in core we wrap request handlers with an asserting one to ensure we can serialize messages
with different versions. Yet, xpack uses the same functionality to add security aspects to
the network layer. These tests assert that the right handlers are in-place.
Original commit: elastic/x-pack-elasticsearch@e39c8995ae
Fixes to x-plugins code now that DateMathParser accepts a LongSupplier rather than a Callable to get the value of now
Relates to elastic/elasticsearchelastic/elasticsearch#20796
Original commit: elastic/x-pack-elasticsearch@99fc47a8a7
This change moves to using SSLParameters as the configuration source for SSLEngine and SSLSocket
objects that are configured by the SSLService. Previously we used a mix of specific methods and
SSLParameters, which resulted in issues where ordering of calls is important. For example, if configuring
client authentication directly on the engine prior to setting the SSLParameters resulted in the client
authentication configuration being reset to the default.
Additionally, this change also sets use cipher suite order to true to ensure preferred ciphers will be used.
Original commit: elastic/x-pack-elasticsearch@8ddecdc20c
This change ensures we wait for a response before the async http client is closed. Otherwise we can
close the client during the connection to the remote endpoint or never even connect to the remote
endpoint.
Closeselastic/elasticsearch#3640
Original commit: elastic/x-pack-elasticsearch@54900b1b4a
This changes does two things in the tribe tests. The first is that when we split data up between
multiple clusters, we always force create the security index so that randomization does not cause
edge cases like the index not existing in the preferred cluster. The second is we look at the cluster
state of the nodes and ensure the tribe node sees the indices and has all primaries active.
Separate tests were also added to cover the scenario where the security index only exists in the non
preferred node.
Original commit: elastic/x-pack-elasticsearch@17b78ec837
This is the last action that needs additional support for proxies.
You can set a proxy in the JSON like this:
```
"actions" : {
"notify-pagerduty" : {
"pagerduty" : {
"description" : "Main system down, please check!",
"proxy" : { "host" : "localhost", "port" : 8080 }
}
}
}
```
Closeselastic/elasticsearch#3372
Original commit: elastic/x-pack-elasticsearch@b99969fd6b
You can set it like this in the JSON
"actions" : {
"notify-slack" : {
"slack" : {
"account" : "integration-account",
"proxy" : {
"host" : "localhost",
"port" : 8080
},
"message" : {
...
}
}
}
}
Relates elastic/elasticsearch#3372
Original commit: elastic/x-pack-elasticsearch@de86233d4f
Watcher uses a custom thread pool. This is because executing watches can
be long-running tasks that often block on I/O and it is best to not
consume the core thread pools with these tasks. Today this thread pool
is fixed, and sized at five times the bounded number of cores (so 160 on
a 32-core box). It makes sense for there to possibly be so many threads,
again because these tasks can block on I/O and having excess capacity
lets unblocked watches execute. It's the fixed size that can cause
problem, all these threads are always consuming resources even when
there are no or not that many watches running. This commit changes this
thread pool to be a scaling thread pool.
Relates elastic/elasticsearch#3660
Original commit: elastic/x-pack-elasticsearch@3cafab6e83
We need to special case IndicesAliasesRequest as it doesn't implement CompositeIndicesRequest anymore. Note that the similar loop for CompositeIndicesRequests's subrequests will soon go away
Relates to elastic/elasticsearch#3638
Original commit: elastic/x-pack-elasticsearch@50d119ff61
This change allows native users and roles to be used on tribe nodes. The tribe node will actually
use the security index of one of the tribes, which must be specified with the `tribe.on_conflict`
setting. User and role modifications are not permitted when running on a tribe node.
Closeselastic/elasticsearch#3451
Original commit: elastic/x-pack-elasticsearch@2b762ca648
When deleting a watch the version was used as part of
the delete request. However a watch deletion means the
user really wants to get rid of it and not accidentally
run into a version exception because the watch was running in
between.
