2.8.1: Fix installation on Ruby 2.6 for libc and musl library builds. Read more about it in our changelog and issue report
1.9.0: Better Erlang error reporting, improved memory host metrics reporting on containers and improvements the diagnostics reports we use to debug any issue you may be having. Read more about it on our blog.
2.8.0: Improved memory host metrics reporting on containers, support for JRuby on Alpine Linux and improvements the diagnostics reports we use to debug any issue you may be having. Read more about it on our blog.
1.8.2: Add support for Ecto 3, fix issues loading :poison and :hackney, allow errors without backtraces in send_error/1-7. Read more about it in our changelog.
2.7.2: Fix missing linked libraries on two stage deploys. Read more about it in our changelog and issue report.
1.8.1: Fix missing linked libraries on two stage deploys. Read more about it in our changelog and issue report.
We've improved the navigation to help you focus on different tasks and to make it easier to find the data you need. Read more about these improvements on our blog.
2.7.1: Improved performance when GC instrumentation is disabled and better error message when a combination of architecture and build is not supported. Read more about it in our changelog.
1.8.0: Longer tags and doubles for custom metrics. Read all about it on our blog.
2.7.0: Longer tags, doubles for custom metrics and new working_directory_path option. Read all about it on our blog.
Add a description field to Anomaly Detection triggers that will be shown with alerts. Can be used to explain why an alert was triggered and what to do to resolve the issue.
1.7.0: Support for custom namespaces, HTTPoison dependency dropped and a bundled CA certificate. Read all about it on our blog.
It's time for the big picture! Every app now has its own live updating dashboard! Read more about it on our blog.
1.6.6: Add support for OTP 21 by using the agent's local environment instead of global environment variables. See the changelog.
1.6.5: Make sure Transactions can be started and completed without the TransactionRegistry running. See the changelog.
1.6.4: Overwrite message for Phoenix.ActionClauseError. See the changelog.
1.6.3: Remove script_name, query_string and peer from Plug transaction metadata. See the changelog.
1.6.2: Merge, instead of ignore, Phoenix's :filter_parameters if also configured in AppSignal. See the changelog.
New GDPR Data Processing Agreement form and updated Privacy Policy. Read all about it on our blog.
2.6.0 & 2.6.1: We've released a new gem version that allows you to filter more data and set the app revision in the config. Read all about it on our blog.
1.6.0 & 1.6.1: We've released a new Elixir package that allows you to filter more data and set the app revision in the config. Read all about it on our blog.
2.5.3: Recommended update for Sidekiq users: Fix unique action names issue for Sidekiq jobs. For more information see full changelog.
2.5.2: Better support Sidekiq delay extension for ActiveRecord instances. Also includes a fix for custom event formatters. For more information see full changelog.
Use GitHub App integration instead of OAuth with very broad scopes, see this blog post for more details.
1.5.0: Fixes for Plug action names, restores the revision configuration, fixes a compilation issue and fixes an issue with unsetting configuration values when hot reloading. Read all about it on our blog.
2.5.1: Improve internal sample storage in the AppSignal agent. See the full changelog.
1.4.10: Fix missing POST parameters in errors in Plug and Phoenix apps. See the changelog.
- Allow for comparing and inspecting multiple hosts in Host usage and Host metrics (blog).
- Refactor account overview to make it faster for large organisations, 24h graphs are loaded separately
2.5.0: jRuby support! Read more about it on our blog.
1.4.9: Add x-real-ip to request header whitelist and handle noise over handle_info without warnings in ErrorHandler. See the changelog.
We just split out hosts into two new screens: "Host usage" and "Host metrics". Read all about it here.
1.4.8: Fix transaction metadata for send_error, fix DataEncoder.encode error for improper lists, fix locking issue in diagnose. See the changelog.
1.4.7: Fix compile errors on Elixir 1.6. See the changelog.
1.4.6: Only reference Plug.WapperError when Plug is available, and make sure exceptions are caught after the Plug response is sent. See the changelog.
