categories and can click to any of them. There's even a keyboard shortcut for each category.

Error-categories

Notices

Notices are the individual error events that are sent from your app. We also refer to them as occurrences as well.

The Notices section allows you to see the distribution of occurrences over time. You can constrain the dates and filter using the same search criteria as our error search. When you view an error after doing a search, the search criteria will be applied to the notice list.

Notices

Comments

Communicate with team members, or record notes to retain context around an error. Your comments can be formatted in GitHub-flavored markdown.

Comments

Backtrace

Each line in the backtrace links to the GitHub or Bitbucket repo. You can even click a button to open the file in your local editor.

Backtrace

Request Params, Cookies, ENV, etc.

When an error occurs during a web request, we record all the relevant debug info like params, cookies, the session, etc.

Params

Context Data

Our clients support providing a special set of data called context. This data is provided by you to help with debugging. You can use our search functionality to find errors with specific context.

user_id and user_email are special context keys that we use to help with looking up users or sending emails (as you can see below.)

Context

Browsing errors

Once you start collecting errors, it can be difficult to keep track of them. To make it easier, we've built a great interface for browsing and searching errors.

Navigating Errors

With advanced search, you can search by any field of params, environment, cookies or the session. Check out our dedicated search guide for more info.

Batch actions

You can use the Bulk Update dropdown to update multiple errors simultaneously. You can apply actions to all search results (this is the default), or you can use the checkboxes in the error list to apply the actions to selected errors.

  1. Use the search to select which errors you'd like change.
  2. Click on "Actions" and select one or more options from the drop-down.

Image of actions drop-down

Merging by batch

You can merge two or more errors to create a single error. There are two ways to merge errors:

  1. Click the Merge button in the Actions panel on the error detail page, as described above.
  2. Select two or more errors to merge from the error index page using the checkboxes and use the "Merge to" action in the Bulk Update dropdown.

Note: The Bulk Update method can merge up to ten thousand error occurrences simultaneously. You can perform multiple updates if necessary.

Local Edits

One of the advanced features in Honeybadger is configuring your local editor settings so that we can display links to open files locally when showing you an error backtrace:

To configure your editor, visit the Local Editor tab under My Settings. Select the editor you use (if you use a different editor that supports opening files via a custom protocol, let us know about it). Next, enter the absolute path to each project you wish to enable local editor links for and click "Save" at the bottom. Re-visit your error pages - they should now have the "Open in editor" links!

Most editors work without any additional setup, but Sublime and Visual Studio Code don't ship with a custom protocol by default. In order to make them work, you'll need to install one of these libraries:

If you're using Atom, make sure you're on version 1.23 or later for this to work.

Disclaimer: We do not endorse nor support these libraries; they are 3rd party open source software.

Content Security Policy Reports

If you use Content Security Policy (CSP) headers to help mitigate XSS attacks, you can use Honeybadger to track policy violations. You can configure the required headers via code when using Rails or Sinatra, or you can configure your web server or framework to emit the headers. Either way, configuring CSP reporting requires specifying the report-uri directive:

Content-Security-Policy: ...; report-uri https://api.honeybadger.io/v1/browser/csp?api_key=Your project API key&env=production Content-Security-Policy-Report-Only: ...; report-uri https://api.honeybadger.io/v1/browser/csp?api_key=Your project API key&report_only=true&env=production

The env and report_only parameters are optional. When report_only is true, CSP report payloads will show up as "CSP Report" in our UI; otherwise, the label will be "CSP Error".

Since CSP violations can be noisy, we strongly suggest you track them in a separate Honeybadger project.