Summary:
Stolos is a task dependency scheduler that helps build distributed pipelines. It shares similarities with Chronos, Luigi, and Azkaban, yet remains fundamentally different from all three.
The goals of Stolos are the following:
- Manage the order of execution of interdependent applications, where each application may run many times with different input parameters.
- Provide an elegant way to define and reason about job dependencies.
- Built for fault tolerance and scalability.
- Applications are completely decoupled from and know nothing about Stolos.
How does Stolos work?
Stolos consists of three primary components:
- a Queue (stores job state)
- a Configuration (defines task dependencies. this is a JSON file by default)
- the Runner (ie. runs code via bash command or a plugin)
Stolos manages a queuing system to decide, at any given point in time, if the current application has any jobs, if the current job is runnable, whether to queue child or parent jobs, or whether to requeue the current job if it failed.
Stolos "wraps" jobs. This is an important concept for three reasons. First, before the job starts and after the job finishes, Stolos updates the job's state in the queueing system. Second, rather than run a job directly (ie from command-line), Stolos runs directly from the command-line, where it will check the application's queue and run a queued job. If no job is queued, Stolos will wait for one or exit. Third, Stolos must run once for every job in the queue. Stolos is like a queue consumer, and an external process must maintain a healthy number of queue consumers. This can be done with crontab (meh), Relay.Mesos, or any auto scaler program.
Stolos lets its users define deterministic dependency relationships between applications. The documentation explains this in detail. In the future, we may let users define non-deterministic dependency relationships, but we don't see the benefits yet.
Applications are completely decoupled from Stolos. This means applications can run independently of Stolos and can also integrate directly with it without any changes to the application's code. Stolos identifies an application via a Configuration, defined in the documentation.
Unlike many other dependency schedulers, Stolos is decentralized. There is no central server that "runs" things. Decentralization here means a few things. First, Stolos does not care where or how jobs run. Second, it doesn't care about which queuing or configuration backends are used, provided that Stolos is able to communicate with these backends. Third, and perhaps most importantly, the mission-critical questions about Consistency vs Availability vs Partition Tolerance (as defined by the CAP theorem) are delegated to the queue backend (and to some extent, the configuration backend).
Stolos, in summary:
- manages job state, queues future work, and starts your applications.
- language agnostic (but written in Python).
- "at least once" semantics (a guarantee that a job will successfully complete or fail after n retries)
- designed for apps of various sizes: from large hadoop jobs to jobs that take a second to complete
What this is project not:
- not aware of machines, nodes, network topologies and infrastructure
- does not (and should not) auto-scale workers
- not (necessarily) meant for "real-time" computation
- This is not a grid scheduler (ie this does not solve a bin packing problem)
- not a crontab. (in certain cases this is not entirely true)
- not meant to manage long-running services or servers (unless order in which they start is important)
Similar tools out there:
Requirements:
- A Queue backend (Redis or ZooKeeper)
- A Configuration backend (JSON file, Redis, ...)
- Some Python libraries (Kazoo, Networkx, Argparse, ...)
Optional requirements:
- Apache Spark (for Spark plugin)
- GraphViz (for visualizing the dependency graph)
Background: Inspiration
The inspiration for this project comes from the notion that the way we manage dependencies in our system defines how we characterize the work that exists in our system.
This project arose from the needs of Sailthru's Data Science team to manage execution of pipeline applications. The team has a complex data pipeline (build models, algorithms and applications that support many clients), and this leads to a wide variety of work we have to perform within our system. Some work is very specific. For instance, we need to train the same predictive model once per (client, date). Other tasks might be more complex: a cross-client analysis across various groups of clients and ranges of dates. In either case, we cannot return results without having previously identified data sets we need, transformed them, created some extra features on the data, and built the model or analysis.
Since we have hundreds or thousands of instances of any particular application, we cannot afford to manually verify that work gets completed. Therefore, we need a system to manage execution of applications.
Concept: Application Dependencies as a Directed Graph
We can model dependencies between applications as a directed graph, where nodes are apps and edges are dependency requirements. The following section explains how Stolos uses a directed graph to define application dependencies.
We start with an assumption that our applications depend on each other:
Scenario 1: Scenario 2:
App_A App_A
| / \
v v v
App_B App_B App_C
| |
| App_D
| |
v v
App_E
In Scenario 1, App_B cannot run until App_A completes. In
Scenario 2, App_B and App_C cannot run until App_A completes,
but App_B and App_C can run in any order. Also, App_D requires
App_C to complete, but doesn't care if App_B has run yet.