- Add email notifier for deploys
- Improve Anomaly detection alerts page
- Improve Anomaly detection notification emails
- Improve visibility of error messages when creating a Anomaly detection trigger
- Add advanced form for Anomaly detection triggers
- Update OpsGenie integration (uses OpsGenie API V2)
1.4.5: Ensure the appsignal application is loaded before running the diagnose task. See the changelog.
1.4.4: Unwrap Plug.Conn.WrapperErrors and fix Plug request ID mismatches. See the changelog.
- Add free memory to the host memory usage anomaly detection form.
- Link to specific samples from exception and performance notification e-mails.
2.4.3: Store more details for Redis events. See the full changelog.
1.4.3: Fix Dialyzer violations and fix logger error on failed installation. See the changelog.
1.4.2: Change log level from info to debug for value comparing failures and collect free memory host metric. See the changelog.
- Add throughput anomaly detection
- Add queue time anomaly detection
- Improve number rounding in graphs (Y-axis and Hover)
- Batch alert emails
2.4.2: Fix CLI crashes, collect free memory host metric for new anomaly detection trigger and other small fixes. See the full changelog.
1.4.1: Use musl for operating systems that use a libc version older than 2.12. See the changelog.
2.4.1: Sidekiq delayed extension support, Que integration added and fix the installation on older Linux systems after upgrading to 2.4.0. For more information about the Operating Systems we support, see our documentation. Also see the full changelog.
1.4.0: Add support for FreeBSD systems, fix a crash when using multiple processes and fix the DNS timeouts issue. Read all about it on our blog.
2.4.0: More Exceptions captured, FreeBSD support, DNS timeouts fixed and more. For more details about this release, please see the the blog post.
You can now have multiple notifiers and have alerts for triggered anomalies sent to all notifiers. Read more on our blog.
- Add performance form to Anomaly Detection, get notified when mean or 90th percentile rises above a certain threshold.
- You can now add multiple notifiers for the same service, e.g. 2 Slack notifiers that each notify a different channel.
Enable two-factor authentication on your user account 🔒 Read more on our blog.
Add option to select and mutate multiple incidents at once.
To keep security sensitive information safe, we now ask for a password confirmation before granting access to the security page.
1.3.4: Allow configuration of permissions of working directory, fix bug that delayed extension shutdown and log extension start with app revision if present. Read the full changelog.
2.3.6: Allow configuration of permissions of working directory, fix bug that delayed extension shutdown and log extension start with app revision if present. Read the full changelog.
1.3.3: Support new Alpine Linux docker image architecture name and don't capture Phoenix Channel payloads. Read the full changelog.
2.3.4: Fix naming for ActiveJob integration with DelayedJob. Read the full changelog.
- Allow users to unlink their GitHub account in their profiles
- Notify organization owners when we receive payloads with a timestamp in future
- Fix performance incidents table column sorting
1.3.2: Don't crash with unbound Phoenix channel payloads. Read the full changelog.
- Improve app switcher
- Always show full environment label of an app
- Update Phabricator integration
2.3.3: Improve agent connection handling, accept mixed case environment variable values for the "true" value, support Capistrano dry run option and more. Read the full changelog.
- Fix Heroku Single Sign On logging out after a few requests
- Remove limit of incidents on GraphQL endpoint, it now returns all incidents for selected deploy
- Fix exception incident table "errors in deploy" sorting when selected deploys is "all"
- Convert performance/exception incident pages to GraphQL
- Convert Action page to GraphQL
1.3.1: Save Phoenix channel parameters. Read the full changelog.
Introducing a much improved version of our Snapshot feature 🎉 Read all about it on our blog.
2.3.2: Improve Rake argument handling. Allow for more detailed view of which arguments a tasks receives. Read the full changelog.
1.3.0: Support for Plug-only Elixir apps, request header filtering, channels namespace and more. Read all about it on our blog.
2.3.1: Fix ActiveSupport::Notifications hook not supporting non-string names for events. Read the full changelog.
