All notable changes to the LaunchDarkly Ruby SDK will be documented in this file. This project adheres to Semantic Versioning.
8.12.2 (2026-03-02)
8.12.1 (2026-02-23)
8.12.0 (2026-01-30)
- Support FDv2 through the configs new data_system_config option (368aa3c)
- Update feature store interface doc to align with existing code that expects symbols not strings (595f91a)
8.11.3 (2026-01-20)
8.11.2 (2025-12-05)
- Fix diagnostic logging for connection results in stream (#343) (d49eaa8)
- Prevent command injection in FileDataSourceImpl (#341) (9ca4b98)
8.11.1 (2025-10-10)
8.11.0 (2025-07-15)
8.10.2 (2025-07-14)
8.10.1 (2025-06-13)
8.10.0 (2025-04-08)
8.9.0 (2025-03-13)
8.8.4 (2025-03-10)
8.8.3 (2025-02-24)
8.8.2 (2024-11-06)
- Allow gem to be installed from source (#305) (76d06bc)
- Improve DataKind eql? to consider object type (#304) (5893245)
8.8.1 (2024-10-30)
8.8.0 (2024-10-16)
8.7.1 (2024-09-06)
8.7.0 (2024-07-25)
This release introduces the ability to enable compression of event payloads. When enabled, the SDK will compress events before sending them to the LaunchDarkly servers. This can reduce the bandwidth required to send events, which can be useful in high-traffic environments to reduce egress traffic costs.
Important
Relay Proxy users MUST upgrade to version 8.9 or higher prior to enabling this option to prevent loss of event data.
However, enabling this feature is NOT required when using the Relay Proxy as it will manage compression automatically.
8.6.0 (2024-06-26)
8.5.0 (2024-06-10)
- Increment flag & segment versions when reloading from file data source (#285) (7d5b051)
- Log warning if client init timeout is considered high (#278) (61f4c7e)
8.4.2 (2024-05-03)
8.4.1 (2024-04-29)
8.4.0 (2024-04-05)
This release introduces a Hooks API. Hooks are collections of user-defined callbacks that are executed by the SDK at various points of interest. You can use them to augment the SDK with metrics or tracing.
8.3.1 (2024-03-28)
- Ensure Rails.logger exists before using it (#258) (2f0f7ed)
- Remove invalid prereq
check_variation_rangecheck (#261) (#263) (74ca206), closes #260
8.3.0 (2024-03-14)
- Inline contexts for all evaluation events (#245) (bd30c3b)
- Redact anonymous attributes within feature events (#246) (640ac39)
8.2.0 (2024-02-20)
8.1.0 (2024-01-03)
- Add hash access and equality support to LDContext (#241) (1ba6bde)
- Define equality methods for LDContext and Reference (#232) (1b2ba7b)
7.3.2 (2023-12-26)
7.3.1 (2023-12-26)
The latest version of this SDK supports the ability to manage migrations or modernizations, using migration flags. You might use this functionality if you are optimizing queries, upgrading to new tech stacks, migrating from one database to another, or other similar technology changes. Migration flags are part of LaunchDarkly's Early Access Program. This feature is available to all LaunchDarkly customers but may undergo additional changes before it is finalized.
For detailed information about this version, refer to the list below. For information on how to upgrade from the previous version, read the migration guide.
- A new
Migratortype which provides an out-of-the-box configurable migration framework. - For more advanced use cases, added new
migration_variationandtrack_migration_opmethods onLDClient.
- Ruby 2.7 support was removed.
- The legacy user format for contexts is no longer supported. To learn more, read the Contexts documentation.
- Previously deprecated config options
user_keys_capacity,user_keys_flush_interval,private_attribute_names,default_user_keys_capacity,user_cache_size,user_cache_time, anddefault_user_keys_flush_intervalhave been removed. - Previously deprecated test data flag builder methods
variation_for_all_users,value_for_all_users, andclear_user_targetshave been removed.
