Migration Guide
Learn about how to migrate from the deprecated sentry-raven SDK.
Unified Interfaces With Other SDKs: The design of
sentry-raven
is outdated compared with our modern Sentry SDKs. If you also use other Sentry SDKs, such as Sentry's JavaScript SDK for your frontend application, you'll notice that their interfaces are quite different from the one provided forsentry-raven
. The newsentry-ruby
SDK provides a more consistent user experience across all different platforms.Tracing: The Sentry Ruby SDK includes tracing, which you can enable if you haven't already as (discussed here).
Future Support:
sentry-raven
has entered maintenance mode, which means it won't receive any new feature supports or aggressive bug fixes.Better Extensibility: Unlike
sentry-raven
,sentry-ruby
is built with extensibility in mind and will allow the community to build extensions for different integrations/features.
Ruby 2.3 and Rails 4 are no longer supported.
Sentry's Ruby SDK still supports integration with Rack
by providing built-in middleware, but you'll need to install gems for integrations with Rails
, sidekiq
, delayed_job
, and other libraries.
Currently available integrations are:
In sentry-raven
we have different processor classes for data scrubbing. But updated Sentry Ruby SDK doesn't support these processors. Instead, to protect users' sensitive data, the Ruby SDK adds a new configuration option of send_default_pii
. When the value is set to false
(default), sensitive information, such as
- user ip
- user cookie
- request body
- query parameters
will not be sent to Sentry.
You can re-enable it by setting:
config.send_default_pii = true
As for scrubbing sensitive data, please use Sentry's Advanced Data Scrubbing feature.
Senty's Ruby SDK uses a unified structure, which introduces two new components: Hub
and Scope
(which are both documented here). Most users won't interact with Hub
directly, but will need Scope
to configure contextual data.
In sentry-raven
, we provided helpers like Raven.user_context
for setting contextual data. In our updated Ruby SDK, those helpers were removed, and you'll need to use a different approach for setting those data like:
Configure data globally
# Before
Raven.user_context(id: 1)
# After
Sentry.set_user(id: 1)
Configure data in a local scope
# Before
Raven.tags_context(foo: "bar") do
Raven.capture_message("test")
end
# After
Sentry.with_scope do |scope|
scope.set_tags(foo: "bar")
Sentry.capture_message("test")
end
config.sanitize_credit_cards
config.sanitize_fields
config.sanitize_fields_excluded
config.sanitize_http_headers
config.scheme
config.secret_key
config.server
config.tags
config.logger
config.encoding
config.silence_ready
# please only use config.before_send
config.should_capture
config.transport_failure_callback
config.current_environment #=> config.environment
config.environments #=> config.enabled_environments
config.rails_report_rescued_exceptions #=> config.rails.report_rescued_exceptions with sentry-rails installed
config.ssl #=> config.transport.ssl
config.ssl_ca_file #=> config.transport.ssl_ca_file
config.ssl_verification #=> config.transport.ssl_verification
config.timeout #=> config.transport.timeout
config.open_timeout #=> config.transport.open_timeout
config.proxy #=> config.transport.proxy
# These options are present in sentry-ruby 4.* but were removed in 5.0 with the faraday removal
config.http_adapter #=> config.transport.http_adapter
config.faraday_builder #=> config.transport.faraday_builder
This section only lists a few important additions. See the full list of configuration options here
# this behaves similar to the old config.scheme option
config.transport.transport_class = MyTransportClass
# sentry-ruby sends events asynchronously with its own background workers
# the default number of workers equals to your machine's processor count
# you can adjust the number with
config.background_worker_threads = 10
# to send events synchronously like sentry-raven does, set it to 0
config.background_worker_threads = 0
In this section, we provide code examples to guide you through the changes required for the migration.
Installation
Old:
gem "sentry-raven"
New:
gem "sentry-ruby"
# and the integrations you need
gem "sentry-rails"
gem "sentry-sidekiq"
gem "sentry-delayed_job"
gem "sentry-resque"
Configuration
Old:
Raven.configure do |config|
config.dsn = "DSN"
end
New:
Sentry.init do |config|
config.dsn = "DSN"
end
Set Contextual Data (global)
Old:
Raven.user_context(id: 1)
Raven.context.tags = { foo: "bar" }
Raven.context.extra = { debug: true }
New:
Sentry.set_user(id: 1)
Sentry.set_tags(foo: "bar")
Sentry.set_extras(debug: true)
Set Contextual Data (local)
Old:
Raven.user_context(id: 1) do
# send event
end
Raven.tag_context(foo: "bar") do
# send event
end
Raven.extra_context(debug: true) do
# send event
end
New:
Sentry.with_scope do |scope|
scope.set_user(id: 1)
scope.set_tags(foo: "bar")
scope.set_extras(debug: true)
# send event
end
Manual Message/Exception Capturing
Old:
Raven.capture_message("test", extra: { debug: true })
New:
Sentry.capture_message("test", extra: { debug: true })
The options for these methods are also changed. Currently available options are:
contexts
tags
extra
user
level
fingerprint
backtrace
To set transaction_name
(transaction
in sentry-raven
) of the event, please use
Sentry.get_current_scope.set_transaction_name("NewTransaction")
sentry-ruby
doesn't capture raven_context
from exceptions anymore. However, you can use before_send
to replicate the same behavior:
config.before_send = lambda do |event, hint|
if exception = hint[:exception]
exception.raven_context.each do |key, value|
# here I assume the event would be a Sentry::Event object
# however, it'll be a hash if you use the async callback (which will be removed in version 6.0)
event.send("#{key}=", value)
end
end
event
end
- Rails example
- Sinatra example
- Sidekiq examples:
A community maintained fork of sentry-ruby
that works with older Ruby versions (2.1, 2.2 and 2.3) is available here. You can use this in cases where it is impossible for you to upgrade but please note that we don't officially support it.
Our documentation is open source and available on GitHub. Your contributions are welcome, whether fixing a typo (drat!) or suggesting an update ("yeah, this would be better").