You're right, I didn't really think through the specific configs that each
sensor has across enrichment and indexing (because even indexing has the
various batch tuning configurations). I was thinking in terms of topology,
which isn't complete. Really it should (probably) be in terms of sensor all
the way through.
Here's some thoughts on that.
- Users have permissions based on sensor throughout (parsing,
enrichment, and indexing).
- Ideally, this is fine-grained enough that administrators might be
able to restrict at each level (e.g. I don't want a user to say the batch
size is either 1 or MAX_INT, that could cause global performance issues
even if it's just on a single sensor).
- Do we need to be more fine-grained in the first cut (e.g. allow
users to do indexing level configs, but
- Global configs also need to be able to have permissions to apply to
them (after all, I don't want a user to change something for everyone with
no notice). I would expect a lot of these to be more admin-level things. If
they aren't, I question if they should actually be in the global config.
- I'm personally fine with either holding off or doing read perms. Right
now, there aren't read perms, and it's obviously much harder for a user to
cause larger scale issues with read perms on configuration itself. I guess
I could potentially see users not wanting other tenants to see their
transforms, but that seems fairly unlikely for most use cases. Read perms
on data an obviously be applied outside of Metron on the data stores
themselves as necessary.
On Mon, Dec 3, 2018 at 4:15 PM Michael Miklavcic <
michael.miklav...@gmail.com> wrote:
> > @Mike Yeah, I agree. The Jira for that doesn't exist yet, pretty much
> pending this exact conversation winding down.
>
> Yup, understood.
>
> On Fri, Nov 30, 2018 at 7:36 AM Justin Leet <justinjl...@gmail.com> wrote:
>
> > To start with, I'm thinking just the configuration, in particular
> anything
> > that touches the ConfigurationsUtils. I think for the first pass, we
> could
> > leave reading from configs out of it and focus on who can write configs
> > (since that's more disruptive to the system).
> >
> > To the best of my knowledge, this is:
> >
> > - Configuration through the Management UI
> > - Configuration through the CLI (e.g. the push script + Stellar + any
> > other scripts that have to do it).
> >
> > In terms of how fine-grained, I would essentially expect this to be at
> the
> > users+groups level and the individual sensor / topology level (e.g.
> parsers
> > can be authorized individually by users/group, indexing is on the whole).
> > Essentially, I don't propose trying to rearchitect things to try to have
> > multiple indexing topologies for complete separation, etc. Just to take
> > the existing setup and be able to apply authorization on top of it.
> >
> > I would expect anything covered by the ZK configs to be done as part of
> > this effort (possibly in stages). As noted, I would expect this to be a
> > feature branch rather than piecemeal replacement.
> >
> > @Mike Yeah, I agree. The Jira for that doesn't exist yet, pretty much
> > pending this exact conversation winding down.
> >
> > On Tue, Nov 20, 2018 at 12:08 PM Michael Miklavcic <
> > michael.miklav...@gmail.com> wrote:
> >
> > > Justin, It probably makes sense to provide a list of these
> configuration
> > > items as subtasks on the FB Jira so that we can crosscheck what entry
> > > points have been implemented against the test scripts. Do you think
> this
> > > will impact streaming enrichments or the profiler at all? That is to
> say,
> > > as Ali asked, just how far are you looking to take the fine grained
> auth
> > > scope for this?
> > >
> > > M
> > >
> > > On Mon, Nov 19, 2018 at 11:37 PM Ali Nazemian <alinazem...@gmail.com>
> > > wrote:
> > >
> > > > Hi Justin,
> > > >
> > > > By configuration do you mean the sensor related configurations only?
> > Are
> > > > you limiting the scope of this activity to the management-UI or also
> > > > Alert-UI as well? For example, defining different roles (pre-defined
> > > > or customizable) and the fine-grained integration with Ranger?
> > > >
> > > > Cheers,
> > > > Ali
> > > >
> > > > On Wed, Nov 14, 2018 at 1:25 AM Justin Leet <justinjl...@gmail.com>
> > > wrote:
> > > >
> > > > > Hi all,
> > > > >
> > > > > Right now, our various configs can be modified by anyone with
> access
> > to
> > > > the
> > > > > various scripts. I'd like to start a discussion around building out
> > > some
> > > > > authorization to be able to add some more fine grained controls
> > around
> > > > > this.