- The documentation for the
default_connect_timeoutsetting stated it defaulted to 10 seconds when the code actually defaults to 2. (Thanks, zmagg-figma!)
- Creating an
LDContextusing the legacy user format has been deprecated and will be removed in the next major release. To learn more, read the Contexts documentation.
- You can monitor the status of the SDK's data source (which normally means the streaming connection to the LaunchDarkly service) with
LaunchDarkly::LDClient.data_source_status_provider. This allows you to check the current connection status, and to be notified if this status changes. - You can monitor the status of a data store with
LaunchDarkly::LDClient.data_store_status_provider. This allows you to check whether updates are succeeding and to be notified if this status changes. - You can tell the SDK to notify you whenever a feature flag's configuration has changed (either in general, or in terms of its result for a specific context), using
LaunchDarkly::LDClient.flag_tracker.
- Support for Payload Filtering in streaming and polling modes. Payload Filtering is a beta feature that allows SDKs to download a subset of environment data, rather than full environments.
- Added flag key to log message to ease debugging. (Thanks, matt-dutchie!)
- Updated underlying event source library to address issue with
Content-Typeheader detection in some customer environments.
- Fixed JSON serialization error on internal models.
- Improved logging of feature flag data validation errors so that they are logged once at the time the SDK receives the data, rather than during each evaluation of the flag.
- Removed a misleading error message about a missing attribute that was being logged when a flag rule used the "is in segment" operator.
The latest version of this SDK supports LaunchDarkly's new custom contexts feature. Contexts are an evolution of a previously-existing concept, "users." Contexts let you create targeting rules for feature flags based on a variety of different information, including attributes pertaining to users, organizations, devices, and more. You can even combine contexts to create "multi-contexts."
For detailed information about this version, please refer to the list below. For information on how to upgrade from the previous version, please read the migration guide.
- The type
LaunchDarkly::LDContextdefines the new context model. - All SDK methods that took a hash representing the user now also accept an
LDContext.
- The
secondaryattribute which existed in the user hash is no longer a supported feature. If you set an attribute with that name inLDContext, it will simply be a custom attribute like any other. - Analytics event data now uses a new JSON schema due to differences between the context model and the old user model.
- The minimum language version is now Ruby 2.7, or jRuby 9.4.
- Several optimizations within the flag evaluation logic have improved the performance of evaluations. For instance, target lists are now stored internally as sets for faster matching.
- Removed support for the
secondarymeta-attribute in the user hash. - The
aliasmethod no longer exists because alias events are not needed in the new context model. - The
inline_users_in_eventsoption no longer exists because it is not relevant in the new context model. - Removed all types and options that were deprecated as of the most recent 6.x release.
- Config options
user_keys_capacityanduser_keys_flush_intervalare being replaced withcontext_keys_capacityandcontext_keys_flush_interval. - Config constants
default_user_keys_capacityanddefault_user_keys_flush_intervalare being replaced withdefault_context_keys_capacityanddefault_context_keys_flush_interval.
- New
Configpropertyapplication_info, for configuration of application metadata that may be used in LaunchDarkly analytics or other product features. This does not affect feature flag evaluations.
- The SDK now produces fewer short-lived objects as a side effect of flag evaluations, causing less work for the garbage collector in applications that evaluate flags very frequently. This change applies to all flag evaluations, regardless of whether analytics events are enabled.
- Miscellaneous improvements to memory usage in analytics event processing: the SDK now allocates somewhat fewer short-lived objects when computing the analytics data for flag evaluations. This does not affect baseline memory usage by the SDK, but could somewhat reduce the need for garbage collection over an application's lifetime.
- The decrease in allocation of short-lived objects is much more significant if analytics events are completely disabled. Previously, the SDK created and then discarded event-related objects in this case even though they were not being used.
- Improved efficiency of SSE parsing to reduce transient memory/CPU usage spikes when receiving flag/segment data for a large LaunchDarkly environment. (Thanks, sq-square!)
