This adds a new mechanism to allow us to handle compatibility issues between Orbit, Fleet Server and Fleet Desktop.
The general idea is to _always_ send a custom header of the form:
```
fleet-capabilities-header = "X-Fleet-Capabilities:" capabilities
capabilities = capability * (,)
capability = string
```
Both from the server to the clients (Orbit, Fleet Desktop) and vice-versa. For an example, see: 8c0bbdd291
Also, the following applies:
- Backwards compat: if the header is not present, assume that orbit/fleet doesn't have the capability
- The current capabilities endpoint will be removed
### Motivation
This solution is trying to solve the following problems:
- We have three independent processes communicating with each other (Fleet Desktop, Orbit and Fleet Server). Each process can be updated independently, and therefore we need a way for each process to know what features are supported by its peers.
- We originally implemented a dedicated API endpoint in the server that returned a list of the capabilities (or "features") enabled, we found this, and any other server-only solution (like API versioning) to be insufficient because:
- There are cases in which the server also needs to know which features are supported by its clients
- Clients needed to poll for changes to detect if the capabilities supported by the server change, by sending the capabilities on each request we have a much cleaner way to handling different responses.
- We are also introducing an unauthenticated endpoint to get the server features, this gives us flexibility if we need to implement different authentication mechanisms, and was one of the pitfalls of the first implementation.
Related to https://github.com/fleetdm/fleet/issues/7929
Configuration and fixes for the Fleet server and frontend to add support
for https://github.com/Uptycs/kubequery.
Co-authored-by: Michal Nicpon <michal@fleetdm.com>
* Bump go to 1.19.1
* Bump remaining go-version to the 1.19.1
* Add extra paths for test-go
* Oops, putting the right path in the right place
* gofmt file
* gofmt ALL THE THINGS
* Moar changes
* Actually, go.mod doesn't like minor versions
- Add a new "Configuration for contributors" doc page. Move settings that are not recommended for production use
- Remove settings modified in the `config` YAML document from the deploying/configuration doc page
- Document all keys in `config` and `teams` YAML documents
- Add comments to several `.go` files and remove unused struct
This adjusts functions related to applying team specs to make sure the `features` key works as expected, this is:
1. Create Team flow
1. if no features are provided, we copy whatever features are set in the global config
2. otherwise, we generate a new config merging whatever is provided with the defaults for new installs.
2. Edit Team flow: we always generate a new config merging whatever is provided with the defaults for new installs.
* Ingest installed Windows updates and store them in the windows_updates table.
* Added config option for enabling/disabling Windows update ingestion and Windows OS vuln. detection.
Related to https://github.com/fleetdm/fleet/issues/7312, the motivation behind these changes is to introduce a way to deprecate configurations in `AppConfig`, while still preserving backwards compatibility.
From the Epic:
> NOTE: `host_settings` is now replaced by `features`. We should still support `host_settings` as an alias to `features` for backwards compatibility, but in our communications, we should use and recommend features as the canonical way forward.
this adjusts the TestInstallerHeadCheck check to use `DoRaw` instead of `Do`, which sends a body with null when nil is passed for the params parameter.
* add google analytics to sandbox instances
* Add serverType variable to frontend handler
* update version of html-webpack-plugin
Co-authored-by: Lucas Rodriguez <lucas@fleetdm.com>
Related to #6894, this entirely replaces FLEET_DEMO with the server config added in #6597
As part of this, I also implemented a small refactor to the integration test suite to allow setting a custom config when the server is initialized.
This commit replaces `os.Setenv` with `t.Setenv` in tests. The
environment variable is automatically restored to its original value
when the test and all its subtests complete.
Reference: https://pkg.go.dev/testing#T.Setenv
Signed-off-by: Eng Zer Jun <engzerjun@gmail.com>
* Add rate limits for device authed endpoints
* Fix lint
* Add missing test
* Fix test
* Increase the quota for desktop endpoints
* Add comment about quota
* Add server support for Fleet Sandbox demo login
This adds an endpoint `/api/latest/fleet/demologin` that provides a
redirect for the fleetdm.com portion of Fleet Sandbox to automatically
log in a user. The username and password must be provided as form
values. The endpoint is only enabled if `FLEET_DEMO=1` is set in the
server environment.