New graph legend design! Hover over any graph and you'll see a fresh legend design that's optimized to display more data. 📈
Improved Slack notifications are live! Read more about it on our blog!
2.3.0: ActionCable support, background job parameter filtering, ignore namespaces and more. For more details about this release, please see the the blog post.
1.2.3: Fix mapping for 32bit architectures. Read the full changelog.
Allow users to set their own timezone. All times in the AppSignal.com UI will be displayed using that timezone. You can configure yours in your profile.
1.2.2: Better backtraces for linked processes. Read the full changelog.
2.2.1: Support for Rails 5.1 was added, a bug with DataMapper query durations was fixed and improvements to the Appsignal.instrument helper. Read the full changelog.
2.2.0: Added support for custom namespaces, support Ruby 2.4, better Padrino support, fix DNS timeouts issue and fix front-end monitoring gem endpoint issues. For more details about this release, please see the the blog post.
1.2.1: Improve registry lookup performance, ErrorHandler handles errors in tuples and more. Read the full changelog.
1.2.0: Revamped Error handling; we now catch more errors, and store more information about them. Read more on our blog.
1.1.1: Add Instrumentation.Helper.instrument function, fix permission issue during installation on shared directories and internal refactoring. Read the full changelog.
1.1.0: Phoenix controller actions in error samples and support for hot reloading. Read more on our blog.
2.1.2: Fix error with Grape request methods defined with symbols. Read the full changelog.
1.0.4: Fix propagation of decorator instrumenter helper return values. Read the full changelog.
2.1.1: Fix DNS issue for certain config in resolv.conf files and fix redis and sequel environment variable config. Read the full changelog.
1.0.3: Print missing APPSIGNAL_APP_ENV during install, deactivate test environment upon install (if available) and run agent in diagnostics mode regardless of app config. Read the full changelog.
1.0.2: Fix broken installer output, broken in 1.0.1 😖. Read the full changelog.
1.0.1: Remove revision config option. Wasn't supposed to be used this way. Read the full changelog.
AppSignal for Elixir 1.0.0 released 🎉 Read all about it on our blog.
0.13.0: Send demo samples on install, make mix tasks available in releases, rename Phoenix framework events, improve Phoenix integration. Read the full changelog.
0.12.3: Fix warning about redefining a module on compilation. Read the full changelog.
Add incident numbers to incidents, change incident urls from /action_name to /incident_number (e.g. web/exceptions/blogpostController#show to web/incidents/12
0.12.2: Check installed agent version and update if using an older version. Read the full changelog.
0.12.1: Support more versions of the HTTPoison dependency. Read the full changelog.
0.12.0: Fix a DNS issue in the agent, add mix appsignal.diagnose task, change the way the APPSIGNAL_PUSH_API_KEY environment variable is used, support big ints, drop Poison dependency, and more. Read the full changelog.
You can now see the request usage per app under Profile & settings -> Billing -> Show usage.
0.11.4 & 0.11.5 & 0.11.6: Fix a segfault in C-extension. Read the full changelog. Read our post-mortem about the issue.
0.11.3: Fix support for DNS search configurations. Read the full changelog.
2.1.0: Add support for Alpine Linux 🎉 and deprecate the is_ignored_*? methods. Read the full changelog.
2.0.6: Fix Appsignal::Transaction#record_event method call used in DataMapper integration. Read the full changelog.
0.11.2: Add support for non-strings as map values in sample payload. Read the full changelog.
0.11.1: Fix phoenix dependency issue, add module to sample when using decorators. Read the full changelog.
0.11.0: No longer require Phoenix as a dependency, add "mix appsignal.demo" task, support Alpine Linux. Read the full changelog.
0.10.0: Enable host metrics by default, format stacktraces on Appsignal.Transaction.set_error/4, check if the AppSignal agent is started before sending transactions and more. Read the full changelog.
2.0.5: Improved logging for agent connection issues and handling of nil request/environments in transactions. Read the full changelog.