App_E requires App_D and App_B to have completed.
By design, we also support the scenario where one application expands into multiple subtasks, or jobs. The reason for this is that if we run a hundred or thousand variations of the one app, the results of each job (ie subtask) may bubble down through the dependency graph independently of other jobs.
There are several ways jobs may depend on other jobs, and this system captures all deterministic dependency relationships (as far as we can tell).
Imagine the scenario where App_A --> App_B
Scenario 1:
App_A
|
v
App_B
Let's say App_A becomes multiple jobs, or subtasks, App_A_i. And App_B
also becomes multiple jobs, App_Bi. Scenario 1 may transform
into one of the following:
Scenario1, Situation I
becomes App_A1 App_A2 App_A3 App_An
-------> | | | |
+----------+------+---------+
| | | |
v v v v
App_B1 App_B2 App_B3 App_Bn
Scenario1, Situation II
App_A1 App_A2 App_A3 App_An
or becomes | | | |
-------> | | | |
v v v v
App_B1 App_B2 App_B3 App_Bn
In Situation 1, each job, App_Bi, depends on completion of all of App_A's
jobs before it can run. For instance, App_B1 cannot run until all App_A
jobs (1 to n) have completed. From Stolos's point of view, this is not
different than the simple case where App_A(1 to n) --> App_Bi. In this
case, we create a dependency graph for each App_Bi. See below:
Scenario1, Situation I (view 2)
becomes App_A1 App_A2 App_A3 App_An
-------> | | | |
+----------+------+---------+
|
v
App_Bi
In Situation 2, each job, App_Bi, depends only on completion of
its related job in App_A, or App_Ai. For instance,
App_B1 depends on completion of App_A1, but it doesn't have any
dependency on App_A2's completion. In this case, we create n
dependency graphs, as shown in Scenario 1, Situation II.
As we have just seen, dependencies can be modeled as directed acyclic
multi-graphs. (acyclic means no cycles - ie no loops. multi-graph
contains many separate graphs). Situation 2 is the
default in Stolos (App_Bi depends only on App_Ai).
Concept: Job IDs
For details on how to use and configure job_ids, see the section, Job
ID Configuration This section
explains what job_ids are.
Stolos recognizes apps (ie App_A or App_B) and jobs (App_A1,
App_A2, ...). An application, or app, represents a group of jobs. A
job_id identifies jobs, and it is made up of "identifiers" that we mash
together via a job_id template. A job_id identifies all possible
variations of some application that Stolos is aware of. To give some
context for how job_id templates characterize apps, see below:
App_A "{date}_{client_id}_{dataset}"
|
v
App_B "{date}_{your_custom_identifier}"
Some example job_ids of App_A and App_B, respectively, might be:
App_A: "20140614_client1_dataset1" <---> "{date}_{client_id}_{dataset}"
App_B: "20140601_analysis1" <---> "{date}_{your_custom_identifier}"
A job_id represents the smallest piece of work that Stolos can
recognize, and good choices in job_id structure identify how work is
changing from app to app. For instance, assume the second job_id
above, 20140601_analysis1, depends on all job_ids from 20140601
that matched a specific subset of clients and datasets. We chose to
identify this subset of clients and datasets with the name analysis1.
But our job_id template also includes a date because we wish to run
analysis1 on different days. Note how the choice of job_id
clarifies what the first and second apps have in common.
Here's some general advice for choosing a job_id template:
- What results does this app generate? The words that differentiate this
app's results from other apps' results are great candidates for identifers
in a
job_id. - What parameters does this app expect? The command-line arguments
to a piece of code can be great
job_ididentiers. - How many different variations of this app exist?
- How do I expect to use this app in my system?
- How complex is my data pipeline? Do I have any branches in my
dependency tree? If you have a very simple pipeline, you may simply
wish to have all
job_idtemplates be the same across apps.
It is important to note that the way(s) in which App_B depends on
App_A have not been explained in this section. A job_id does not
explain how apps depend on each other, but rather, it characterizes how
we choose to identify a app's jobs in context of the parent and child
apps.
Concept: Bubble Up and Bubble Down
"Bubble Up" and "Bubble Down" refer to the direction in which work and app state move through the dependency graph.