This was tested locally with the following HTML served by `python3 -m
http.server`, and the Fleet server running with `FLEET_DEMO=1
./build/fleet serve --dev`:
```
<!DOCTYPE html>
<body>
<form
method="post"
action="https://localhost:8080/api/latest/fleet/demologin"
id="demologin"
>
<input type="hidden" name="email" value="admin@example.com" />
<input type="hidden" name="password" value="admin123123#" />
<input type="submit"/>
</form>
<script type="text/javascript">
document.forms["demologin"].submit();
</script>
</body>
</html>
```
For Fleet sandbox purposes, the `action` should be set to the correct
hostname for the sandbox instance, while the `email` and `password`
should be set to the same credentials that were provided when creating
the instance.
* lucas comments
* Add integration tests
* Fix status codes and add comments
Co-authored-by: Martin Angers <martin.n.angers@gmail.com>
* Do not use golangci action for better reproducibility
* Add fix to trigger build
* Fix all reported issues
* fix more lint errors
* Add missing import
* Remove unused method
* Remove change not necessary
Related to #6063, this adds a new device API to get an object with boolean values that we can use as feature flags to manage backwards compatibility in Fleet Desktop.
Feature: Improve our capability to detect vulnerable software on Ubuntu hosts
To improve the capability of detecting vulnerable software on Ubuntu, we are now using OVAL definitions to detect vulnerable software on Ubuntu hosts. If data sync is enabled (disable_data_sync=false) OVAL definitions are automatically kept up to date (they are 'refreshed' once per day) - there's also the option to manually download the OVAL definitions using the 'fleetctl vulnerability-data-stream' command. Downloaded definitions are then parsed into an intermediary format and then used to identify vulnerable software on Ubuntu hosts. Finally, any 'recent' detected vulnerabilities are sent to any third-party integrations.
This removes policy information from `GET /api/_version_/fleet/device/{token}` from non-premium Fleet instances.
Starting the server with `./build/fleet serve --dev --dev_license`
```bash
$ curl -s https://localhost:8080/api/latest/fleet/device/1804e808-171f-4dda-9bec-f695b2f2371a | jq '.host.policies // "not present"'
[
{
"id": 3,
"name": "Antivirus healthy (Linux)",
"query": "SELECT score FROM (SELECT case when COUNT(*) = 2 then 1 ELSE 0 END AS score FROM processes WHERE (name = 'clamd') OR (name = 'freshclam')) WHERE score == 1;",
"description": "Checks that both ClamAV's daemon and its updater service (freshclam) are running.",
"author_id": 1,
"author_name": "Roberto",
"author_email": "test@example.com",
"team_id": null,
"resolution": "Ensure ClamAV and Freshclam are installed and running.",
"platform": "darwin,linux",
"created_at": "2022-05-23T20:53:36Z",
"updated_at": "2022-06-03T13:17:42Z",
"response": ""
}
]
```
Starting the server with `./build/fleet serve --dev`
```bash
$ curl -s https://localhost:8080/api/latest/fleet/device/1804e808-171f-4dda-9bec-f695b2f2371a | jq '.host.policies // "not present"'
"not present"
```
This adds a new API client named DeviceClient to server/service, meant to consume device endpoints and be used from Fleet Desktop.
Some of the logic to make requests and parse responses was very repetitive, so I introduced a private baseClient type and moved some of the logic of the existent Client there.
Related to #5685 and #5697
This adds a new device authenticated endpoint, `/api/_version_/fleet/device/{token}/policies` to retrieve the device policies.