- Fix NaN issue on sample page when swap is empty
- Add support for custom namespaces
- Use Intercom Oauth instead of API key
- Add field on site settings to override GitHub/GitLab base url
2.0.4: Make appsignal diagnose output even more information, change default log path to ./log, load APPSIGNAL_APP_ENV consistently for all integrations, and more. Read the full changelog.
- Add support for Elixir backtrace links to GitHub
- Add namespace to Notification E-mails
0.9.2: Set APPSIGNAL_IGNORE_ACTIONS from config and send metadata in Appsignal.ErrorHandler.submit_transaction/6. Read the full changelog.
Review your active browser sessions and keep your user account secure. Read more about it on our blog
Add custom markers from any graph! Read about this new feature in our blog post.
- Custom metric screens can be viewed fullscreen and auto-refresh.
- Improved notification settings screens
- Add support for Elixir backtraces (app/full backtrace)
2.0.3: Fix JavaScript exception catcher throwing an error on finishing a transaction. Read the full changelog.
2.0.2: Fix Sequel instrumentation overriding existing Sequel logic from the logger and other extensions. Read the full changelog. More documentation about how to instrument Sequel and examples applications can be found in our documentation.
0.9.0 / 0.9.1: Various updates, and some deprecations (like the removal of instrumentation macro's and Channel related macro's). Do read the release notes.
2.0.1: Fix config load order regression for the APPSIGNAL_PUSH_API_KEY environment variable's activation behavior introduced in 2.0.0. Read the full changelog.
You can now add Custom Markers to all the graphs in AppSignal!
– When selecting a timeframe, a deploy or an issue state filter we now show the selection in the URL so you can share what you're looking at.
– Filters are now persistent everywhere, so when setting a timeframe to 1H for exceptions it will also be 1H when switching to Custom Metrics.
– Your preferred graph grid layout (full width or half width graphs) is now persisted in a cookie.
– You can now set default notification settings for Web and Background separately, instead of having to use the same for both. Also, you can set organization defaults now so newly added apps inherit your preferred settings instead of ours.
2.0.0: ActiveSupport dependency removed, performance improvements, Grape namespace support, automatic container detection, diagnose improvements and configuration load order changes. Read our blog post about this release.
0.8.0: Add experimental support for Phoenix channels
Remember original stacktrace in phoenix endpoint (🎩👌 Manuel Kallenbach!)
0.7.1: Parameters are now be sanitized
0.7.0: Send session information to AppSignal
Add parameter filtering
0.6.2: Agent download retries
Fixed a bug in where `Appsignal.Transaction.set_meta_data/3` sent the key twice instead of the key and the value. (🎩👌 Martin Schurig)
We've launched our Public Beta for Elixir! Check it out at appsignal.com/elixir.
0.6.0: Improved error extracting
You can now configure whether you'll receive e-mail notifications by default for new sites here.
1.3.6: Support blocks arguments on method instrumentation, support APPSIGNAL_APP_ENV for Sinatra, minor improvements to installer, more robust handing of non-writable log files and support for some more constructs in SQL lexing. Read the full changelog.
1.3.5: Fix SSL certificate config in appsignal-agent. Remove mounted_at Sinatra middleware option. Now detected by default. Sinatra applications with middleware loading before AppSignal's middleware would crash a request. Read the full changelog.
1.3.4: Fix argument order for record_event in the AppSignal extension. Read the full changelog.
1.3.3: Output AppSignal environment on appsignal diagnose. Fixes for Sinatra transactions that could crash. Capistrano 2 "stage" and "appsignal_env" options added. Read the full changelog.
1.3.2: Transaction discard method, fixes for some warnings. Read the full changelog.
1.3.1: Bug fix in CLI install command. Read the full changelog.
Add configuration options to "notifier" integrations, you can now select what type of notification you'd like to receive (Deploy markers, Exceptions or Performance notifications).
1.3.0: Updated Sinatra support, non-Rails parameter filtering, custom hostnames and method instrumentation. Plus, new Webmachine and DataMapper integrations. Read all about it!
1.2.5: Bugfix in CPU utilization calculation for host metrics.
Add ability to search for events in sample event timeline.
Improve sample breakdown.
Stop graph legend from moving when hovering over the legend, prevents "cursor trapping" .
Highlight graph line when overing over a legend row.
1.2.4: Support for adding a namespace when mounting Sinatra apps in Rails and for negative numbers and ILIKE in the sql lexer.
Improved graph hover.
1.2.3: Catch nil config for installer and diag, minor performance improvements and support for arrays, literal value types and function arguments in sql lexer.
1.2.2: Handle out of range numbers in queue lenght and metrics api.
1.2.1: Support for Sequel 4.35, env option to skip errors in Sinatra, fix for queue time calculation in Sidekiq and other minor fixes.
New event timeline, read about it here.
1.2.0: Host metrics beta, Shuryoken support. Read all about it!
1.1.9: Fix for race condition when creating working dir exactly at the same time, make diag Rake task resilient to missing config.
1.1.8: Require json to fix problem with using from Capistrano.
1.1.7: Logging fix when using daemons gem, support for Resque/ActiveJob, rescue more errors in json generation.
Fixed a bug where session data was not properly stored, samples between 11 May and today may miss session data.
1.1.6: Generic Rack instrumentation middleware, event formatter for Faraday and rescue and log errors in transaction complete and fetching params
Persistent filters on account overview. Filter by environment, namespace or text
1.1.5: Support for null in sql sanitization, Capistrano installation fix, minor logging and compatibility fixes.
1.1.4: Logging improvements, fix for exception with nil messages and a fix for using structs as job params in Delayed Job
1.1.3: Fix for issue where Appsignal.send_exception clears the current transaction if it is present, Rails 3 compatibility fix.
1.1.2: Bug fix in notify of deploy cli
1.1.1: Collects global metrics for GC durations (in beta, disabled by default), more reliably collects params from Delayed Job, collects params for Delayed Job and Sidekiq when using ActiveJob, adds official Grape support and comes with easier CLI installation using bundle exec appsignal install.
1.0.6: Bug fix in sql sanitizer.
1.0.5: Major improvements in sanitization and in some cases memory usage of the agent.
Ability to only give owners access to new sites by default (not team "Everyone")
More comprehensive data in the Event Tree
Syntax highlighting for SQL and MongoDB queries
1.0.4: Make working directory configurable, see the docs for details.
Graphs can now be shown in one or two columns!
Active sample in sample list is now a clear bright blue
Last deploy marker on exception sample page now also shows year
Event tree "show event" link is now located after expanding to avoid missclicks
More info
1.0.3: Fix in Resque support and in Javascript error monitoring.
The event groups percentage graph is now based on total durations instead of averages to give better insight into how groups affect the overall performance.
The Controllers and Jobs tables now have a row with aggregate values of the selected timeframe.
1.0.2: Backwards compatibility fixes and a bug fix for duplicate events in samples that can occur with forking web servers.
1.0.1: Fix for bug in gem initialization when using safe_yaml gem
1.0.0: Our new agent goes mainstream, and it brings four new features with it. Read all about it!
0.11.18: Fixes an issue with the Rake integration and refactors the Resque integration.
New: AppSignal now integrates with Phabricator.
Bugfix: At certain viewport heights the main navigation wouldn't allow scrolling. We've fixed that.
Beta: The PagerDuty integration is now in Beta, head over to the integrations tab to try it out.
0.11.17: Fixes a bug with using Rack request in `send_exception`.
0.11.16: Adds tagging for front-end error catching, and fixes an issue with ActiveAdmin sending incompatible instrumentation events.
New: Snapshots are now scoped by namespace (frontend/backend).
New: You can now turn request tags into automagic links (eg. to quickly go to a customer profile in your back-office). Find out more in our documentation on Link Templates.
New: Does your Performance overview seem off, because of the long-time averages? We now show you metrics for the current deploy by default (screenshot).
0.11.15: Improves Sinatra support.