Recall Scenario 1, which defines two apps. App_B depends on
App_A. The following picture is a dependency tree:
Scenario 1:
App_A
|
v
App_B
"Bubble Down"
By analogy, the "Bubble Down" approach is like "pushing" work through a pipe.
Assume that App_A and App_B each had their own job_id queue. A
job_id, job_id_123 is submitted to App_A's queue, some worker
fetches that job, completes required work, and then marks the
(App_A, job_id_123) pair as completed.
The "Bubble Down" process happens when, just before (App_A, job_id_123) is marked complete, we queue (App_B, f(job_id_123))
where f() is a magic function that translates App_A's job_id to
the equivalent job_id(s) for App_B.
In other words, the completion of App_A work triggers the completion
of App_B work. A more semantically correct version is the following:
the completion of (App_A, job_id_123) depends on both the successful
execution of App_A code and then successfully queuing some App_B work.
"Bubble Up"
The "Bubble Up" approach is the concept of "pulling" work through a pipe.
In contrast to "Bubble Down", where we executed App_A first, "Bubble
Up" executes App_B first. "Bubble Up" is a process of starting at
some child (or descendant job), queuing the furthest uncompleted and
unqueued ancestor, and removing the child from the queue. When
ancestors complete, they will queue their children via "Bubble Down" and
re-queue the original child job.
For instance, we can attempt to execute (App_B, job_id_B) first.
When (App_B, job_id_B) runs, it checks to see if its parent,
(App_A, g(job_id_B)) has completed. Since (App_A, g(job_id_B))
has not completed, it queues this job and then removes job_id_B from
the App_B queue. Finally, App_A executes and via "Bubble Down",
App_B also completes.
Magic functions f() and g()
Note that g(), mentioned in the "Bubble Up" subsection, is the inverse of
f(), mentioned in the "Bubble Down" subsection. If f() is a magic function
that translates App_A's job_id to one or more job_ids for App_B, then
g() is a similar magic function that transforms a App_B job_id to one or
more equivalent ones for App_A. In reality, g() and f() receive one
job_id as input and return at least one job_id as output.
These two functions can be quite complex:
- If the parent and child app have the same
job_idtemplate, thenf() == g(). In other words,f()andg()return the samejob_id. - If they have different templates, the functions will attempt to use the metadata available from configuration metadata (ie in TASKS_JSON)
- If a parent has many children,
f(parent_job_id)returns ajob_idfor each child andg(child_id)returns at least 1job_idfor that parent app. This may involve calculating the crossproduct ofjob_ididentifier metadata listed in dependency configuration for that app.- If a child has many parents,
gandfperform similar operations.
- If a child has many parents,
Why perform a "Bubble Up" operation at all?
If Stolos was purely a "Bubble Down" system (like many task dependency
schedulers), executing App_B first means we
would have to wait indefinitely until App_A successfully completed
and submitted a job_id to the queue. This can pose many problems: we
don't know if App_A will ever run, so we sit and wait; waiting
processes take up resources and become non-deterministic (we have no
idea if the process will hang indefinitely); we can create locking
scenarios where there aren't enough resources to execute App_A;
App_B's queue size can become excessively high; we suddenly need a
queue prioritization scheme and other complex algorithms to manage
scaling and resource contention.
If a task dependency system, such as Stolos, supports a "Bubble Up" approach, we can simply pick and run any app in a dependency graph and expect that it will be queued to execute as soon as possible to do so. This avoids the above mentioned problems.
Additionally, if Stolos is used properly, "Bubble up" will never queue particular jobs that would otherwise be ignored.
The tradeoff to this approach is that if you request to run a leaf node of a dependency graph that has never run before, due to bubble up, you will eventually queue all tasks in that tree that have never run before, which may result in quite a bit of computation for one simple request.
Do I need to choose between "Bubble Up" and "Bubble Down" modes?
Stolos performs "Bubble Up" and "Bubble Down" operations simultaneously, and as a user, you do not need to choose whether to set Stolos into "Bubble Up" mode or "Bubble Down" mode, as both work by default. Currently, Stolos does not let users disable "Bubble Up."
The key question you do need to answer is how do you want to start the jobs in your dependency graph. You can start by queueing jobs at the very top of a tree and then "bubble down" to all of your child jobs. You could also start by queueing the last node in your tree, or you can start jobs from somewhere in the middle. In the latter two cases, jobs would "Bubble Up" and then "Bubble Down" as described earlier in this section.
