Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Auditbeat] Process: Add hash of executable #11722

Merged
merged 13 commits into from
May 9, 2019
Merged

Conversation

cwurm
Copy link
Contributor

@cwurm cwurm commented Apr 9, 2019

Adds the hash(es) of the process executable to process.hash.*. The default is to add SHA-1 only as process.hash.sha1.

Design

Re-using the hashing code from file_integrity (see next section for some of the copied places) introduces a FileHasher type in a new package auditbeat/helper/hasher. Configured using its own Config and created using NewFileHasher(config, done), it has only one other useful method, HashFile(path) that does the hashing incl. throttling.

In the future, we could re-use this hasher in other places (e.g. auditd.paths, Winlogbeat's process.[parent.]executable, the add_process_metadata processor).

Executable hashing can be quite expensive. When hashing all execs of all about 450 running processes on my macOS it takes about 10 seconds. That's why I moved the hashing logic to an enrichProcess() function that is only called when a new process is discovered. When the dataset is run for the first time, it will take a bit, but afterwards, it should be much better.

Code

The code is mostly copied from the file_integrity module:

Configuration

3 new parameters, same as for file_integrity just with a process.hash prefix:

  • process.hash.hash_types (default [sha1])
  • process.hash.max_file_size (default 100 MiB)
  • process.hash.scan_rate_per_sec (default 50 MiB)

Follow up

Two immediate ones:

  1. Refactor the file_integrity module to use the new FileHasher. I tried doing this in this PR, but it became quite complex surprisingly quickly. The hashing logic is strewn across various files and types.
  2. Introduce a hash_max_delay (naming tbd) parameter to configure how long to wait for a hashing to happen (based on scan_rate_per_sec which we honor here even for regular process events, something we might want to do in file_integrity as well). At the moment, it seems the hashing time is essentially unbounded - there is max_file_size, but given enough files hashing could take over CPU and I/O.

@andrewkroh @adriansr - adding you both as reviewers since you're both familiar with the hashing code from file_integrity.

@elasticmachine
Copy link
Collaborator

Pinging @elastic/secops

@cwurm cwurm requested a review from a team as a code owner April 9, 2019 13:58
auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
@cwurm cwurm removed the request for review from a team April 10, 2019 14:14
Copy link
Contributor

@adriansr adriansr left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM except for the CHANGELOG entry. Feel free to ignore the other two comments if you don't have the time to add new behavior or code cleanup. It's a refactor after all.

auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
@@ -104,6 +104,7 @@ https://github.com/elastic/beats/compare/v7.0.0-alpha2...master[Check the HEAD d
- Auditd module: Add `event.outcome` and `event.type` for ECS. {pull}11432[11432]
- Package: Enable suse. {pull}11634[11634]
- Add support to the system package dataset for the SUSE OS family. {pull}11634[11634]
- Process: Hash executable. {pull}11722[11722]
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this is too cryptic. Can you make it more descriptive? "Support hashing the process executable" or similar.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've changed to Process: Add file hash of process executable..

auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
@adriansr
Copy link
Contributor

jenkins, test this.

@codecov
Copy link

codecov bot commented May 1, 2019

Codecov Report

Merging #11722 into master will increase coverage by <.01%.
The diff coverage is 38.59%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master   #11722      +/-   ##
==========================================
+ Coverage   57.25%   57.26%   +<.01%     
==========================================
  Files         901      904       +3     
  Lines       56056    56176     +120     
==========================================
+ Hits        32095    32167      +72     
- Misses      21036    21089      +53     
+ Partials     2925     2920       -5
Impacted Files Coverage Δ
auditbeat/helper/hasher/hasher.go 38.59% <38.59%> (ø)
auditbeat/module/file_integrity/action.go 68.42% <0%> (-12.29%) ⬇️
metricbeat/module/docker/memory/helper.go 89.47% <0%> (-10.53%) ⬇️
libbeat/publisher/queue/memqueue/ackloop.go 93.1% <0%> (-3.45%) ⬇️
metricbeat/module/zookeeper/server/data.go 55.75% <0%> (ø) ⬆️
metricbeat/module/docker/memory/memory.go 70.37% <0%> (ø) ⬆️
metricbeat/module/system/system_other.go 0% <0%> (ø)
metricbeat/module/http/json/data.go 17.14% <0%> (ø)
libbeat/publisher/queue/spool/inbroker.go 68.35% <0%> (+1.17%) ⬆️
... and 14 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update bb3f116...99cd81d. Read the comment docs.

auditbeat/helper/hasher/hasher.go Show resolved Hide resolved
@cwurm
Copy link
Contributor Author

cwurm commented May 2, 2019

  1. Rebased on master to resolve conflicts.
  2. Simplified hash construction - thanks @adriansr!
  3. Moved detailed hash config to auditbeat.reference.yml - same as for file_integrity.