- Add
initial_reconnect_delayoption to config which controls the initial delay for reconnecting the streaming connection. - CI builds now include a cross-platform test suite implemented in https://github.com/launchdarkly/sdk-test-harness. This covers many test cases that are also implemented in unit tests, but may be extended in the future to ensure consistent behavior across SDKs in other areas.
- The
HTTP_PROXY/HTTPS_PROXYenvironment variables are now correctly applied for all requests to LaunchDarkly. Previously, this setting worked for streaming flag requests, but did not work for analytics event delivery (or flag polling), causing the latter to be attempted without using the proxy. - Rules targeting
secondaryattribute on user will now reference the correct value. all_flags_statewill return invalid flag state if the store hasn't initialized properly.- When using
all_flags_stateto produce bootstrap data for the JavaScript SDK, the Ruby SDK was not returning the correct metadata for evaluations that involved an experiment. As a result, the analytics events produced by the JavaScript SDK did not correctly reflect experimentation results. - The info level message logged when using DynamoDB now correctly identifies the feature store description. (#195)
- Providing a configuration hash when instantiating a persistent store is now optional.
- Fixed a bug that could cause a streaming connection to fail intermittently if the feature flag data contained UTF-8 characters outside of the ASCII character set. This would happen if a multi-byte character happened to be split across two successive reads from the stream, so the chances of it happening varied according to how often international characters appeared in the data and how much buffering of reads was done by the OS.
- In JRuby only, stream reconnections would fail if the application explicitly set the initial reconnect delay to zero.
- The SDK now supports evaluation of Big Segments. See: https://docs.launchdarkly.com/home/users/big-segments
LaunchDarkly::Integrations::TestDatais a new way to inject feature flag data programmatically into the SDK for testing—either with fixed values for each flag, or with targets and/or rules that can return different values for different users. UnlikeFileData, this mechanism does not use any external resources, only the data that your test code has provided.
- To use the file data source feature, the preferred entry point is now
LaunchDarkly::Integrations::FileData.data_sourcerather thanLaunchDarkly::FileDataSource.factory. This makes the Ruby SDK more consistent with other SDKs, grouping together all of the optional "connecting the SDK to something else" features underIntegrations, and using the method namedata_sourcefor consistency with theConfigproperty that will receive the object.
LaunchDarkly::FileDataSource(see above).
- Fixed a bug that caused unnecessarily heavy CPU usage when receiving very large sets of flag data from LaunchDarkly.
- Improved the speed of making the initial streaming connection to LaunchDarkly. The delay that happens before reconnecting after a connection failure was mistakenly being applied before the first connection.
- The dependency version constraint for the
httpgem is now looser: it allows 5.x versions as well as 4.x. The breaking changes inhttpv5.0.0 do not affect the SDK. (#184) - The dependency version constraint for the
jsongem is also looser: it allows any 2.x version that is higher than the SDK's minimum dependency version, not just 2.3. (#184) - The project's build now uses v2.2.10 of
bundlerdue to known vulnerabilities in other versions.
- Diagnostic events did not properly set the
usingProxyattribute when a proxy was configured with theHTTPS_PROXYenvironment variable. (#182)
- Enabling debug logging in polling mode could cause polling to fail with a
NameError. (Thanks, mmurphy-notarize!)
- If
variationorvariation_detailis called with a user object that has nokey(an invalid condition that will always result in the default value being returned), the SDK now logs awarn-level message to alert you to this incorrect usage. This makes the Ruby SDK's logging behavior consistent with the other server-side LaunchDarkly SDKs. (#177)
- The SDK now supports the ability to control the proportion of traffic allocation to an experiment. This works in conjunction with a new platform feature now available to early access customers.
- Calling
variationwith a nil user parameter is invalid, causing the SDK to log an error and return a fallback value, but the SDK was still sending an analytics event for this. An event without a user is meaningless and can't be processed by LaunchDarkly. This is now fixed so the SDK will not send one.
- Added the
aliasmethod. This can be used to associate two user objects for analytics purposes by generating an alias event.