You can even start jobs from the top and the bottom of a tree at the same time, though there is not much point to that. If you have multiple dependency trees defined in the graph, you can start some from the top (and bubble down) and others from the bottom (and bubble up). It really doesn't matter how you queue jobs because whater you choose to do, Stolos will eventually run all parent and then child jobs from your chosen starting point.
Concept: Job State
There are 4 recognized job states. A job_id should be in any one of these states at any given time.
completed-- When a job has successfully completed the work defined by ajob_id, and children have been queued, thejob_idis marked as completed.pending-- Ajob_idis pending when it is queued for work or otherwise waiting to be queued on completion of a parent job.failed-- Failedjob_ids have failed more than the maximum allowed number of times. Children of failed jobs will never be executed.skipped-- Ajob_idis skipped if it does not pass valid_if_or criteria defined for that app. A skipped job is treated like a "failed" job.
Concept: Dependency Graph
An underlying graph representation defines how apps depend on each other. This information is stored in the Configuration Backend.
You should think about the graph stored in the configuration backend as an abstract, generic template that that may define multiple independent DAGs (Directed Acyclic Graphs). Because it's just a template, the graph typically represents an infinite number of DAGS. Here's what we mean:
If you recall the Task_Ai --> Task_Bi relationship, for each ith job_id,
there is a 2 node DAG. If we think about the dependency graph as defined
in the Configuration Backend as a template, we can imagine just saying that
Task_A --> Task_B. Dropping the i term, and assuming that i could be
infinitely large, we have now created an overly simplistic representation of
an infinite number of 2 node DAGs.
Stolos extends this concept quite a further by using components of the
job_id template to define how more intricate dependencies that may exist
between two apps. For instance, if Task_B depends on "all" of Task_A
job_ids, we must be able to enumerate all possible Task_A job_ids. Further
details on the configuration of this graph are in Setup: Configuration
Backends.
Another key assumption of this graph is that it is fully deterministic. This
means that, given an app_name and job_id, we can compute the full DAG
for that job_id. An example of a non-deterministic graph is one where the
parents (or children) of an app change randomly.
Concept: Queue
Every app known to Stolos has a queue. In fact, Stolos can be thought of as an ordered queueing system. When jobs are consumed from the first queue, they queue further work in child queues. Taking this idea one step further, the queue is an area where the pending work for a given App may come from many different DAGs (as defined by the Dependency Graph in Configuration Backend).
The queue means a few things. First, that Stolos will not consume from an app's queue unless explicitly told to do so. Second, it does not manage the queues or solve the auto-scaling problem. There are many other ways to solve this. At Sailthru, we use a tool called Relay.Mesos to autoscale our Stolos queues. Third, Stolos is only as reliable, fault tolerant and scalable as the queuing system used.
Usage: Quick Start
This will get you started playing around with the pre-configured version of Stolos that we use for testing.
git clone [email protected]:sailthru/stolos.git
cd ./stolos
python setup.py develop
export $(cat conf/stolos-env.sh |grep -v \# )
Submit a job_id to app1's queue
stolos-submit -a app1 --job_id 20140101_123_profile
Consume the job
stolos -a app1
Take a look at some examples
Usage: Installation and Setup (Detailed):
-
The first thing you'll want to do is install Stolos
pip install stolos # If you prefer a portable Python egg, clone the repo and then type: # python setup.py bdist_egg -
Next, define environment vars. You decide here which configuration backend and queue backend you will use. Use this sample environment configuration as a guide.
export STOLOS_JOB_ID_DEFAULT_TEMPLATE="{date}_{client_id}_{collection_name}" export STOLOS_JOB_ID_VALIDATIONS="my_python_codebase.job_id_validations" # assuming you choose the default JSON configuration backend: export STOLOS_TASKS_JSON="/path/to/a/file/called/tasks.json" # and assuming you use the default Redis queue backend: export STOLOS_QUEUE_BACKEND="redis" export STOLOS_QB_REDIS_HOST="localhost" -
Then, you need to define the dependency graph that informs Stolos how your applications depend on each other. This graph is stored in the "Configuration Backend." In the environment vars defined above, we assume you're using the default configuration backend, a JSON file. See the links below to define the dependency graph within your configuration backend.