Original commit: elastic/x-pack-elasticsearch@e585f717f1
In preparation for elastic/elasticsearch#2957, I found we have things both in the root level
dev-tools, as well as elasticsearch/x-dev-tools. Most of this stuff can
be removed as it had to do with the old manual release process. There
was also a signed license file checked in. I removed it here, we really
should not have licenses checked in IMO, and it is unclear what the
purpose of this license was for. The two remaining scripts were moved to
the root dev-tools.
Original commit: elastic/x-pack-elasticsearch@3e24ea2d56
When running as a node, we check the `xpack.security.transport.filter.enabled` setting to see
if we should create the IPFilter but this check is not really correct. The HTTP filter could be
enabled or a profile filter could be enabled so there are times when we may not be filtering connections
when we should. Additionally, since we do not bind the IPFilter to a null provider, Guice will try to create
one during startup to inject into the security transport. This results in an exception and startup fails.
This change always creates the IPFilter when running as a node. This IPFilter has its own settings and
logic to determine whether it should be filtering on a given network transport.
Closeselastic/elasticsearch#3592
Original commit: elastic/x-pack-elasticsearch@95c25651c4
Some methods have been renamed in elastic/elasticsearchelastic/elasticsearch#20560. This commit change a .bytes() call to a .getBytes() call.
Original commit: elastic/x-pack-elasticsearch@4a0ff77361
This commit fixes the cat.templates REST tests so that it works when other templates exist (like monitoring)
Original commit: elastic/x-pack-elasticsearch@2e27ad88b4
This change switches the build to use the licensing prod key when
building the xpack jar for release.
Original commit: elastic/x-pack-elasticsearch@54a21dae5b
This adds an "interval" placeholder parameter that is required to the Monitoring Bulk API, and adds it to the Kibana side of the plumbing.
Having this will allow us to add it to all incoming documents and start to report against it with the Insights, as well as to detect the _lack_ of incoming documents.
By adding it now, we can avoid having a non-BWC API change for Kibana in 5.1. We'll just pickup new data in our documents.
Original commit: elastic/x-pack-elasticsearch@5ba8aafe03
This is a followup for elastic/elasticsearchelastic/elasticsearch#20526 removes the pluggability of
transport / http server transport via guice.
Original commit: elastic/x-pack-elasticsearch@5fb84949aa
The x-pack client jar and api jar are exactly the same: the entirety of
x-pack. Since we added the x-pack-transport jar, we no longer really
want the client jar as it is confusing. Additionally, it causes jar hell
when a test for an extension uses the transport client. This change
removes the client jar, and makes the x-pack transport client use the
api jar instead. This sounds odd at first, but since transport client is
going away eventually, it is a stopgap, and works.
closeselastic/elasticsearch#3309
Original commit: elastic/x-pack-elasticsearch@ee7a2c12c0
Due to untested code there was an NPE happening in production,
when a chained input execution failed, but the chained input
tried to access the resulting payload (which is never set on
failures). This payload now defaults to being empty.
This commit also drive-by fixes a broken logging statement, that
on the one side returned not the watch id, but a useless watch
toString() representation, and on the other hand only logs an error
message, but not a stack trace into the log, as this is what the
history is for.
Original commit: elastic/x-pack-elasticsearch@7dbe1afd90
Whenever a watch is updated (put, delete, set state), until now we
happened to reject those operations when a watch was executed at the
same time. However with long running reporting this might mean, that a
watch can never be changed, because it always gets executed.
* Removes the ability of write requests to obtain a lock at all (executing watches is still protected by a lock)
* Replaced the FairKeyedLock in watcher with the KeyedLock in Elasticsearch, which also takes a fair option, removed the FairKeyedLock
* Removed all the timeout parameters that are no longer needed, because there is no lock anymore
* Removed also the force parameter for watch deletion. Just do it[tm]
* Added a test that deleting a watch while it is being executed does not leave any leftovers
In case of a deletion of a watch during an execution, so that updating the status of the watch fails,
a warning is logged.