An example request / response looks like:
```bash
curl https://localhost:8080/api/latest/fleet/device/1804e808-171f-4dda-9bec-f695b2f2371a/policies
```
```json
{
"policies": [
{
"id": 3,
"name": "Antivirus healthy (Linux)",
"query": "SELECT score FROM (SELECT case when COUNT(*) = 2 then 1 ELSE 0 END AS score FROM processes WHERE (name = 'clamd') OR (name = 'freshclam')) WHERE score == 1;",
"description": "Checks that both ClamAV's daemon and its updater service (freshclam) are running.",
"author_id": 1,
"author_name": "Admin",
"author_email": "admin@example.com",
"team_id": null,
"resolution": "Ensure ClamAV and Freshclam are installed and running.",
"platform": "darwin,windows,linux",
"created_at": "2022-05-23T20:53:36Z",
"updated_at": "2022-05-23T20:53:36Z",
"response": "fail"
}
]
}
```
Related to [#5685](https://github.com/fleetdm/fleet/issues/5685), in another changeset I will be adding "client" endpoints so we can consume this endpoint from Fleet Desktop
this change prevents errors from being automatically cleared once they are read. A new flag `-flush` is introduced to flush errors on read if necessary.
To add support for #400, we're using the macadmins/osquery-extension to gather Google Chrome profiles from hosts.
Under the hood, the extension looks and parses a json file in which Chrome stores a bunch of data. Given that emails are not required to create Chrome profiles, some of the profiles stored in this file and returned by the query we're using contain empty emails.
The idea after this change is to prevent empty emails from being ingested in the first place instead of filtering them after the fact. I have also included a migration to clean the rows with empty email columns.
Fixes#4780
* Fix access control issues with users
* Fix access control issues with packs
* Fix access control issues with software
* Changes suggested by Martin
* All users can access the global schedule
* Restrict access to activities
* Add explicit test for team admin escalation vuln
* All global users should be able to read all software
* Handbook editor pass - Security - GitHub Security (#5108)
* Update security.md
All edits are recorded by line:
395 replaced “open-source” with “open source”
411 replaced “open-source” with “open source”
439 added “the” before “comment”; replaced “repositories,” with “repositories”
445 deleted “being” before “located”
458 added “and” after “PR”
489 replaced “on” with “in”
493 replaced “open-source” with “open source”; Replaced “privileges,” with “privileges”
* Update security.md
line 479
* Update security.md
added (static analysis tools used to identify problems in code) to line 479
* Fix UI
* Fix UI
* revert api v1 to latest in documentation (#5149)
* revert api v1 to latest in documentation
* Update fleetctl doc page
Co-authored-by: Noah Talerman <noahtal@umich.edu>
* Add team admin team policy automation; fix e2e
* Update to company page of the handbook (#5164)
Updated "Why do we use a wireframe-first approach?" section of company.md
* removed extra data on smaller screens (#5154)
* Update for team automations; e2e
* Jira Integration: Cypress e2e tests only (#5055)
* Update company.md (#5170)
This is to update the formatting under "empathy" and to fix the spelling of "help text."
This was done as per @mikermcneil .
This is related to #https://github.com/fleetdm/fleet/pull/4941 and https://github.com/fleetdm/fleet/issues/4902
* fix update updated_at for aggregated_stats (#5112)
Update the updated_at column when using ON DUPLICATE UPDATE so that
the counts_updated_at is up to date
* basic sql formatting in code ie whitespace around operators
* Fix e2e test
* Fix tests in server/authz
Co-authored-by: gillespi314 <73313222+gillespi314@users.noreply.github.com>
Co-authored-by: Desmi-Dizney <99777687+Desmi-Dizney@users.noreply.github.com>
Co-authored-by: Michal Nicpon <39177923+michalnicp@users.noreply.github.com>
Co-authored-by: Noah Talerman <noahtal@umich.edu>
Co-authored-by: Mike Thomas <78363703+mike-j-thomas@users.noreply.github.com>
Co-authored-by: Martavis Parker <47053705+martavis@users.noreply.github.com>
Co-authored-by: RachelElysia <71795832+RachelElysia@users.noreply.github.com>
This optimizes the query in particular for Domain Controllers, where
there are a number of users with no local accounts and don't need to be
searched for software.
See #4261
not set on the INSERT.
- OUT: Only sets the ID on the passed session and returns it. (`CreatedAt`, `AccessedAt`, are not set.)
New version:
```go
func (ds *Datastore) NewSession(ctx context.Context, userID uint, sessionKey string) (*fleet.Session, error) {
sqlStatement := `
INSERT INTO sessions (
user_id,
` + "`key`" + `
)
VALUES(?,?)
`
result, err := ds.writer.ExecContext(ctx, sqlStatement, userID, sessionKey)
if err != nil {
return nil, ctxerr.Wrap(ctx, err, "inserting session")
}
id, _ := result.LastInsertId() // cannot fail with the mysql driver
return ds.sessionByID(ctx, ds.writer, uint(id))
}
```
- IN: Define arguments that are truly used when creating a session.
- OUT: Load and return the fleet.Session struct with all values set (using the `ds.writer` to support read replicas correctly).
PS: The new `NewSession` version mimics what we already do with other entities, like policies (`Datastore.NewGlobalPolicy`).
* geoip wip
* return nil if ip is empty string or if ParseIP returns nil
* add ui component to render geolocation if available, address PR feedback
* render public ip if available
* add changes file, document geoip in deployment guide
* update rest-api docs
* Error early if email already exists
* Update changes/issue-4361-mail-change-should-error
Co-authored-by: Michal Nicpon <39177923+michalnicp@users.noreply.github.com>
* Fix test
* Lint fixes
* Fix another test
* Also check for invites
* Improve error checks
* Update comment
* Update tests
* Fix test
Co-authored-by: Michal Nicpon <39177923+michalnicp@users.noreply.github.com>
* Add CentOS parsing and post-processing in fleet
* Add tests and amend SyncCPEDatabase
* Add test for centosPostProcessing
* Changes from PR comments
* Amend software test
* Fix sync test
* Add index to source and vendor
* Use os.MkdirTemp
* Rearrange migrations
* Regenerate test schema
* Add support for testing migrations (#4112)
* Add support for testing migrations
* Rename migration in tests
* Changes suggested in PR
* Go mod tidy
* Fix OSVersion to work properly when distribution does not follow symver
Certain distros such as ClearLinux or ArchLinux do not use symver or any
dotted versioning scheme for their releases. Archlinux uses the static
string "Rolling" and ClearLinux uses a single build number such as 35550
for their versions.
In Fleet console, this shows up as a string like "Archlinux 0.0.0.0"
which makes very little sense to the user. This change makes it so that
if OSQuery cannot generate a dotted version number, we should instead
use the build id as an opaque string.
* Add /changes
* Add tests for os_version
* fixup
* fixup
* Add platform filters for MDM/Munki/Chrome queries
This should help quiet warnings that users/customers have reported when
these queries try to run on platforms without the macadmins extension
tables.
For #4123
* Improve documentation
* add changes file
* revert doc formatting
* Update tests
* Yet another test fix
Co-authored-by: Tomas Touceda <chiiph@gmail.com>
* Split hosts in bucket of minutes
* New approach on jitter
* Use minutes to define the amount of buckets
* Add logging to jitter hash creation
* Clean up code and remove unused jitter
* Fix test
* Add docs and address review comments
* Address review comments
* Fix typo in doc
Co-authored-by: Tomas Touceda <chiiph@gmail.com>
* Add sentry
* Fix gosum
* More gosum fixes
* Add missing def for config
* Enrich sentry scope a bit
* Add changes file
* Add goroutine safe scope to errors
* Encapsulate sentry logic
* Add documentation for new flag
* Add sentry capturing to crons and other background tasks
* Only send to sentry when enabled
This helps the period stay under the default request timeouts for most
load balancers.
Some default timeouts:
* AWS ALB - 60s
* Nginx - 60s
* GCP LB - 30s
* wip
* Delete targets when deleting teams, hosts, and labels
* Add changes file
* Fix error message
* Remove unused teamsTable
* Cleanup new pack
* Clean new packs at end of test
@Smjert reported instances of Windows Domain Controllers having massive
resource utilization and being killed by the watchdog when running this
query. In his test environment, this new query performs much better.