-
Learn about Job Ids
-
Help me define app configuration: Configuration: Apps, Dependencies and Configuration
- Last, create a
job_id_validationspython file that should look like this:
-
Use Stolos to run my applications
$ stolos-submit <args> $ stolos <args>
For help configuring non-default options:
-
Non-default configuration backend: Setup: Configuration Backends
-
Non-default queue backend: Setup: Queue Backends
Usage: Command-line and API
Stolos wraps your application code so it can track job state just before and just after the application runs. Therefore, you should think of running an application via Stolos like running the application itself. This is fundamentally different than many projects because there is no centralized server or "master" node from you can launch tasks or control Stolos. Stolos is simply a thin wrapper for your application. It is ignorant of your infrastructure and network topology. You can generally execute Stolos applications the same way you would treat your application without Stolos. Keep in mind that your job will not run unless it is queued.
This is how you can manually queue a job from command-line:
# to use the demo example, run first:
export $(cat conf/stolos-env.sh |grep -v \# )
stolos-submit -a app1 -j 20150101_123_profile
In order to run a job, you have to queue it and then execute it. You can get a job from the application's queue and execute code via:
stolos -a app1 -h
We also provide a way to bypass Stolos and execute a job directly. This isn't useful unless you are testing an app that may be dependent on Stolos plugins, such as the pyspark plugin.
stolos --bypass_scheduler -a my_app --job_id my_job_id
Finally, there's also an API that exposes to users of Stolos the key features. The API is visible in code at this link. To use the API, try something like this:
$ export $(cat conf/stolos-env.sh |grep -v \# )
$ ipython
In [1]: from stolos import api ; api.initialize()
In [2]: api.<TAB>
# Get details on a function by using the question mark
In [4]: api.maybe_add_subtask?
# If you have GraphViz installed, try this out
In [5]: api.visualize_dag()
Setup: Configuration Backends
The configuration backend identifies where you store the dependency graph. By default, and in our examples, Stolos expects to use a a simple json file to store the dependency graph. However, you can choose to store this data in other formats or databases. The choice of configuration backend defines how you store configuration. Also, keep in mind that every time a Stolos app initializes, it queries the configuration.
Currently, the only supported configuration backends are a JSON file or a Redis database. However, it is also simple to extend Stolos with your own configuration backend. If you do implement your own configuration backend, please consider submitting a pull request to us!
These are the steps you need to take to use a non-default backend:
-
First, let Stolos know which backend to load. (You could also specify your own configuration backend, if you are so inclined).
export STOLOS_CONFIGURATION_BACKEND="json" # defaultOR
export STOLOS_CONFIGURATION_BACKEND="redis"OR (to roll your own configuration backend)
export STOLOS_CONFIGURATION_BACKEND="mypython.module.mybackend.MyMapping" -
Second, each backend has its own options.
-
For the JSON backend, you must define:
export STOLOS_TASKS_JSON="$DIR/stolos/examples/tasks.json" -
For the Redis backend, it is optional to overide these defaults:
export STOLOS_REDIS_DB=0 # which redis db is Stolos using? export STOLOS_REDIS_PORT=6379 export STOLOS_REDIS_HOST='localhost'
-
-
The specific backend you use should have a way of storing and representing data as Mappings (key:value dictionaries) and Sequences (lists)
- If using the JSON representation, you simply modify a JSON file when you which to change the graph.
- The Redis configuration backend (not to be confused with queue backend) may require some creativity on your part as to how to initialize and modify the graph. We suggest creating a json file and uploading it to redis using your favorite tools.
For examples, see the file, conf/stolos-env.sh
Setup: Queue Backends
The queue backend identifies where (and how) you store job state.
Currently, the only supported queue backends are Redis and Zookeeper. By default, Stolos uses the Redis backend, as we have found the Redis backend much more scalable and suitable to our needs. Both of these databases have strong consistency guarantees. If using the Redis backend with replication, be careful to follow the Redis documentation about Replication. Subtle configuration errors in any distributed database can result in data consistency and corruption issues that may cause Stolos to running tasks multiple times or in the worst case run tasks infinitely until the database is manually cleaned up.
These are the steps you need to take to choose a backend:
- First, let Stolos know which backend to load.
export STOLOS_QUEUE_BACKEND="redis" # default
OR
export STOLOS_QUEUE_BACKEND="zookeeper"
- Second, each backend has its own options.