- Added a
socket_factoryconfiguration option which can be used for socket creation by the HTTP client if provided. The value ofsocket_factorymust be an object providing anopen(uri, timeout)method and returning a connected socket.
- Switched to the
httpgem instead ofsocketry(with a custom http client) for streaming, and instead ofNet::HTTPfor polling / events. - User keys are required to be a string type. The SDK will no longer automatically perform this conversion.
- Dropped support for Ruby < version 2.5
- Dropped support for JRuby < version 9.2
- Switched the default polling domain from
app.launchdarkly.comtosdk.launchdarkly.com.
- Fixed a warning within the Redis integration when run with version 4.3 or later of the
redisgem. (Thanks, emancu!)
- Updated
jsongem to patch CVE-2020-10663.
- In
LaunchDarkly::Integrations::Redis::new_feature_store, if you pass in an externally createdpool, you can now set the new optionpool_shutdown_on_closetofalseto indicate that the SDK should not shut down this pool if the SDK is shut down. The default behavior, as before, is that it will be shut down. (Thanks, jacobthemyth!)
- Setting a user's
customproperty explicitly tonil, rather than omitting it entirely or setting it to an empty hash, would cause the SDK to log an error and drop the current batch of analytics events. Now, it will be treated the same as an empty hash. (#147)
- Previously, installing the SDK in an environment that did not have
opensslwould cause a failure at build time. The SDK still requiresopensslat runtime, but this check has been removed because it caused therakeproblem mentioned below, and becauseopensslis normally bundled in modern Ruby versions.
- The
LDClientconstructor will fail immediately with a descriptiveArgumentErrorif you provide anilSDK key in a configuration that requires an SDK key (that is, a configuration that will require communicating with LaunchDarkly services). Previously, it would still fail, but without a clear error message. You are still allowed to omit the SDK key in an offline configuration. (#154) - Removed a hidden dependency on
rakewhich could cause your build to fail if you had a dependency on this SDK and you did not haverakeinstalled. (#155) - Previously a clause in a feature flag rule that used a string operator (such as "starts with") or a numeric operator (such as "greater than") could cause evaluation of the flag to completely fail and return a default value if the value on the right-hand side of the expression did not have the right data type-- for instance, "greater than" with a string value. The LaunchDarkly dashboard does not allow creation of such a rule, but it might be possible to do so via the REST API; the correct behavior of the SDK is to simply treat the expression as a non-match.
- Fixed a bug in the 5.7.0 and 5.7.1 releases that caused analytics events not to be sent unless diagnostic events were explicitly disabled. This also caused an error to be logged:
undefined method started?.
- The backoff delay logic for reconnecting after a stream failure was broken so that if a failure occurred after a stream had been active for at least 60 seconds, retries would use no delay, potentially causing a flood of requests and a spike in CPU usage. This bug was introduced in version 5.5.0 of the SDK.
- The SDK now periodically sends diagnostic data to LaunchDarkly, describing the version and configuration of the SDK, the architecture and version of the runtime platform, and performance statistics. No credentials, hostnames, or other identifiable values are included. This behavior can be disabled with
Config.diagnostic_opt_outor configured withConfig.diagnostic_recording_interval. - New
Configpropertieswrapper_nameandwrapper_versionallow a library that uses the Ruby SDK to identify itself for usage data if desired.
- Removed an unused dependency on
rake.
- The SDK now specifies a uniquely identifiable request header when sending events to LaunchDarkly to ensure that events are only processed once, even if the SDK sends them two times due to a failed initial attempt.
- In rare circumstances (depending on the exact data in the flag configuration, the flag's salt value, and the user properties), a percentage rollout could fail and return a default value, logging the error "Data inconsistency in feature flag ... variation/rollout object with no variation or rollout". This would happen if the user's hashed value fell exactly at the end of the last "bucket" (the last variation defined in the rollout). This has been fixed so that the user will get the last variation.
- Added support for upcoming LaunchDarkly experimentation features. See
LDClient.track().
- Under conditions where analytics events are being generated at an extremely high rate (for instance, if an application is evaluating a flag repeatedly in a tight loop on many threads), it was possible for the internal event processing logic to fall behind on processing the events, causing them to use more and more memory. The logic has been changed to drop events if necessary so that besides the existing limit on the number of events waiting to be sent to LaunchDarkly (
config.capacity), the same limit also applies on the number of events that are waiting to be processed by the worker thread that decides whether or not to send them to LaunchDarkly. If that limit is exceeded, this warning message will be logged once: "Events are being produced faster than they can be processed; some events will be dropped". Under normal conditions this should never happen; this change is meant to avoid a concurrency bottleneck in applications that are already so busy that thread starvation is likely.
FileDataSourcewas usingYAML.load, which has a known security vulnerability. This has been changed to useYAML.safe_load, which will refuse to parse any files that contain the!directives used in this type of attack. This issue does not affect any applications that do not useFileDataSource(which is meant for testing purposes, not production use). (#139)
This release was an error; it is identical to 5.5.9.
- Due to the gem name no longer being the same as the
requirename, Bundler autoloading was no longer working in versions 5.5.7 and 5.5.8 of the SDK. This has been fixed. (Thanks, tonyta!)
- In streaming mode, depending on the Ruby version, calling
closeon the client could cause a misleading warning message in the log, such asUnexpected error from event source: #<IOError: stream closed in another thread>. (#135)
- Changed the gem name from
ldclient-rbtolaunchdarkly-server-sdk.
There are no other changes in this release. Substituting ldclient-rb version 5.5.6 with launchdarkly-server-sdk version 5.5.7 will not affect functionality.
- CI tests now include Ruby 2.6.x.
- Running the SDK unit tests is now simpler, as the database integrations can be skipped. See
CONTRIBUTING.md.
The LaunchDarkly SDK repositories are being renamed for consistency. This repository is now ruby-server-sdk rather than ruby-client.
The gem name will also change. In the 5.5.6 release, it is still ldclient-rb; in all future releases, it will be launchdarkly-server-sdk. No further updates to the ldclient-rb gem will be published after this release.
- Setting user attributes to non-string values when a string was expected would cause analytics events not to be processed. Also, in the case of the
secondaryattribute, this could cause evaluations to fail for a flag with a percentage rollout. The SDK will now convert attribute values to strings as needed. (#131)
- Fixed a missing
requirethat could sometimes cause aNameErrorto be thrown when starting the client, depending on what other gems were installed. This bug was introduced in version 5.5.3. (#129) - When an analytics event was generated for a feature flag because it is a prerequisite for another flag that was evaluated, the user data was being omitted from the event. (#128)
- If
trackoridentifyis called without a user, the SDK now logs a warning, and does not send an analytics event to LaunchDarkly (since it would not be processed without a user). - Added a link from the SDK readme to the guide regarding the client initialization.
- The SDK previously used the
faradayandnet-http-persistentgems for all HTTP requests other than streaming connections. Sincefaradaylacks a stable version and has a known issue with character encoding, andnet-http-persistentis no longer maintained, these have both been removed. This should not affect any SDK functionality.
- The SDK was not usable in Windows because of
net-http-persistent. That gem has been removed. - When running in Windows, the event-processing thread threw a
RangeErrordue to a difference in the Windows implementation ofconcurrent-ruby. This has been fixed. - Windows incompatibilities were undetected before because we were not running a Windows CI job. We are now testing on Windows with Ruby 2.5.
- Like 5.5.1, this release contains only documentation fixes. Implementation classes that are not part of the supported API are now hidden from the generated documentation.
- Fixed several documentation comments that had the wrong parameter names. There are no other changes in this release; it's only to correct the documentation.
- It is now possible to use Consul or DynamoDB as a persistent feature store, similar to the existing Redis integration. See the
LaunchDarkly::Integrations::ConsulandLaunchDarkly::Integrations::DynamoDBmodules, and the reference guide Persistent data stores. - There is now a
LaunchDarkly::Integrations::Redismodule, which is the preferred method for creating a Redis feature store. - All of the database feature stores now support local caching not only for individual feature flag queries, but also for
all_flags_state. - The
Configpropertydata_sourceis the new name forupdate_processorandupdate_processor_factory.