Copy link
Member

@andrewkroh andrewkroh left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. I recommend opening an issue to track the follow up items. It would be nice to be able to share this code with the file_integrity module to reduce duplication and simplify maintenance.


// Config contains the configuration of a FileHasher.
type Config struct {
HashTypes []HashType `config:"hash_types,replace"`
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

q: What's does the ,replace option do? I haven't see that used before.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

replace allows process.hash.hash_types to be empty and disable hashing altogether that way (e.g. if there is a performance problem, or simply to restore current behavior). Only when the config option is not present at all does the default (sha1) take effect.

Without replace, the default value will be used even when the option is present but empty, so there would be no way of disabling it.

I would prefer something called allow_empty or similar, but replace is what I found works as well.

@urso maybe you can take a quick look at whether this is the right use.

@cwurm cwurm merged commit c9ffceb into elastic:master May 9, 2019
@cwurm cwurm deleted the process_hash branch May 9, 2019 20:43
@willemdh
Copy link

willemdh commented May 10, 2019

@cwurm

What kind of hashes will the functionality allow to create? Any chance and imphash is on the roadmap?

"The imphash value can be used to identify malware samples that are related to one another. It can also be utilized to identify similar samples created by a certain threat group."

Maybe not ideal for high granluarity process auditing, but combine this with the file_integrity module and you could kind of create your own antivirus with VirusTotal imphash lookups in a Logstash pipeline

https://www.virustotal.com/intelligence/help/

image

ph pushed a commit to ph/beats that referenced this pull request May 21, 2019
Adds the hash(es) of the process executable to `process.hash.*`. The default is to add SHA-1 only as `process.hash.sha1`.
ph added a commit that referenced this pull request May 22, 2019
* Use time.Duration directly in GetStartTimeEndTime function (#12033)

* Remove convertPeriodToDuration and use duration directly in GetStartTimeEndTime

* Pass period in time.Duration type

* Fix memory leak in Filebeat pipeline acker (#12063)

* Fix memory leak in Filebeat pipeline acker

Before this change acker goroutine was kept forever as processed events
count was not correctly updated.

Filebeat sends an empty event to update file states, this event is not
published, but treated as dropped, without updating counters.

This change makes sures that `a.events` count gets updated for dropped
events also, so the acker gets closed after all ACKs happen.

* Add convert processor (#11686)

The `convert` processor converts a field in the event to a different type, such
as converting a string to an integer. For a full description of the processor's
capabilities see the included documentation.

Closes #8124

* Update docs.asciidoc (#11852) (#12045)

* Update docs.asciidoc

Added in a short note on a few of the other configurable variables.

* Make update

* Remove systemd v233 requirement because it's no longer true (#12076)

* Don't generate autodiscover config when no port matches host hints (#12086)

On metricbeat, when the host autodiscover hint is used, and it includes
the port, one of the exposed ports has to match with the one in the
hint. If not, no configuration should be generated. If it is generated,
it will have empty hosts, what would lead to unexpected errors as the
seen in #8264.

* Fix transptest testing (#12091)

Rewrite test for more clarity ensuring that the server and listeners are
stopped before the test function returns (cause for the case condition).
Each test has it's very own socks5 server now. Tests not requiring a
proxy don't spin up one.

* Refactor and add tests for template and ilm handling. (#12065)

Refactor and add more tests related to setting up template and ilm policy. Ensure template and ilm-policy is exported despite enabled=false setting.

* [Auditbeat] Login: Fix re-read of utmp files (#12028)

The `login` dataset is not using the previous file offset when reading new entries in a utmp file. As a result, whenever a new login event occurs, all records are re-read.

Also expands the documentation, moves test files to testdata/, and adds a test case that adds a utmp record to the test file and re-reads it to make sure this bug does not happen again.