-
For the Redis backend, you may define the following:
export STOLOS_QB_REDIS_PORT=6379 export STOLOS_QB_REDIS_HOST='localhost' export STOLOS_QB_REDIS_DB=0 # which redis db is Stolos using? export STOLOS_QB_REDIS_LOCK_TIMEOUT=60 export STOLOS_QB_REDIS_MAX_NETWORK_DELAY=30 export STOLOS_QB_REDIS_SOCKET_TIMEOUT=15
-
For the Zookeeper backend, you can define:
export STOLOS_QB_ZOOKEEPER_HOSTS="localhost:2181" # or appropriate uri export STOLOS_QB_ZOOKEEPER_TIMEOUT=30
-
For examples, see the file, conf/stolos-env.sh
Configuration: Job IDs
This section explains what configuration for job_ids must exist.
These environment variables must be available to Stolos:
export STOLOS_JOB_ID_DEFAULT_TEMPLATE="{date}_{client_id}_{collection_name}"
export STOLOS_JOB_ID_VALIDATIONS="my_python_codebase.job_id_validations"
STOLOS_JOB_ID_VALIDATIONSpoints to a python module containing code to verify that the identifiers in ajob_idare correct.- See stolos/examples/job_id_validations.py for the expected code structure
- These validations specify exactly which identifiers can be used in
job_id templates and what format they take (ie is
datea datetime instance, an int or string?). - They are optional to implement, but you will see several warning
messages for each unvalidated
job_ididentifier.
STOLOS_JOB_ID_DEFAULT_TEMPLATE- defines the defaultjob_idfor an app if thejob_idtemplate isn't explicitly defined in the app's config. You should havejob_idvalidation code for each identifier in your default template.
In addition to these defaults, each app in the app configuration may also
contain a custom job_id template. See section: Configuration: Apps,
Dependencies and Configuration
for details.
Configuration: Apps, Dependencies and Configuration
This section will show you how to define the dependency graph. This graph
answers questions like: "What are the parents or children for this (app_name,
job_id) pair?" and "What general key:value configuration is defined for this
app?". It does not store the state of job_ids and it does not contain
queuing logic. This section exposes how we define the apps and their
relationships to other apps.
Configuration can be defined in different configuration backends: a json file, Redis, a key-value database, etc. For instructions on how to setup different or custom configuration backends, see section "Setup: Configuration Backends."
Each app_name in the graph may define a few different options:
job_type- (optional) Select which plugin this particular app uses to execute your code. The defaultjob_typeisbash.depends_on- (optional) A designation that a (app_name,job_id) can only be queued to run if certain parentjob_ids have completed.job_id- (optional) A template describing what identifiers compose thejob_ids for your app. If not given, assumes the defaultjob_idtemplate.job_idtemplates determine how work changes through your pipelinevalid_if_or- (optional) Criteria thatjob_ids are matched against. If ajob_idfor an app does not match the givenvalid_if_orcriteria, then the job is immediately marked as "skipped"
Here is a minimum viable configuration for an app:
{
"app_name": {
"bash_cmd": "echo 123"
}
}
As you can see, there's not much to it.
Next up is an example of a simple App_Ai --> App_Bi relationship.
Also notice that the bash_cmd performs string interpolation so
applications can receive dynamically determined command-line parameters.
{
"App1": {
"bash_cmd": "echo {app_name} is Running App 1 with {job_id}"
},
"App2": {
"bash_cmd": "echo Running App 2. job_id contains date={date}"
"depends_on": {"app_name": ["App1"]}
}
}
Next, we will see a slightly more complex variant of a App_Ai --> App_Bi
relationship. Notice that the job_id of the child app has changed, meaning a
preprocess job identified by {date}_{client_id} would kick off a
modelBuild job identified by {date}_{client_id}_purchaseRevenue. The
"autofill_values" section informs Stolos that if ever there is a scenario where
the modelBuild's target is undefined, we can fill in the missing
information using values defined in "autofill_values". For instance, when
preprocess 20150101_123 completes, it should queue modelBuild
20150101_123_purchaseRevenue. If autofill_values was not defined, Stolos
would raise an exeption.
{
"preprocess": {
"job_id": "{date}_{client_id}"
},
"modelBuild": {
"job_id": "{date}_{client_id}_{target}"
"autofill_values": {
"target": ["purchaseRevenue"]
},
"depends_on": {
"app_name": ["preprocess"]
}
}
}
Expanding on the above example, we see a dependency graph demonstrating how
App_A queues up multiple App_Bi. In this example, the completion of a
preprocess job identified by 20140101_1234 enqueues two modelBuild jobs:
20140101_1234_purchaseRevenue and 20140101_1234_numberOfPageviews.