- The implementation of the SSE protocol for streaming has been moved into a separate gem,
ld-eventsource. This has no effect on streaming functionality.
- Added or corrected a large number of documentation comments. All API classes and methods are now documented, and internal implementation details have been hidden from the documentation. You can view the latest documentation on RubyDoc.
- Fixed a problem in the Redis feature store that would only happen under unlikely circumstances: trying to evaluate a flag when the LaunchDarkly client had not yet been fully initialized and the store did not yet have data in it, and then trying again when the client was still not ready but the store did have data (presumably put there by another process). Previously, the second attempt would fail.
- In polling mode, the SDK did not correctly handle non-ASCII Unicode characters in feature flag data. (#90)
RedisFeatureStore.new. This implementation class may be changed or moved in the future; useLaunchDarkly::Integrations::Redis::new_feature_store.Config.update_processorandConfig.update_processor_factory; useConfig.data_source.
- The SDK is now compatible with
net-http-persistent3.x. (Thanks, CodingAnarchy!)
- Fixed overly specific dependency versions of
concurrent-rubyandsemantic. (#115) - Removed obsolete dependencies on
hashdiffandthread_safe.
- Fixed a
LoadErrorinfile_data_source.rb, which was added in 5.4.0. (Thanks, kbarrette!)
- It is now possible to inject feature flags into the client from local JSON or YAML files, replacing the normal LaunchDarkly connection. This would typically be for testing purposes. See
file_data_source.rb.
- When shutting down an
LDClient, if in polling mode, the client was usingThread.raiseto make the polling thread stop sleeping.Thread.raisecan cause unpredictable behavior in a worker thread, so it is no longer used.
- The
all_flags_statemethod now accepts a new option,details_only_for_tracked_flags, which reduces the size of the JSON representation of the flag state by omitting some metadata. Specifically, it omits any data that is normally used for generating detailed evaluation events if a flag does not have event tracking or debugging turned on.
- JSON data from
all_flags_stateis now slightly smaller even if you do not use the new option described above, because it omits the flag property for event tracking unless that property is true.
- The new
LDClientmethodvariation_detailallows you to evaluate a feature flag (using the same parameters as you would forvariation) and receive more information about how the value was calculated. This information is returned in anEvaluationDetailobject, which contains both the result value and a "reason" object which will tell you, for instance, if the user was individually targeted for the flag or was matched by one of the flag's rules, or if the flag returned the default value due to an error.
- Evaluating a prerequisite feature flag did not produce an analytics event if the prerequisite flag was off.
- The new
LDClientmethodall_flags_state()should be used instead ofall_flags()if you are passing flag data to the front end for use with the JavaScript SDK. It preserves some flag metadata that the front end requires in order to send analytics events correctly. Versions 2.5.0 and above of the JavaScript SDK are able to use this metadata, but the output ofall_flags_state()will still work with older versions. - The
all_flags_state()method also allows you to select only client-side-enabled flags to pass to the front end, by using the optionclient_side_only: true.
- Unexpected exceptions are now logged at
ERRORlevel, and exception stacktraces atDEBUGlevel. Previously, both were being logged atWARNlevel.
LDClient.all_flags()
Fixed a regression in version 5.0.0 that could prevent the client from reconnecting if the stream connection was dropped by the server.
- The client no longer uses Celluloid for streaming I/O. Instead, it uses socketry.
- The client now treats most HTTP 4xx errors as unrecoverable: that is, after receiving such an error, it will not make any more HTTP requests for the lifetime of the client instance, in effect taking the client offline. This is because such errors indicate either a configuration problem (invalid SDK key) or a bug, which is not likely to resolve without a restart or an upgrade. This does not apply if the error is 400, 408, 429, or any 5xx error.