> > > > >
> > > > > Other projects have some variants on how to accomplish this.
> > > Typically,
> > > > > this follows a pattern of calling out to a interface/class that
> takes
> > > in
> > > > > the operation and context (user and other info) and returns
> > true/false
> > > if
> > > > > something is authorized.
> > > > >
> > > > > In my mind, what we would need out of this is
> > > > >
> > > > > - Ability to apply fine grained permissions
> > > > > - The various scripts and UI should flow this authorization
> > > framework.
> > > > > I believe most (if not all) of our configuration flows through
> > > > > ConfigurationsUtils. Anything that doesn't should either be
> > hooked
> > > in
> > > > > or
> > > > > refactored to flow through the same codepaths.
> > > > > - Pluggability. We shouldn't force only one authorizer.
> > > > >
> > > > > In particular, I'm proposing we use Apache Ranger
> > > > > <https://ranger.apache.org/> as a supported authorization
> framework,
> > > > > implementing it alongside the authorization framework to validate
> > what
> > > we
> > > > > build. In my mind, the main catch with Ranger is that, based on my
> > > > > understanding, we won't be able to restrict users with direct
> access
> > to
> > > > > ZooKeeper via it's CLI (e.g. Ranger can't mirror it's ACLs down
> into
> > > ZK's
> > > > > ACLs). I believe this is a reasonable restriction, especially as
> the
> > > > > management UI gets improved to handle more of the configuration
> > burden
> > > > and
> > > > > the number of users with access to ZK CLI begins to decrease.
> Users
> > > can
> > > > > still add ZK ACLs separately to enforce that access there.
> > > > >
> > > > > For anyone not familiar with Ranger, essentially you build a plugin
> > > that
> > > > > hooks into the existing component's authorization framework (e.g.
> for
> > > > > Storm, the plugin runs through the IAuthorizer
> > > > > <
> > > > >
> > > >
> > >
> >
> https://storm.apache.org/releases/1.2.2/javadocs/org/apache/storm/security/auth/IAuthorizer.html
> > > > > >
> > > > > interface, for Yarn it runs through YarnAuthorizationProvider
> > > > > <
> > > > >
> > > >
> > >
> >
> http://hadoop.apache.org/docs/stable/hadoop-yarn/hadoop-yarn-common/apidocs/org/apache/hadoop/yarn/security/YarnAuthorizationProvider.html
> > > > > >).
> > > > > Additionally, Ranger provides auditing capabilities for this
> > > > authorization
> > > > > and has plugins for a good portion of our stack already (so users
> can
> > > > > already setup ACLs on HDFS, Storm, etc.). Checkout the Ranger
> Github
> > > > > <https://github.com/apache/ranger> for a list of the plugins they
> > have
> > > > > built in.
> > > > >
> > > > > What this means for Metron is building out an authorization setup
> > > similar
> > > > > to Storm or Yarn or whatever we choose. We'll want this anyway, to
> > > allow
> > > > > our solution to be pluggable. At that point, we build an
> > > implementation
> > > > of
> > > > > the authorizer compatible with Ranger along with the plugin.
> > > > >
> > > > > I think this could probably be a fairly small feature branch, which
> > I'm
> > > > > suggesting primarily to do the Ranger implementation alongside the
> > > > general
> > > > > authorization work to validate what's being built. I think the
> main
> > > > > tasking would be something similar to:
> > > > >
> > > > > - Build out pluggable authorization for our configs.
> > > > > - This includes testing (and possibly doing something similar to
> > > > Storm,
> > > > > where they have a some testing IAuthorizers, e.g.
> NoopAuthorizer,
> > > > > DenyAuthorizer, etc.)
> > > > > - Ensure that all the code paths consistently flow through this
> > > > > Authorization.
> > > > > - Build a Ranger compatible version of this.
> > > > > - Define the Ranger plugin for this.
> > > > > - Make sure auditing is defined.
> > > > > - Integration testing (particularly with Kerberos. After all, if
> > > they
> > > > > want to do authorization and auditing, they're almost certainly
> > > using
> > > > > Kerberos).
> > > > >
> > > > > Is there anything missing that we'd need or want for this? Are
> there
> > > any
> > > > > other concerns we'd want to make sure are taken care of?
> > > > >
> > > >
> > > >
> > > > --
> > > > A.Nazemian
> > > >
> > >
> >
>