{
"preprocess": {
"job_id": "{date}_{client_id}"
},
"modelBuild"": {
"autofill_values": {
"target": ["purchaseRevenue", "numberOfPageviews"]
},
"job_id": "{date}_{client_id}_{target}"
"depends_on": {
"app_name": ["preprocess"]
}
}
}
The below configuration demonstrates how multiple App_Ai reduce to App_Bj.
In other words, the modelBuild2 job, client1_purchaseRevenue, cannot run
(or be queued to run) until preprocess has completed these job_ids:
20140601_client1, 20140501_client1, and 20140401_client1. The same
applies to client1_numberOfPageviews. However, it does not matter whether
client1_numberOfPageviews or client1_purchaseRevenue runs first. Looking
at this from the other way around, the children of preprocess
20140501_client1 are these two modelBuild2 jobs: client1_purchaseRevenue
and client1_numberOfPageViews. Also, notice that these dates are hardcoded.
If you would like to implement more complex logic around dates, you should
refer to the section, Configuration: Defining Dependencies with Two-Way
Functions.
{
"preprocess": {
"job_id": "{date}_{client_id}"
},
"modelBuild2": {
"job_id": "{client_id}_{target}"
"autofill_values": {
"target": ["purchaseRevenue", "numberOfPageviews"]
}
"depends_on": {
"app_name": ["preprocess"],
"date": [20140601, 20140501, 20140401]
}
}
}
We also enable boolean logic in dependency structures. We will introduce a
concept of a dependency group and then say that the use of a list specifies AND
logic, while the declaration of different dependency groups specifies OR logic.
An app can depend on dependency_group_1 OR another dependency group. Within
a dependency group, you can also specify that the dependencies come from one
different set of job_ids AND another set. The AND and OR logic can also be
combined in one example, and this can result in surprisingly complex
relationships.
Take a look at the below example. In this example, there are several things
happening. Firstly, note that in order for any of modelBuild3's jobs to be
queued to run, either the dependencies in dependency_group_1 OR those in
dependency_group_2 must be met. Looking more closely at
dependency_group_1, we can see that it defines a list of key-value objects
ANDed together using a list. dependency_group_1 will not be satisfied unless
all of the following is true: the three listed dates for preprocess have
completed AND the two dates for otherPreprocess have completed. In summary,
the value of dependency_group_1 is a list. The use of a list specifies AND
logic, while the declaration of different dependency groups specifies OR logic.
Also take note of how target is defined here. The autofill_values for
target are different depending on which dependency group we are dealing with.
If you find that putting target in the depends_on is confusing, we agree! Open
a GH issue if you have suggestions!
{
"preprocess": {
"job_id": "{date}_{client_id}"
},
"modelBuild": {
"job_id": "{client_id}_{target}"
"autofill_values": {
"target": []
},
"depends_on": {
"dependency_group_1": [
{"app_name": ["preprocess"],
"date": [20140601, 20140501, 20140401],
"target": ["purchaseRevenue", "purchaseQuantity"]
},
{"app_name": ["otherPreprocess"],
"target": ["purchaseRevenue", "purchaseQuantity"],
"date": [20120101, 20130101]
}
],
"dependency_group_2": {
"app_name": ["preprocess"],
"target": ["numberOfPageviews"],
"date": [20140615]
}
}
Finally, the last two examples worth exploring are the use of depends_on "all"
and also the use of ranges in autofill_values. Here's an example where App2
jobs are identified by day. Each date, App2 cannot run until all of that
date's App1 jobs run. That includes the cross product of 10 stage_ids with
all even numbered response_ids greater than or equal to 10 and less than 50.
This means job_id like App2 20150101 depends on 200 (10 * 20) App1 jobs.
You can see how complexity builds fairly quickly.
{
"App1": {
"job_id_template": "{date}_{stage_id}_{response_id}",
"autofill_values": {
"stage_id": "0:10",
"response_id": "10:50:2"
}
},
"App2": {
"bash_cmd": "echo Running App 2. job_id contains date={date}"
"job_id_template": "{date}",
"depends_on": {
"app_name": ["App1"],
"response_id": "all",
"stage_id": "all"
}
}
}
There are many structures that depends_on can take, and some are better than
others. We've given you enough building blocks to express many
deterministic batch processing pipelines. That said, there are several things
Stolos does not currently support:
- Complex logic in depends_on (Two-Way functions directly solve this problem)
- Dependency on failure conditions (Two-way functions might be able to solve)
Finally, for more examples, consult stolos/examples/tasks.json
Configuration: Defining Dependencies with a Two-Way Functions
TODO This isn't implemented yet.
Two way functions WILL allow users of Stolos to define arbitrarily complex
dependency relationships between jobs. The general idea of a two-way function
is to define how the job, App_Ai, can spawn one or more children, App_Bj.
Being "two-way", this function must be able to identify child and parent
jobs. One risk of introducing two-way functions is that users can define
non-deterministic dependencies. Do this at your own risk. It's not recommended!
A second risk, or perhaps a feature, of these functions is that one could
define dependency structures that are one directional. For instance, given a
parent app_name and parent job_id components, the function should return
children. It could be crafted such that given children, the function returns
no parents. The details still need to be ironed out.
```
depends_on:
{_func: "python.import.path.to.package.module.func", app_name: ...}
```
Configuration: Job Types
The job_type specifier in the config defines how your application code
should run. For example, should your code be treated as a bash job (and
executed in its own shell), or should it be an Apache Spark (python) job
that receives elements of a stream or a textFile instance? The
following table defines different job_type options available. Each
job_type has its own set of configuration options, and these are
available at the commandline and (possibly) in the app configuration.
For most use-cases, we recommend "bash" job type. However, if a plugin seems particularly useful, remember that running the application without Stolos may require some extra code on your part.
job_type="bash"bash_cmd
job_type="pyspark"pymodule- a python import path to python application code. ie.stolos.examples.tasks.test_task,spark_conf- a dict of Spark config keys and valuesenv- a dict of environment variables and valuesenv_from_os- a list if os environment variables that should exist on the Spark driveruris- a list of Spark files and pyFiles
(Developer note) Different job_types correspond to specific "plugins"
recognized by Stolos. One can extend Stolos to support custom job_types.
You may wish to do this if you determine that it is more convenient have
similar apps re-use the same start-up and tear-down logic. Keep in mind that
plugins generally violate Stolos's rule that it is ignorant of your runtime
environment, network topology, infrastructure, etc. As a best practice,
try to make your plugin completely isolated from the rest of Stolos's codebase.
Refer to the developer documentation for writing custom plugins.
Developer's Guide
Submitting a Pull Request
We love that you're interested in contributing to this project! Hopefully, this section will help you make a successful pull request to the project.
If you'd like to make a change, you should:
- Create an issue and form a plan with maintainers on the issue tracker
- Fork this repo, clone it to you machine, make changes in your fork, and then submit a Pull Request. Google "How to submit a pull request" or follow this guide.
-
Before submitting the PR, run tests to verify your code is clean:
./bin/test_stolos_in_docker # run the tests
- Get code reviewed and iterate until PR is closed
Creating a plugin
Plugins define how Stolos should execute an Application's jobs. By default, Stolos supports executing bash applications and Python Spark (pyspark) applications. In general, just using the bash plugin is fine for most scenarios. However, we expose a plugin api so that you may more tightly couple your running application to Stolos's runtime environment. It might make sense to do this if you want your application to share the same process as Stolos, or perhaps you wish to standardize the way different applications are initialized.
If you wish to add another plugin to Stolos or use your own, please follow these instructions. If you wish to create a custom plugin, create a python file that defines exactly two things:
- It must define a
main(ns)function.nsis anargparse.Namespaceinstance. This function should use the values defined by variablesns.app_nameandns.job_id(and whatever other ns variables) to execute some specific piece of code that exists somewhere. - It must define a
build_arg_parserobject that will populate thens. Keep in mind that Stolos will also populate this ns and it will force you to avoid naming conflicts in argument options.
Boilerplate for a Stolos plugin is this:
from stolos.plugins import at, api
def main(ns):
pass
build_arg_parser = at.build_arg_parser([...])
To use your custom plugin, in your application's configuration, set the
job_type to python.import.path.to.your.module.
Roadmap:
Here are some improvements we are considering in the future:
- Support additional configuration backends
- Some of these backends may support a web UI for creating, viewing and managing app config and dependencies
- We currently support storing configuration in json file xor in Redis.
- A web UI showing various things like:
- Interactive dependency graph
- Current job status