0.11.14: Supports ActiveJob wrapped jobs, and has improved proxy and Rake support.
New: AppSignal now integrates with OpsGenie, the notification system with on-call schedules. Find out more.
New: The Actions tree now shows the 10 slowest events by default, with options to show the 50th, 75th, 80th, 90th, 95th or 99th percentile slowest (or just all events) instead (screenshot).
0.11.13: Adds Padrino, Rake task monitoring and http proxy support. Fixes bugs with sending of empty queues, retries, Capistrano 2 detection, and the Sinatra integration.
New: Today we introduce issue states, interface improvements and better notification settings. Find out more.
New: We added a user setting to turn off in-app announcements and promotions sent using Intercom (screenshot). This doesn't disable chat/one-on-one messaging.
New: The Actions tree now shows bars to indicate time spent by an event on itself, and its children (screenshot).
New: If you use Intercom, you can have every error a user encounters create an event on their activity stream. Find out more.
0.11.12: Sanitizer will no longer inspect unknown objects, because implementations of inspect sometimes trigger unexpected behavior.
Enhancement: A new design for deploy markers in the sample navigation makes it easier to distinguish them from links to individual samples (screenshot).
0.11.11: Reliably intercept errors for Sinatra in production.
We added an Upgrade Policy to explain how and when you will be asked to upgrade your plan.
0.11.10: Handle UTF-8 encoded strings that contain invalid UTF-8 characters when sending data.
Privacy is important, so we updated our Privacy Policy to include a detailed overview of all the (third party) services we use.
Some people were having issues with integrating AppSignal into their Sinatra app, because of incomplete documentation. We've fixed that.
New: Revamped Hosts section, with the ability to compare hosts and separate Web and Background.
0.11.9: Fix for a regression we found in 0.11.8 that prevented processes from exiting if there was no valid push API key.
0.11.8: Uses less memory when retrying transmissions; please upgrade. Also includes additional background job metadata, beta front-end error catching and Apache HTTP queue times improvements.
New: You can now manage AppSignal errors and performance issues in JIRA. Find out more.
New: Ability to display dates in the YYYY-MM-DD format by changing your settings.
New: Error messages can be wrapped & unwrapped, to prevent excessive horizontal scrolling.
New: We now fully support Sinatra! Find out more.
New: Show Key Performance Indicators on "jobs" and "controllers" pages instead of sparklines.
New: Ability to choose a custom time range to display results on the "jobs" and "controller" pages.
0.11.7: Adds option to override Job name for Delayed Job (see docs).
New: AppSignal now integrates with Zapier, so you can route your notifications to external apps not (yet) covered by our other integrations.
New: Fresh homepage with a better overview of features and a more comprehensive pricing page, as well as the introduction of new 25M, 250M and 1B plans.
0.11.6: Uses APPSIGNAL_APP_NAME and APPSIGNAL_ACTIVE environment variables in config. Adds better Sinatra support by using routes as action and setting session data.
New: Date & number formatting is now based on your locale. The default setting can be overridden in your personal settings.
0.11.5: Adds support for the Sequel gem.
New: The app now shows deploy markers in all graphs. Hovering over them display info about when and who, clicking them takes you to a diff on GitHub or GitLab (if you have those integrations enabled).
0.11.4: Makes without_instrumentation thread safe.
New: The app now shows deploy markers (gray lines) in the list of saved samples.
Enhancement: You won't be notified about slow requests during the first 5 minutes after a deploy, to prevent false positives due to Rails restarting.
0.11.3: Adds support for Ruby 1.9+ (instead of 1.9.3+).
Bugfix: The list of saved samples is now limited to samples within the trace retention period for a plan, preventing huge lists and browser crashes.
Enhancement: Email notifications any error are now limited to once per 5 minutes, to prevent flooding mailboxes with a gazillion emails.
New: AppSignal now fully integrates with GitLab, for issue creation as well as linking to diffs, etc.
0.11.2: If the APP_REVISION environment variable is set, it's now include in the log entry.
