Moves the application of jackson defaults overrides into pure java, and
applies them statically _before_ the `org.logstash.ObjectMappers` has a chance
to start initializing object mappers that rely on the defaults.
We replace the runner's invocation (which was too late to be fully applied) with
a _verification_ that the configured defaults have been applied.
* pipeline bus: add deadlock test for unlisten/unregisterSender
* pipeline bus: eliminate deadlock
Moves the sync-to-notify out of the `AddressStateMapping#mutate`'s effective
synchronous block to eliminate a race condition where unlistening to an address
and unregistering a sender could deadlock.
It is safe to notify an AddressState's attached input without exclusive access
to the AddressState, because notifying an input that has since been disconnected
is net zero harm.
(cherry picked from commit 8af6343a26)
Co-authored-by: Ry Biesemeyer <yaauie@users.noreply.github.com>
Introduce a new flag setting `xpack.monitoring.allow_legacy_collection` which eventually enable the legacy monitoring collector.
Update the method to test if monitoring is enabled so that consider also `xpack.monitoring.allow_legacy_collection` to determine if `monitoring.*` settings are valid or not.
By default it's false, the user has to intentionally enable it to continue to use the legacy monitoring settings.
---------
Co-authored-by: kaisecheng <69120390+kaisecheng@users.noreply.github.com>
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
Switch the default value of `pipeline.buffer.type` to use the heap memory instead of direct one.
Change the default value of the setting `pipeline.buffer.type` from direct to heap and update consequently the documentation.
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
The current max inflight error message focuses on a single pipeline and on a maximum amount of 10k events regardless of the heap size.
The new warning will take into account all loaded pipelines and also consider the heap size, giving a warning if the total number of events consumes 10% or more of the total heap.
For the purpose of the warning events are assumed to be 2KB as it a normal size for a small log entry.
Adds a JMH benchmark to measure the peformances of BufferedTokenizerExt.
Update also Gradle build script to remove CMS GC flags and fix deprecations for Gradle 9.0.
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
Fixes the behaviour of the tokenizer to be able to work properly when buffer full conditions are met.
Updates BufferedTokenizerExt so that can accumulate token fragments coming from different data segments. When a "buffer full" condition is matched, it record this state in a local field so that on next data segment it can consume all the token fragments till the next token delimiter.
Updated the accumulation variable from RubyArray containing strings to a StringBuilder which contains the head token, plus the remaining token fragments are stored in the input array.
Furthermore it translates the `buftok_spec` tests into JUnit tests.
* [health] bootstrap HealthObserver from agent to API (#16141)
* [health] bootstrap HealthObserver from agent to API
* specs: mocked agent needs health observer
* add license headers
* Merge `main` into `feature/health-report-api` (#16397)
* Add GH vault plugin bot to allowed list (#16301)
* regenerate webserver test certificates (#16331)
* correctly handle stack overflow errors during pipeline compilation (#16323)
This commit improves error handling when pipelines that are too big hit the Xss limit and throw a StackOverflowError. Currently the exception is printed outside of the logger, and doesn’t even show if log.format is json, leaving the user to wonder what happened.
A couple of thoughts on the way this is implemented:
* There should be a first barrier to handle pipelines that are too large based on the PipelineIR compilation. The barrier would use the detection of Xss to determine how big a pipeline could be. This however doesn't reduce the need to still handle a StackOverflow if it happens.
* The catching of StackOverflowError could also be done on the WorkerLoop. However I'd suggest that this is unrelated to the Worker initialization itself, it just so happens that compiledPipeline.buildExecution is computed inside the WorkerLoop class for performance reasons. So I'd prefer logging to not come from the existing catch, but from a dedicated catch clause.
Solves #16320
* Doc: Reposition worker-utilization in doc (#16335)
* settings: add support for observing settings after post-process hooks (#16339)
Because logging configuration occurs after loading the `logstash.yml`
settings, deprecation logs from `LogStash::Settings::DeprecatedAlias#set` are
effectively emitted to a null logger and lost.