* [Metricbeat][postgresql] Update lib/pq to fix #11393 (#12094)

This updates github.com/lib/pq to the latest version to fix a socket
leak when SSL is not enabled on the PostgreSQL server.

Fixes #11393

* Make breaking changes separate files (#12002)

* Change image references to use block syntax not inline (#11911)

* Change image references to use block syntax not inline

* Fix format for inline links

* Revert "Make breaking changes separate files (#12002)" (#12116)

This reverts commit c17586a.

* fix queue.spool.write.flush.events config type (#12080)

* Revert printing template and policy name on export. (#12067)

Ensures behavior on running the `export` cmd does not change compared to last released version.

* [Filebeat] Add -expected files by default (#12041)

So far expected files in Filebeat tests were only generated and compared when a file exists. This changes to create a generated for all example logs. This will add a few more files to the repository but I think there the benefits outweight the costs as it means the modules are tested in more detail. Also minor changes will be detected easier.

* Update vendored gosigar to 0.10.2 (#12101)

#11924 duplicated some code from go-sysinfo that is affected by a memory leak (fixed in #12100)

In this case, only master is affected as the PR that introduced the leaky gosigar wasn't backported.

* Fix various memory leaks under Windows (#12100)

A function in go-sysinfo used under Windows to split a command-line into arguments was leaking memory.

This code was used in various places:

add_process_metadata processor. (enabled by default in all Beats)
Packetbeat's process monitor. (disabled by default).
Auditbeat's system/process metricset.
This PR updates vendored go-sysinfo to the most recent version, which fixes the leak.

* New processor extract_array (#11761)

This adds a new processor, extract_array, that allows accessing values
inside arrays and copying them to target fields.

* [Heartbeat] Remove not needed flags from setup command (#11856)

The setup command until now contained all the possible options from the other Beats. As Heartbeat does not ship anymore with dashboards, the --dashboards command is not needed anymore and is only confusing. I also removed all the other commands except `--ilm-policy` and `--template`. I'm not aware that `--pipelines` or `--machine-learning` would be used.

Here the comparison between `./heartbeat setup -h` from before and after.

Before:

```
This command does initial setup of the environment:

 * Index mapping template in Elasticsearch to ensure fields are mapped.
 * Kibana dashboards (where available).
 * ML jobs (where available).
 * Ingest pipelines (where available).
 * ILM policy (for Elasticsearch 6.5 and newer).

Usage:
  heartbeat setup [flags]

Flags:
      --dashboards         Setup dashboards
  -h, --help               help for setup
      --ilm-policy         Setup ILM policy
      --machine-learning   Setup machine learning job configurations
      --pipelines          Setup Ingest pipelines
      --template           Setup index template
```

After:

```
This command does initial setup of the environment:
 * Index mapping template in Elasticsearch to ensure fields are mapped.
 * ILM Policy

Usage:
  heartbeat setup [flags]

Flags:
  -h, --help         help for setup
      --ilm-policy   Setup ILM policy
      --template     Setup index template
```

In this PR I did not include a check for the config option `setup.dashboards` to make sure they are not there like apm-server does (https://github.com/elastic/apm-server/blob/2baefab778fdfe70c47bc2fb488677b2e43e4635/beater/beater.go#L60) as I don't think it's necessary.

* Skip Windows testing if magefile.go does not exist (#12099)

Changes the jenkins_ci.ps1 script to skip testing when magefile.go does not
exist. This will allow us to add projects like x-pack/winlogbeat to the test
matrix because not all branches have an x-pack/winlogbeat/magefile.go
file.

* Fix goroutine leak on initialization failures of log input (#12125)

Outlets are created during log input initialization, and if it
fails they were never freed. Handle this case.

* Document and improve permission checks when running socket metricset from Docker (#12039)

Update instructions for system/socket metricset on Docker. And base
permission checks on capabilities rather than on the effective uid.
Running a process as root doesn't mean that it has all privileges,
specially when run as container.

* [metricbeat] added CPU usage check to docker memory stats (#12062)

* Change type from scaled_float to long and add format (#11982)

* Change type from scaled_float to long and add format

* [libbeat] Add unit tests for libbeat's client proxy settings (#12044)

These tests set up server listeners and create libbeat clients with varying proxy settings, and verify that the clients ping the correct target URL.