Closeselastic/elasticsearch#3417
Original commit: elastic/x-pack-elasticsearch@22fad1b797
As part of the review of elastic/elasticsearch#3287, the stopping of the IndexAuditTrail was moved to the tearDown
method. This works sometimes but other times it fails because tearDown is run after
ESIntegTestCase#after, so the IndexAuditTrail is still running during the after checks which will
cause the test to fail since the shard lock cannot be obtained.
Closeselastic/elasticsearch#3520
Original commit: elastic/x-pack-elasticsearch@4cb52b15a2
Today the operation mode can be set to default for a short amout of
time until it's reset to the actual mode this can cause weird sideeffects
for users if it's read concurrently. Also the test relies on a certain
happens before relationship that is not guaranteed since the operation
mode is set before the listerner is run. This change also rewrites the test
to not use busy waiting but wait for the actual listern to be executed.
Original commit: elastic/x-pack-elasticsearch@a2a42b89e5
1. We only support indexes created by Marvel 2.3+. All other indexes
are just ignored.
2. The tests don't assert a ton of interesting stuff because there
isn't a java API for Monitoring that we can just use. Instead we assert
that a few objects are there and look sane.
3. We don't migrate the contents of the data index. Instead we just
rely on Monitoring recreating it.
Original commit: elastic/x-pack-elasticsearch@86216c2d61
The IndexAuditTrail had both a stop and close method that needed to be called in order
to stop the service. There was a race where we called either flush or close in a non
blocking fashion and then immediately closed the underlying client. This change makes
the stop method wait for up to 10 seconds when closing the bulk processor.
Closeselastic/elasticsearch#3279
Original commit: elastic/x-pack-elasticsearch@0d776bc91a
TransportService is not pluggable anymore in core. Instead we now have a interceptor
infrastructure that allows to intercept send and receive calls on the transport layer.
Relates to elastic/elasticsearchelastic/elasticsearch#20505
Original commit: elastic/x-pack-elasticsearch@04194ecb09
This adds proxy support to the hipchat action. Right now
neither hipchat nor slack nor pagerduty allow for this,
but if you dont need a proxy for internal http connections,
but you do for external, then this configuration cannot be done
without setting a proxy for those actions.
You can set it like this in the JSON
```
"actions" : {
"notify-hipchat" : {
"hipchat" : {
"account" : "integration-account",
"proxy" : {
"host" : "localhost",
"port" : 8080
},
"message" : {
...
}
}
}
}
```
Relates elastic/elasticsearch#3372
Original commit: elastic/x-pack-elasticsearch@4e8447ce37
This cleans up some of the reported stats to be a little clearer, including making the JVM Heap chart behave like Kibana's memory chart. This solves two problems: you can now determine the max heap size and you know what "x%" actually means relative to it.
Original commit: elastic/x-pack-elasticsearch@450f6fd546
This commit fixes the logging audit tests which were broken due to an
upstream change in core Elasticsearch relating to the fact that prefixes
are no longer considered part of the log message, but are instead
implemented via markers.
Original commit: elastic/x-pack-elasticsearch@abd7ec23d8
To deny access to a fields users can name exceptions to field permissions with the following syntax:
"fields": {
"grant": [list of field names patterns],
"except": [list of patterns that are forbidden]
}
See doc for the rules for this.
This commit also reverts elastic/elasticsearch#2720closeselastic/elasticsearch#2681
Original commit: elastic/x-pack-elasticsearch@d6537028ec
The enabled and username fields are both now allowed in the request body for the put user
request. This makes it easier to perform a get and update a user without needing to edit more
of the request body than necessary.
Closeselastic/elasticsearch#3391
Original commit: elastic/x-pack-elasticsearch@ab763e843b
Start dates are a required feature for cloud. This functionality adds support
for specifying and enforcing a start date on licenses.
Behaviour: If the start date is > than now, the license will be rejected.
Due to another field in the license class, the version of the License class as well
as its serialization methods are adapted to this.
Closeselastic/elasticsearch#3370
Original commit: elastic/x-pack-elasticsearch@eb2a6f5be3
We were starting nodes at weird times and then shutting them down again,
slowing down the tests and causing the watcher tests to fail because
watcher wasn't being shut down with its traditional kid gloves.
Original commit: elastic/x-pack-elasticsearch@2fd81b3eaf
When the HTTP attachment was not able to successfully retrieve the
data from and endpoint, there was no indication in the watch history
of what went wrong. Instead a logger was used, which is not useful
for the person running the watches.
This commit removes the logger statement and throws an exception,
so that the exception message can be stored in the watch history.
Source of this issue was a forum post:
https://discuss.elastic.co/t/sending-e-mail-with-generated-report-fails/60263/6
Original commit: elastic/x-pack-elasticsearch@acdaf7abef
In our tests you have to explicitly shut down watcher rather than shut
down the node it is running on because of thread leak detection. Just
shutting down the node that it is running on will cause it to start up
on another node if there is another one running and then not properly
shut down. This is probably something that should be fixed in watcher
somehow but for now lets just be more careful with the tests.
Closeselastic/elasticsearch#2365Closeselastic/elasticsearch#2588
Original commit: elastic/x-pack-elasticsearch@fb8a172972
Previously core Elasticsearch had methods in a test class for removing
and adding appenders. However, these methods were moved to production
code. This commit adjusts x-plugins for this change.
Original commit: elastic/x-pack-elasticsearch@83e37ef65a
This particular change focuses on upgrading the source of a watch when it comes to scripts that have no language specified explicitly.
The default language in version 5 changed to painless from whatever is specified in `script.default_lang` setting (this defaulted to groovy). In order to make sure that scripts in watcher remain to work we should rewrite the search source upon startup and set the legacy default language explicitly. The legacy script language is now controlled by `script.legacy.default_lang` setting and that defaults to groovy.
Changing the source upon startup should do the trick and only change the source of watches with scripts that don't have an explicit language set. For new watches the default language used in scripts is painless and because we now always serialize the language explicitly in scripts these watches won't be changed on startup.
The upgrade logic added here tries to upgrade scripts in the following places in a watch:
* script condition
* script transform
* any script defined inside of a search input
Original commit: elastic/x-pack-elasticsearch@4d578819eb
This commit cleans most of the methods of XContentBuilder so that:
- Jackson's convenience methods are used instead of our custom ones (ie field(String,long) now uses Jackson's writeNumberField(String, long) instead of calling writeField(String) then writeNumber(long))
- null checks are added for all field names and values
- methods are grouped by type in the class source
- methods have the same parameters names
- duplicated methods like field(String, String...) and array(String, String...) are removed
- varargs methods now have the "array" name to reflect that it builds arrays
- unused methods like field(String,BigDecimal) are removed
- all methods now follow the execution path: field(String,?) -> field(String) then value(?), and value(?) -> writeSomething() method. Methods to build arrays also follow the same execution path.
Original commit: elastic/x-pack-elasticsearch@d83f3aa6e2
This fixes a bug I found with a customer when he updated from 1.x to 2.x.
Due to an BWC incompatible change in the watch history mapping and a thread
pool rejection during execution a watch was not removed from the triggered
watches and tried to be executed again.
While trying to fix it it turned out that the execution of the failure
test case was still done in the transport thread and thus required some
offloading to another thread pool.
Original commit: elastic/x-pack-elasticsearch@df04ce31f2
If someone deletes the watch index (i.e. by deleting all indices), the watcher
in memory store still contains all the watches and tries to execute watches -
which results in exceptions as the watch itself cannot be updated anymore.
In order to minimize this problem (it cant be get rid of completely), we should
act accordingly if the watch index goes missing (either deleted or closed) and
clear out the memory representation of watches in the watchstore as well as trying
to finish all the current executions.
Closeselastic/elasticsearch#2794
Original commit: elastic/x-pack-elasticsearch@12d98cd566
This change moves the logfile audit output from determining what to log based on the
logger level to a enum based configuration that is used by the index output.
A few notable changes were made:
* We alway log all the information we have except for the request body
* The request body is no longer logged by default for REST events; the user needs to
explicitly opt in as there could be sensitive data in the body
* Added a `realm_authentication_failed` event that separates overall authentication
failure from that of an individual realm
Original commit: elastic/x-pack-elasticsearch@343a2bcdd9
This change adds support for disabling users. Users can be disabled by setting the enabled
property to false and the AuthenticationService will check to make sure that the user is enabled.
If the user is not enabled, this will be audited as an authentication failure.
Also as part of this work, the AnonymousUser was cleaned up to remove having a static instance
that caused issues with tests.
Finally, the poller of users was removed to simplify the code in the NativeUsersStore. In our other
realms we rely on the clear cache APIs and the timeout of the user cache. We should have the
same semantics for the native realm.
Closeselastic/elasticsearch#2172
Original commit: elastic/x-pack-elasticsearch@0820e40183
This rewrites the HTTP Exporter to use the REST client underneath. Functionality is improved in resource blocking (templates and pipelines existing) and the majority of the code fundamentall simplified by removing direct HTTP calls.
This is blocked by the SSLService pull request. After that is merged, the I will update this PR to reflect those changes and it could possibly allow us to remove the security privileges required for monitoring.
Original commit: elastic/x-pack-elasticsearch@1ad25f17f8
Basic backwards compatibility support for watcher.
Closeselastic/elasticsearch#3230
Relates to elastic/elasticsearch#3231 - this actually should fix all the failures caused
by fractional time values but it does so by being able to parse them.
Being able to parse them is important for 2.x compatibility but 5.0
watches shouldn't produce fractional time values. This fixes the
particular way of making fractional time values mentioned in elastic/elasticsearch#3231
but I expect there are a half dozen more places to fix. The actual
watcher tests are fairly basic.
Original commit: elastic/x-pack-elasticsearch@328717455c
This publishes X-Pack usage data to the cluster info from the elected master node. This allows phone home to retrieve this data from the index, rather than fetching it live from the connected cluster (thereby not getting it from any n - 1 clusers that are not connceted).
Original commit: elastic/x-pack-elasticsearch@79bfaaaf0b
This removes the "agent" package from org.elasticsearch.xpack.monitoring.agent.*, so that now everything is simply org.elasticsearch.xpack.monitoring.*.
Follow-on work will be refactoring some of the other code, but this is a first step now that it's always the agent (in effect).
Original commit: elastic/x-pack-elasticsearch@14025cb17c
This change migrates xpack (security, watcher, and monitoring) to use the common ssl
configuration for the elastic stack. As part of this work, several aspects of how we deal
with SSL has been modified.
From a functionality perspective, an xpack wide configuration for SSL was added and
all of the code that needs SSL uses the SSLService now. The following is a list of all
of the aspects of xpack that can have their own SSL configuration, which are separate
from the xpack wide configuration:
* Transport
* Transport profiles
* HTTP Transport
* Realms
* Monitoring Exporters
* HTTP Client
In terms of the code, some cleanups were made with these changes. SSLConfiguration is
now a concrete class and SSLConfiguration.Custom and SSLConfiguration.Global have been
removed. The validate method on key and trust configurations has been removed and these
classes will now throw exceptions when they are constructed with bad values. The
OptionalSettings helper class has been removed as it was just a file with one line functions
that made the code harder to understand. The SSL configuration and service classes have
been moved from the security source directories to the main xpack source set. The SSLService
now handles more of the configuration of the SSLEngine it returns to prevent callers from
having to handle those aspects. The settings that get registered for SSL have been moved to
XPackSettings.
Also included in this PR is a update to the docs around SSL. This includes a large simplification to
the documentation in that the certificate authority configuration section has been removed and the
process that is documented for generating certificates only includes the CLI tool that we bundle.