- During initialization, if the client receives any of the unrecoverable errors described above, the client constructor will return immediately; previously it would continue waiting until a timeout. The
initialized?method will return false in this case.
- The SDK no longer supports Ruby versions below 2.2.6, or JRuby below 9.1.16.
- To reduce the network bandwidth used for analytics events, feature request events are now sent as counters rather than individual events, and user details are now sent only at intervals rather than in each event. These behaviors can be modified through the LaunchDarkly UI and with the new configuration option
inline_users_in_events.
- JRuby 1.7 is no longer supported.
- Greatly reduced the number of indirect gem dependencies by removing
moneta, which was previously a requirement for the Redis feature store.
- In the Redis feature store, fixed a synchronization problem that could cause a feature flag update to be missed if several of them happened in rapid succession.
- Improved efficiency of logging by not constructing messages that won't be visible at the current log level. (Thanks, julik!)
- Fixed a bug that could prevent very large feature flags from being updated in streaming mode.
- Support for a new LaunchDarkly feature: reusable user segments.
- The feature store interface has been changed to support user segment data as well as feature flags. Existing code that uses
InMemoryFeatureStoreorRedisFeatureStoreshould work as before, but custom feature store implementations will need to be updated.
- Adds support for a future LaunchDarkly feature, coming soon: semantic version user attributes.
- It is now possible to compute rollouts based on an integer attribute of a user, not just a string attribute.
- Reduce logging level for missing flags
- Relax json and faraday dependencies
- Wrap redis bulk updates in a transaction
- Fixed documentation links
- Will use feature store if already initialized even if connection to service could not be established. This is useful when flags have been initialized in redis.
- Increase default and minimum polling interval to 30s
- Strip out unknown top-level attributes
- Make sure redis store initializations are atomic
- Include source code for changes described in 2.3.0
- Add
closemethod to Ruby client to stop processing events - Add support for Redis feature store
- Add support for LDD mode
- Allow user to disable outgoing event stream.
- Stop retrying on 401 responses (due to bad sdk keys)
- Update Readme to fix instructions on installing gem using command line
- Cleaned up formatting on various files (Rubocop)
- Added proxy support to streaming and http connections. Respects
HTTP_PROXYandhttp_proxyenvironment variables as well as the:proxy => protocol://user:pass@hostconfiguration parameter.
- Updated changelog
- Bumped nio4r to 2.0
- Improved handling of http status codes that may not be integers.
- Improved error handling when connected to flag update stream.
- Indirect stream events are now correctly processed
- The default logger now logs at
infolevel
- Support for multivariate feature flags. In addition to booleans, feature flags can now return numbers, strings, dictionaries, or arrays via the
variationmethod. - New
all_flagsmethod returns all flag values for a specified user. - If streaming is disabled, the client polls for feature flag changes. If streaming is disabled, the client will default to polling LaunchDarkly every second for updates. The poll interval is configurable via
poll_interval. - New
secure_mode_hashfunction computes a hash suitable for the new LaunchDarkly JavaScript client's secure mode feature. - Support for extremely large feature flags. When a large feature flag changes, the stream will include a directive to fetch the updated flag.
- You can now initialize the LaunchDarkly client with an optional timeout (specified in seconds). This will block initialization until the client has finished bootstrapping and is able to serve feature flags.
- The streaming implementation (
StreamProcessor) uses Celluloid under the hood instead of EventMachine. The dependency on EventMachine has been removed. - The
storeoption has been renamed tocache_store. - Offline mode can no longer be set dynamically. Instead, at configuration time, the
offlineparameter can be set to put the client in offline mode. It is no longer possible to dynamically change whether the client is online and offline (viaset_onlineandset_offline). Calloffline?to determine whether or not the client is offline. - The
debug_streamconfiguration option has been removed. - The
log_timingsconfiguration option has been removed.
- The
togglecall has been deprecated in favor ofvariation.
update_user_flag_settinghas been removed. To change user settings, use the LaunchDarkly REST API.