This is a preparation for #11713, since most of the logic (and work) is in testing the proxy settings; the much simpler PR adding the proxy-disable flag will be a followup to this one, to keep the functional changes isolated in case of rollbacks etc.

* [Metricbeat](Etcd-Leader)Followers wont report leader metrics (#12004)

* manage leader metricset so that followers don't report errors nor events
* add debug message when skipping leader events from non leader members

* Add package libbeat/common/cleanup (#12134)

* Add package libbeat/common/cleanup

The cleanup package adds helpers for deferred optional cleanup on
errors.

For example:

```
ok := False
defer cleanup.IfNot(&ok, func() { ... })

// continue initialization

ok = True
return // some value
```

* Add changelog entry

* [docs] add make fmt to contributing guide (#12118)

* Move one changelog entry from breaking change to bug fix (#12146)

* Sysmon and Security "modules" for Winlogbeat (#11651)

Add pipelines for the Security and Sysmon event logs

The Security processor handles just three events to start with - 4624, 4625, and 4648.
These are event.category=authentication events.

The Sysmon processor handles all event IDs found in the sysmon manifest (sysmon -s).
It moves the event_data fields that are in ECS and does some type conversions.

* [Auditbeat] Process: Add hash of executable (#11722)

Adds the hash(es) of the process executable to `process.hash.*`. The default is to add SHA-1 only as `process.hash.sha1`.

* [Docs] Comment out section that contains bad link (#12152)

* [Filebeat] Introduce UTC as default timezone for modules tests (#12120)

Currently all our modules have convert_timezone disable by default. The reason in 6.x for this was probably that 6.0 did not support convert_timezone and we did not want to introduce a breaking changes. New modules should have convert_timezone enabled by default.

If a module has convert_timezone enabled by default the tests will fail as it takes the timezone of the local computer. To circumvent this, this PR sets the timezone of the tests to UTC so the same time zone is always used.

No generated files were changed in this PR as all modules have convert_timezone off by default. But it will affect #12079 and #12032

* Add number of goroutines to reported metrics (#12135)

* Add minimal ES template functionality. (#12103)

When loading a template without fields, create a minimal template only applying given configuration, without any default values for mappings and settings. This allows to create additional templates only defining specific values.

* Refactor logging in pgsql module (#12151)

Guard debug logging statements with "isDebug" checks. And switch the module over to using named loggers.

Fixes #12150

* Ignore doc type in ES search API for ES 8 (#12171)

* [Docs] Make breaking changes separate files for each version (#12173)

* [Filebeat] module for palo_alto (pan-os) logs (#11999)

This is a module for Palo Alto Networks PAN-OS logs received via Syslog.

It has been tested with logs for PAN-OS version 7.1 to 9.0. However, it is expected to work with earlier versions as the log format is compatible.

* [cmd setup] Add and deprecate setup cmds for index handling (#12132)

Deprecate `setup --template` and `setup --ilm` in favour of newly introduced `setup --index-management` command.
Fix bug in template and write alias creation order to ensure creating properly managed indices. 

implements #12095

Co-authored-by: [email protected]

* Zdd zfs beat (#12136)

Querying ZFS Storage and Pool Status

* [metricbeat] Expand metricbeat dev guide for testing (#12105)

* update dev guide to add examples and expand testing

* Bugfix set template.order to 1 by default. (#12160)

To ensure default order is not changed to 0, set it to 1 in default config.

* [Auditbeat] Fix issues with multiple calls to rpmReadConfigFiles (#12168)

This patch fixes two issues in Auditbeat's system/package on RPM
distros:

- Multiple calls to rpmReadConfigFiles lead to a crash (segmentation
  fault). It is necessary to call rpmFreeRpmrc after each call to
  rpmReadConfigFiles.

  See [1] for a similar issue.

- In addition, it is also necessary to call rpmFreeMacros (when
  available) to avoid leaking memory after each
  rpmReadConfigFiles call.

1: https://lists.fedorahosted.org/pipermail/anaconda-patches/2015-February/015826.html

Fixes #12147

* [Filebeat] Palo_alto module improvements (#12182)

This PR adds some missing features to the recently merged palo_alto module:

Dashboards (One for traffic logs, one for threats).
Sets network.type to either ipv4 or ipv6.
Renames palo_alto.pan_os.threat_file_or_url to palo_alto.pan_os.threat.resource.
Splits palo_alto.pan_os.threat_id into palo_alto.pan_os.threat.id and palo_alto.pan_os.threat.name.