Closeselastic/elasticsearch#3104Closeselastic/elasticsearch#2971Closeselastic/elasticsearch#3164
Original commit: elastic/x-pack-elasticsearch@5bd9e5ef38
* master:
Changes tests to conform with new cluster health API, calling setWaitForNoRelocatingShards(true) instead of setWaitForRelocatingShards(0)
Original commit: elastic/x-pack-elasticsearch@bde6ad8c8a
* master:
Use releasable locks in NativeRolesStore
security: limit the size of the role store cache
security: remove explicit handshake wait in netty4 transport
test: smoke-test-plugins-ssl no longer relies on logging to start
kibana monitoring/uuid config key reference update
Docs: Updated release date for 2.4 in RNs.
Update README.md
Build: Add apijar task to assemble so it gets built with other artifacts
monitoring ui/license: cluster listing status cell treatment for basic/unsupported cluster
monitoring ui:fix cluster overview when cluster has no indices/shards
monitoring ui/license: logic cleanup per feedback
monitoring ui/license: primary cluster asterisk styling
monitoring ui/license: allow clicking into primary cluster if all are basic
monitoring ui: add isPrimary property to cluster listing response
Security: throw exception if we cannot extract indices from an indices request
Security: add tests for delete and update by query
Original commit: elastic/x-pack-elasticsearch@3cb41739ee
Previously the roles store cache was unbounded as it was a just using a ConcurrentHashMap,
which could lead to excessive memory usage in cases where there are a large number of roles
as we tried to eagerly load the roles into the cache if they were not present. The roles store now
loads roles on demand and caches them for a finite period of time.
Additionally, the background polling of roles has been removed to reduce complexity. A best effort
attempt is made to clear the roles cache upon modification and if necessary the cache can be
cleared manually.
See elastic/elasticsearch#1837
Original commit: elastic/x-pack-elasticsearch@450dd779c8
Netty 4's SslHandler does not require the application to wait for the handshake to
be completed before data is written. This change removes the explicit wait on each
handshake future.
Original commit: elastic/x-pack-elasticsearch@c19bcebb83
This commit removes some unnecessary calls to Path#toString when logging
a path in OperationFileModeWatcher#onChage. The calls to Path#toString
are not necessary since the logging infrastructure will do this anyway.
Original commit: elastic/x-pack-elasticsearch@da1bc30d60
This commit removes an unnecessary logging of an exception message from
HttpEmailAttachementParser#toAttachment since the full exception is
logged anyway.
Original commit: elastic/x-pack-elasticsearch@55d4bada00
This commit modifies the call sites that allocate a parameterized
message to use a supplier so that allocations are avoided unless the log
level is fine enough to emit the corresponding log message.
Original commit: elastic/x-pack-elasticsearch@955ab89b8b
This used to be an assertion but we move it to an exception to be able to catch this at all times without requiring assertion enabled
Original commit: elastic/x-pack-elasticsearch@fcb5fbe852
When migrating the package and class names from shield to xpack.security, the logging.yml
file was updated incorrectly which resulted in audit entries going to the wrong file. This change
specifies the correct package name so audit logs go to the correct file.
Closeselastic/elasticsearch#3063
Original commit: elastic/x-pack-elasticsearch@5c07108392
This change allows extensions to define their own filtered settings. This is used when there
may be sensitive credentials stored in a custom realm that should be filtered out of some
API calls.
Closeselastic/elasticsearch#2847
Original commit: elastic/x-pack-elasticsearch@952474daba
With elastic/elasticsearchelastic/elasticsearch#19865 the XContentBuilder has changed to support both inclusive and exclusive filters and now uses Set<String> instead of arrays of Strings. This change updates the various places in x-plugins where string arrays were used.
Original commit: elastic/x-pack-elasticsearch@1f8d4485f4
Nik Everett
jaymode
Simon Willnauer
Simon Willnauer
Colin Goodheart-Smithe
Igor Motov
Jay Modi
jaymode
Alexander Reelsen
Jason Tedor
Luca Cavanna
Ryan Ernst
Tanguy Leroux
Chris Earle
Boaz Leskes
Yannick Welsch
Britta Weber
Martijn van Groningen
javanna
Britta Weber
Jack Conradson
Ali Beyad