By re-emitting after the post-process hooks, we can ensure that they make
their way to the deprecation log. This change adds support for any setting
that responds to `Object#observe_post_process` to receive it after all
post-processing hooks have been executed.
Resolves: elastic/logstash#16332
* fix line used to determine ES is up (#16349)
* add retries to snyk buildkite job (#16343)
* Fix 8.13.1 release notes (#16363)
make a note of the fix that went to 8.13.1: #16026
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
* Update logstash_releases.json (#16347)
* [Bugfix] Resolve the array and char (single | double quote) escaped values of ${ENV} (#16365)
* Properly resolve the values from ENV vars if literal array string provided with ENV var.
* Docker acceptance test for persisting keys and use actual values in docker container.
* Review suggestion.
Simplify the code by stripping whitespace before `gsub`, no need to check comma and split.
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
---------
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
* Doc: Add SNMP integration to breaking changes (#16374)
* deprecate java less-than 17 (#16370)
* Exclude substitution refinement on pipelines.yml (#16375)
* Exclude substitution refinement on pipelines.yml (applies on ENV vars and logstash.yml where env2yaml saves vars)
* Safety integration test for pipeline config.string contains ENV .
* Doc: Forwardport 8.15.0 release notes to main (#16388)
* Removing 8.14 from ci/branches.json as we have 8.15. (#16390)
---------
Co-authored-by: ev1yehor <146825775+ev1yehor@users.noreply.github.com>
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
* Squashed merge from 8.x
* Failure injector plugin implementation. (#16466)
* Test purpose only failure injector integration (filter and output) plugins implementation. Add unit tests and include license notes.
* Fix the degrate method name typo.
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
* Add explanation to the config params and rebuild plugin gem.
---------
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
* Health report integration tests bootstrapper and initial tests implementation (#16467)
* Health Report integration tests bootstrapper and initial slow start scenario implementation.
* Apply suggestions from code review
Renaming expectation check method name.
Co-authored-by: kaisecheng <69120390+kaisecheng@users.noreply.github.com>
* Changed to branch concept, YAML structure simplified as changed to Dict.
* Apply suggestions from code review
Reflect `help_url` to the integration test.
---------
Co-authored-by: kaisecheng <69120390+kaisecheng@users.noreply.github.com>
* health api: expose `GET /_health_report` with pipelines/*/status probe (#16398)
Adds a `GET /_health_report` endpoint with per-pipeline status probes, and wires the
resulting report status into the other API responses, replacing their hard-coded `green`
with a meaningful status indication.
---------
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
* docs: health report API, and diagnosis links (feature-targeted) (#16518)
* docs: health report API, and diagnosis links
* Remove plus-for-passthrough markers
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
---------
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
* merge 8.x into feature branch... (#16519)
* Add GH vault plugin bot to allowed list (#16301)
* regenerate webserver test certificates (#16331)
* correctly handle stack overflow errors during pipeline compilation (#16323)
This commit improves error handling when pipelines that are too big hit the Xss limit and throw a StackOverflowError. Currently the exception is printed outside of the logger, and doesn’t even show if log.format is json, leaving the user to wonder what happened.
A couple of thoughts on the way this is implemented:
* There should be a first barrier to handle pipelines that are too large based on the PipelineIR compilation. The barrier would use the detection of Xss to determine how big a pipeline could be. This however doesn't reduce the need to still handle a StackOverflow if it happens.
* The catching of StackOverflowError could also be done on the WorkerLoop. However I'd suggest that this is unrelated to the Worker initialization itself, it just so happens that compiledPipeline.buildExecution is computed inside the WorkerLoop class for performance reasons. So I'd prefer logging to not come from the existing catch, but from a dedicated catch clause.
Solves #16320
* Doc: Reposition worker-utilization in doc (#16335)
* settings: add support for observing settings after post-process hooks (#16339)
Because logging configuration occurs after loading the `logstash.yml`
settings, deprecation logs from `LogStash::Settings::DeprecatedAlias#set` are
effectively emitted to a null logger and lost.
By re-emitting after the post-process hooks, we can ensure that they make
their way to the deprecation log. This change adds support for any setting
that responds to `Object#observe_post_process` to receive it after all
post-processing hooks have been executed.
Resolves: elastic/logstash#16332
* fix line used to determine ES is up (#16349)
* add retries to snyk buildkite job (#16343)
* Fix 8.13.1 release notes (#16363)
make a note of the fix that went to 8.13.1: #16026
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
* Update logstash_releases.json (#16347)
* [Bugfix] Resolve the array and char (single | double quote) escaped values of ${ENV} (#16365)
* Properly resolve the values from ENV vars if literal array string provided with ENV var.
* Docker acceptance test for persisting keys and use actual values in docker container.
* Review suggestion.
Simplify the code by stripping whitespace before `gsub`, no need to check comma and split.
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
---------
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
* Doc: Add SNMP integration to breaking changes (#16374)
* deprecate java less-than 17 (#16370)
* Exclude substitution refinement on pipelines.yml (#16375)
* Exclude substitution refinement on pipelines.yml (applies on ENV vars and logstash.yml where env2yaml saves vars)
* Safety integration test for pipeline config.string contains ENV .
* Doc: Forwardport 8.15.0 release notes to main (#16388)
* Removing 8.14 from ci/branches.json as we have 8.15. (#16390)
* Increase Jruby -Xmx to avoid OOM during zip task in DRA (#16408)
Fix: #16406
* Generate Dataset code with meaningful fields names (#16386)
This PR is intended to help Logstash developers or users that want to better understand the code that's autogenerated to model a pipeline, assigning more meaningful names to the Datasets subclasses' fields.
Updates `FieldDefinition` to receive the name of the field from construction methods, so that it can be used during the code generation phase, instead of the existing incremental `field%n`.
Updates `ClassFields` to propagate the explicit field name down to the `FieldDefinitions`.
Update the `DatasetCompiler` that add fields to `ClassFields` to assign a proper name to generated Dataset's fields.
* Implements safe evaluation of conditional expressions, logging the error without killing the pipeline (#16322)
This PR protects the if statements against expression evaluation errors, cancel the event under processing and log it.
This avoids to crash the pipeline which encounter a runtime error during event condition evaluation, permitting to debug the root cause reporting the offending event and removing from the current processing batch.
Translates the `org.jruby.exceptions.TypeError`, `IllegalArgumentException`, `org.jruby.exceptions.ArgumentError` that could happen during `EventCodition` evaluation into a custom `ConditionalEvaluationError` which bubbles up on AST tree nodes. It's catched in the `SplitDataset` node.
Updates the generation of the `SplitDataset `so that the execution of `filterEvents` method inside the compute body is try-catch guarded and defer the execution to an instance of `AbstractPipelineExt.ConditionalEvaluationListener` to handle such error. In this particular case the error management consist in just logging the offending Event.
---------
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
* Update logstash_releases.json (#16426)
* Release notes for 8.15.1 (#16405) (#16427)
* Update release notes for 8.15.1
* update release note
---------
Co-authored-by: logstashmachine <43502315+logstashmachine@users.noreply.github.com>
Co-authored-by: Kaise Cheng <kaise.cheng@elastic.co>
(cherry picked from commit 2fca7e39e8)
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
* Fix ConditionalEvaluationError to do not include the event that errored in its serialiaxed form, because it's not expected that this class is ever serialized. (#16429) (#16430)
Make inner field of ConditionalEvaluationError transient to be avoided during serialization.
(cherry picked from commit bb7ecc203f)
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
* use gnu tar compatible minitar to generate tar artifact (#16432) (#16434)
Using VERSION_QUALIFIER when building the tarball distribution will fail since Ruby's TarWriter implements the older POSIX88 version of tar and paths will be longer than 100 characters.
For the long paths being used in Logstash's plugins, mainly due to nested folders from jar-dependencies, we need the tarball to follow either the 2001 ustar format or gnu tar, which is implemented by the minitar gem.
(cherry picked from commit 69f0fa54ca)
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
* account for the 8.x in DRA publishing task (#16436) (#16440)
the current DRA publishing task computes the branch from the version
contained in the version.yml
This is done by taking the major.minor and confirming that a branch
exists with that name.
However this pattern won't be applicable for 8.x, as that branch
currently points to 8.16.0 and there is no 8.16 branch.
This commit falls back to reading the buildkite injected
BUILDKITE_BRANCH variable.
(cherry picked from commit 17dba9f829)
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
* Fixes the issue where LS wipes out all quotes from docker env variables. (#16456) (#16459)
* Fixes the issue where LS wipes out all quotes from docker env variables. This is an issue when running LS on docker with CONFIG_STRING, needs to keep quotes with env variable.
* Add a docker acceptance integration test.
(cherry picked from commit 7c64c7394b)
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
* Known issue for 8.15.1 related to env vars references (#16455) (#16469)
(cherry picked from commit b54caf3fd8)
Co-authored-by: Luca Belluccini <luca.belluccini@elastic.co>
* bump .ruby_version to jruby-9.4.8.0 (#16477) (#16480)
(cherry picked from commit 51cca7320e)
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
* Release notes for 8.15.2 (#16471) (#16478)
Co-authored-by: andsel <selva.andre@gmail.com>
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
(cherry picked from commit 01dc76f3b5)
* Change LogStash::Util::SubstitutionVariables#replace_placeholders refine argument to optional (#16485) (#16488)
(cherry picked from commit 8368c00367)
Co-authored-by: Edmo Vamerlatti Costa <11836452+edmocosta@users.noreply.github.com>
* Use jruby-9.4.8.0 in exhaustive CIs. (#16489) (#16491)
(cherry picked from commit fd1de39005)
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
* Don't use an older JRuby with oraclelinux-7 (#16499) (#16501)
A recent PR (elastic/ci-agent-images/pull/932) modernized the VM images
and removed JRuby 9.4.5.0 and some older versions.
This ended up breaking exhaustive test on Oracle Linux 7 that hard coded
JRuby 9.4.5.0.
PR https://github.com/elastic/logstash/pull/16489 worked around the
problem by pinning to the new JRuby, but actually we don't
need the conditional anymore since the original issue
https://github.com/jruby/jruby/issues/7579#issuecomment-1425885324 has
been resolved and none of our releasable branches (apart from 7.17 which
uses `9.2.20.1`) specify `9.3.x.y` in `/.ruby-version`.
Therefore, this commit removes conditional setting of JRuby for
OracleLinux 7 agents in exhaustive tests (and relies on whatever
`/.ruby-version` defines).
(cherry picked from commit 07c01f8231)
Co-authored-by: Dimitrios Liappis <dimitrios.liappis@gmail.com>
* Improve pipeline bootstrap error logs (#16495) (#16504)
This PR adds the cause errors details on the pipeline converge state error logs
(cherry picked from commit e84fb458ce)
Co-authored-by: Edmo Vamerlatti Costa <11836452+edmocosta@users.noreply.github.com>
* Logstash Health Report Tests Buildkite pipeline setup. (#16416) (#16511)
(cherry picked from commit 5195332bc6)
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
* Make health report test runner script executable. (#16446) (#16512)
(cherry picked from commit 2ebf2658ff)
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
* Backport PR #16423 to 8.x: DLQ-ing events that trigger an conditional evaluation error. (#16493)
* DLQ-ing events that trigger an conditional evaluation error. (#16423)
When a conditional evaluation encounter an error in the expression the event that triggered the issue is sent to pipeline's DLQ, if enabled for the executing pipeline.
This PR engage with the work done in #16322, the `ConditionalEvaluationListener` that is receives notifications about if-statements evaluation failure, is improved to also send the event to DLQ (if enabled in the pipeline) and not just logging it.
(cherry picked from commit b69d993d71)
* Fixed warning about non serializable field DeadLetterQueueWriter in serializable AbstractPipelineExt
---------
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
* add deprecation log for `--event_api.tags.illegal` (#16507) (#16515)
- move `--event_api.tags.illegal` from option to deprecated_option
- add deprecation log when the flag is explicitly used
relates: #16356
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
(cherry picked from commit a4eddb8a2a)
Co-authored-by: kaisecheng <69120390+kaisecheng@users.noreply.github.com>
---------
Co-authored-by: ev1yehor <146825775+ev1yehor@users.noreply.github.com>
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
Co-authored-by: kaisecheng <69120390+kaisecheng@users.noreply.github.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Luca Belluccini <luca.belluccini@elastic.co>
Co-authored-by: Edmo Vamerlatti Costa <11836452+edmocosta@users.noreply.github.com>
Co-authored-by: Dimitrios Liappis <dimitrios.liappis@gmail.com>
---------
Co-authored-by: ev1yehor <146825775+ev1yehor@users.noreply.github.com>
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
Co-authored-by: Andrea Selva <selva.andre@gmail.com>
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
Co-authored-by: kaisecheng <69120390+kaisecheng@users.noreply.github.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Luca Belluccini <luca.belluccini@elastic.co>
Co-authored-by: Edmo Vamerlatti Costa <11836452+edmocosta@users.noreply.github.com>
Co-authored-by: Dimitrios Liappis <dimitrios.liappis@gmail.com>
(cherry picked from commit 7eb5185b4e)
Co-authored-by: Ry Biesemeyer <yaauie@users.noreply.github.com>
- move `--event_api.tags.illegal` from option to deprecated_option
- add deprecation log when the flag is explicitly used
relates: #16356
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>
Update Settings to_hash method to also skip Java DeprecatedAlias and not just the Ruby one.
With PR #15679 was introduced org.logstash.settings.DeprecatedAlias which mirrors the behaviour of Ruby class Setting::DeprecatedAlias. The equality check at Logstash::Settings, as descibed in #16505 (comment), is implemented comparing the maps.
The conversion of Settings to the corresponding maps filtered out the Ruby implementation of DeprecatedAlias but not the Java one.
This PR adds also the Java one to the list of filter.
Suppress some warnings compared with JDK 21
- this-escape uses this before it is completely initialised.
- avoid a non serialisable DeadLetterQueueWriter field from serialisable instance.
When a conditional evaluation encounter an error in the expression the event that triggered the issue is sent to pipeline's DLQ, if enabled for the executing pipeline.
This PR engage with the work done in #16322, the `ConditionalEvaluationListener` that is receives notifications about if-statements evaluation failure, is improved to also send the event to DLQ (if enabled in the pipeline) and not just logging it.
Reimplement the root Ruby Setting class in Java and use it from the Ruby one moving the original Ruby class to a shell wrapping the Java instance.
In particular create a new symmetric hierarchy (at the time just for `Setting`, `Coercible` and `Boolean` classes) to the Ruby one, moving also the feature for setting deprecation. In this way the new `org.logstash.settings.Boolean` is syntactically and semantically equivalent to the old Ruby Boolean class, which replaces.
* Fixes the issue where LS wipes out all quotes from docker env variables. This is an issue when running LS on docker with CONFIG_STRING, needs to keep quotes with env variable.
* Add a docker acceptance integration test.
If a user runs Logstash with a hosted JDK and not the one bundled with Logstash distribution, like setting a specific LS_JAVA_HOME, which is minor than JDK 17 then Logstash refuses to start. Has to provide at least a JDK 17 or unset the LS_JAVA_HOME and let Logstash uses the bundled JDK.
Updates the jvm.options and JvmOptionsParser to remove support for JDK 11. If the options parser identifies that the running JVM is less than 17, it refuses to start.
---------
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
This PR protects the if statements against expression evaluation errors, cancel the event under processing and log it.
This avoids to crash the pipeline which encounter a runtime error during event condition evaluation, permitting to debug the root cause reporting the offending event and removing from the current processing batch.
Translates the `org.jruby.exceptions.TypeError`, `IllegalArgumentException`, `org.jruby.exceptions.ArgumentError` that could happen during `EventCodition` evaluation into a custom `ConditionalEvaluationError` which bubbles up on AST tree nodes. It's catched in the `SplitDataset` node.
Updates the generation of the `SplitDataset `so that the execution of `filterEvents` method inside the compute body is try-catch guarded and defer the execution to an instance of `AbstractPipelineExt.ConditionalEvaluationListener` to handle such error. In this particular case the error management consist in just logging the offending Event.
---------
Co-authored-by: Karen Metts <35154725+karenzone@users.noreply.github.com>
This PR is intended to help Logstash developers or users that want to better understand the code that's autogenerated to model a pipeline, assigning more meaningful names to the Datasets subclasses' fields.
Updates `FieldDefinition` to receive the name of the field from construction methods, so that it can be used during the code generation phase, instead of the existing incremental `field%n`.
Updates `ClassFields` to propagate the explicit field name down to the `FieldDefinitions`.
Update the `DatasetCompiler` that add fields to `ClassFields` to assign a proper name to generated Dataset's fields.
* Exclude substitution refinement on pipelines.yml (applies on ENV vars and logstash.yml where env2yaml saves vars)
* Safety integration test for pipeline config.string contains ENV .
* Properly resolve the values from ENV vars if literal array string provided with ENV var.
* Docker acceptance test for persisting keys and use actual values in docker container.
* Review suggestion.
Simplify the code by stripping whitespace before `gsub`, no need to check comma and split.
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
---------
Co-authored-by: João Duarte <jsvd@users.noreply.github.com>
Because logging configuration occurs after loading the `logstash.yml`
settings, deprecation logs from `LogStash::Settings::DeprecatedAlias#set` are
effectively emitted to a null logger and lost.
By re-emitting after the post-process hooks, we can ensure that they make
their way to the deprecation log. This change adds support for any setting
that responds to `Object#observe_post_process` to receive it after all
post-processing hooks have been executed.
Resolves: elastic/logstash#16332
This commit improves error handling when pipelines that are too big hit the Xss limit and throw a StackOverflowError. Currently the exception is printed outside of the logger, and doesn’t even show if log.format is json, leaving the user to wonder what happened.
A couple of thoughts on the way this is implemented:
* There should be a first barrier to handle pipelines that are too large based on the PipelineIR compilation. The barrier would use the detection of Xss to determine how big a pipeline could be. This however doesn't reduce the need to still handle a StackOverflow if it happens.
* The catching of StackOverflowError could also be done on the WorkerLoop. However I'd suggest that this is unrelated to the Worker initialization itself, it just so happens that compiledPipeline.buildExecution is computed inside the WorkerLoop class for performance reasons. So I'd prefer logging to not come from the existing catch, but from a dedicated catch clause.
Solves #16320
* licenses: allow elv2, standard abbreviation for Elastic License version 2
* json-dump: reduce unicode normalization cost
Since the underlying JrJackson now properly (and efficiently) encodes the
UTF-8 transcode of whichever strings it is given, we no longer need to
pre-normalize to UTF-8 in ruby _except_ when the string is flagged as BINARY
because we have alternate behaviour to preserve valid UTF-8 sequences.
By emitting a _copy_ of binary-flagged strings that have been re-flagged as
UTF-8, we allow the downstream (efficient) encoding operation in jrjackson
to produce equivalent behaviour at much lower cost.
* cleanup: remove orphan unicode normalizer
This commit fixed the configuration reload process to clean up the pipeline's metric store, so it does not retain references to failed pipelines components.
* Add RubyEvent#dup support and unit test case to keep Json#dump(Event) safe.
Co-authored-by: Ry Biesemeyer <ry.biesemeyer@elastic.co>
---------
Co-authored-by: Ry Biesemeyer <ry.biesemeyer@elastic.co>
* fix: restore support for unicode pipeline- and plugin-id's
JRuby's `Ruby#newSymbol(String)` throws an exception when provided a `String`
that contains characters outside of lower-ASCII because JRuby internals expect
"the incoming String to be one of our mangled ISO-8859-1 strings" as noted in
a comment on jruby/jruby#6217.
Instead, we use `Ruby#newString(String)` to create a new `RubyString` (which
works properly), and then rely on `RubyString#intern` to get our `RubySymbol`.
This fixes a regression introduced in the 8.7 series in which pipeline id's
are consistently represented as ruby symbols in the metrics store, and ensures
similar issue does not exist when specifying a plugin id that contains
characters above the lower-ASCII plane.
* fix: use properly-encoded RubySymbol in PipelineConfig
We cannot rely on `RubySymbol#toString` to produce a properly-encoded `String`
whe the string contains characters above the lower-ASCII plane because the
result is effectively a binary ruby-internal marshal of the bytes that only
holds when the symbol contains lower-ASCII.
Instead, we can use the internally-memoizing `RubySymbol#name` to get a
properly-encoded `RubyString`, and `RubyString#asJavaString()` to get a
properly-encoded java-`String`.
* fix: properly serialize unicode pipeline names in API output
Jackson's JSON serializer leaks the JRuby-internal byte structure of Symbols,
which only aligns with the byte-structure of the symbol's actual string when
that string is wholly-comprised of lower-ASCII characters.
By pre-converting Symbols to Strings, we ensure that the result is readable
and useful.
* spec: bypass monitoring specs for unicode pipeline ids when PQ enabled
* Rework the logic to delete DLQ eldest segments to be more resilient on file not found errors and avoid to log warn messages that there isn't any action the user can do to solve.
* Fixed test case, when path point to a file that doesn't exist, rely always on path name comparator. Reworked the code to simplify, not needing anymore the tri-state variable
This a refactoring of test fixture.
Avoid mocking the value returned in global SETTINGS constant. Use instead the local setting map instance used in subject creation.
* p2p: extract interface from v1 pipeline bus
* p2p: extract pipeline push to abstract
* p2p: add opt-in unblocked "v2" implementation
Adds a v2 implementation that does not synchronize on the sender so that
multiple workers can send events through a common `pipeline` output instance
simultaneously.
In this implementation, an `AddressStateMapping` provides synchronized
mutation and cleanup of the underlying `AddressState`, and allows only
queryable mutable views (`AddressState.ReadOnly`) to escape encapsulation.
The implementation also holds indentity-keyed mapping from `PipelineOutput`s
to the set of `AddressState.ReadOnly`s it is regested as a sender for so
that they can be quickly resolved at runtime.
* p2p: more tests for pipeline restart behaviour
* p2p: make v2 pipeline bus the default
Updates the DLQ reader to create a notification file (`.deleted_segment`) which signal when a segment is deleted in consequence of `clean_consumed` set. Updates the DLQ writer to have a filesystem watch so that can receive the reader's signal and update the exposed metric, loading the size by listing FS segments occupation.
The PR was created to skip resolving environment variable references in comments present in the “config.string” pipelines defined in the pipelines.yml file.
However it introduced a bug that no longer resolves env var references in values of settings like pipeline.batch.size or queue.max_bytes.
For now we’ll revert this PR and create a fix that handles both problems.
* pq: avoid blocking writer when queue is precisely full
A PQ is considered full (and therefore needs to block before releasing the
writer) when its persisted size on disk _exceeds_ its `queue.max_bytes`
capacity.
This removes an edge-case preemptive block when the persisted size after
writing an event _meets_ its `queue.max_bytes` precisely AND its current
head page has insufficient room to also accept a hypothetical future event.
Fixes: elastic/logstash#16172
* docs: PQ `queue.max_bytes` cannot be less than `queue.page_capacity`