* Add mesosbeat to the community beats (#12185)

* Missing module.yml.disabled file for palo_alto (#12191)

* [Filebeat] Add RabbitMQ module (#12032)

* Parses single and multiline messages from [RabbitMQ's unified log file](https://www.rabbitmq.com/logging.html#log-file-location)
* `var.paths` uses the `RABBITMQ_LOGS` env variable if present
* `var.convert_timezone` (enabled by default since the [timestamps are localtime by default](https://github.com/erlang-lager/lager#universal-time)) 
* No dashboard
* Tested with
  ```
  ~/go/src/github.com/elastic/beats/x-pack/filebeat [master @ 092a3f5] ✓ ‣ MODULES_PATH=./module GENERATE=1 INTEGRATION_TESTS=1 TESTING_FILEBEAT_MODULES=rabbitmq nosetests tests/system/test_xpack_modules.py -v
  test_fileset_file_0_rabbitmq (test_xpack_modules.XPackTest) ... ok

  ----------------------------------------------------------------------
  Ran 1 test in 3.320s

  OK
  ```

* [metricbeat] Add linux sockstat data to socket_summary metricset (#12050)

* add sockstat data to socket_summary

* Fix goroutine leak on non-explicit finalization of log inputs (#12164)

If log inputs were finished because their context, or one of their
ouleters have been finished, then it wasn't stopping its harvesters,
leaking resources.

* [Metricbeat] CoreDNS module: Add Kibana Dashboard (#11619)

* [Metricbeat] CoreDNS module: Add Kibana Dashboard

Fixes #10432.

* [Libbeat][Metricbeat]Add IgnoreAllErrors to schema.Conv object (#12089)

* schema conv option for ignore all errors

Co-Authored-By: Jaime Soriano Pastor <[email protected]>

* Add `container` input, deprecate `docker` in favor of it (#12162)

Add `container` input, deprecate `docker` in favor of it

This change adds a new container input for better support of CRI based
scenarios.

`docker` input was acting as a catch all input for all container related
cases, but its config options were very opinionated towards Docker, with
some issues:

 * `containers.ids` setting was good to abstract logs path, but we have
 seen many cases were logs are not under default location, or follow a
 different path pattern (ie CRI logs).
 * `containers.*` settings have shown counter intuitive for many users,
 in many cases we have seen people writing `container.*` instead, ending
 up in a config error.
 * Some existing settings (`combine_partials`, `cri.parse_flags`) were
 introduced as a way to offer a backwards compatible upgrades, but it
 doesn't really make sense to disable them, as they handle actual
 format behaviors.

This new `container` input offers the same wrapper to read log files
from containers with the following changes:

 * It exposes `paths` as the `log` input, instead of `containers.ids`
 and `containers.path`.
 * `parse_flags` and `combine_partials` are hardcoded, as there is no
 good reason to disable them.
 * `stream` selector is still available, under root settings.
 * It allows to select the log format (also atodetect it), giving room
 for future format changes. `format` can be `auto` (default), `docker`
    and `CRI`.

Example configurations:

Get Docker logs:

```
filebeat.inputs:
- type: container
  paths:
    - /var/lib/docker/containers/*/*.log
```

Get Kubernetes logs:

```
filebeat.inputs:
- type: container
  paths:
    - /var/log/pods/*/*/*.log
    # this could also be used:
    #- /var/log/containers/*.log
```

Previous `docker` input is deprecated in favor of this, to be removed in 8.0

* [Auditbeat] Fix formatting of config files on macOS and Windows (#12148)

Fixes formatting of auditbeat.yml and auditbeat.reference.yml across platforms.

* [libbeat] Escape BOM on JsonReader before trying to decode line (#11661)

* fix json bom + testing

* Set beat ID in registries after loading meta file (#12180)

* Reset beat ID in registries in case loaded from meta file

* Set, not reset

* Adding CHANGELOG entry

* [Filebeat] Move dashboards from 8 to 7 directory (#12217)

The CoreDNS and Envoyproxy dashboard were in the 8 instead of 7 directory. This PR fixes this.

* adjust doc
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants