we have a scenario where we need to aggregate data from several services and show in UI. The current scenario is when an agent logins in, we need to show cases assigned to that agent. Case information needs to be aggregated from several microservices. There would be around 1K cases assigned to agent at a time and all of the needs to be shown to agent so that he can perform sorting based on certain case data.
What be best approach to show data in this scenario? should we do API calls to several services for each case and aggregate and show ? Or there are better approaches to achieve this.
No. You'll certainly not call multiple APIs to aggregate data on runtime. Even if you call the apis parallely, it will be a huge latency.
You need to pre-aggregate the case details and cache them in a distributed caching system (e.g. Redis or memcached) using a streaming platform (e.g. Kafka). Also, store the pre-aggregated case details in a persistent database. Basically, it's a kind of materialized views.
Caching will enable you to serve the case details fast to the user without any noticeable latency. And streaming will help you to keep the cache and DB aggregations updated in a near-real time. Storing the materialized view in database will save you from storing everything in memory. You can use an LRU cache. Only the recently used data will be in cache. If you need to show any case data that is not in cache, you'd read it from database and store it in cache for future requests.
I recommend you read these two Martin Kleppmann articles here and here
Related
We are building an application on top of Siddhi (using the Java library) that allows users to dynamically add rules and have all incoming information going forward be run against those rules. My question is if it's better to have one large app with many queries, streams, windows, and partitions, or to break up each query into it's own application.
We have been including everything in one single Siddhi app (SiddhiAppRuntime), but this is starting to become large and I fear things may start interacting with each other in unintended ways. We are also snapshotting the SiddhiAppRuntime and restoring state whenever our application gets restarted. This could likely lead to massive restores if we have hundreds of pattern queries to re-run.
I am considering making a separate SiddhiAppRuntime from a single SiddhiManager for each query. The benefits (as I see them) would be reduce the risk of unintentional interactions, make each query able to function on its own, and restoring the query after a shutdown should be much simpler since it will only need to restore a single query. Potential downsides could be increased overhead for having potentially hundreds of SiddhiAppRuntimes.
What is considered best practice for our scenario? What will offer better performance, both for running data through the rules and for restoring the rules in the case we have to restart.
(If this is too broad or any clarification is needed I will do my best to update this question accordingly)
From the lengthy description that you have given I assume these rules that users add does not interact with each other meaning rules add by user1 will not be interacting with rules added by user2.
In such a case it is recommended to use different Siddhi Apps(SiddhiAppRuntimes) for each user. This wont add much additional performance overhead as apps wont be interacting with each other. This will improve snapshoting process as we will be taking separate snapshopts per each app.
Also this will make sure you will have clear separation between each collection of rules and will be easily manageable.
I want to build an internal dashboard to show the key metrics of a startup.
All data is stored in a mongodb database on Mongolab (SaaS on top of AWS).
Queries to aggregate datas from all documents take 1-10minutes.
What is the best practice to cache such data and make it immediately available?
Should I run a worker thread once a day and store the result somewhere?
I want to build an internal dashboard to show the key metrics of a startup. All data is stored in a mongodb database on Mongolab (SaaS on top of AWS). Queries to aggregate datas from all documents take 1-10minutes.
Generally users aren't happy to wait on the order of minutes to interact with dashboard metrics, so it is common to pre-aggregate using suitable formats to support more realtime interaction.
For example, with time series data you might want to present summaries with different granularity for charts (minute, hour, day, week, ...). The MongoDB manual includes some examples of Pre-Aggregated Reports using time-based metrics from web server access logs.
What is the best practice to cache such data and make it immediately available?
Should I run a worker thread once a day and store the result somewhere?
How and when to pre-aggregate will depend on the source and interdependency of your metrics as well as your use case requirements. Since use cases vary wildly, the only general best practice that comes to mind is that a startup probably doesn't want to be investing too much developer time in building their own dashboard tool (unless that's a core part of the product offering).
There are plenty of dashboard frameworks and reporting/BI tools available in both open source and commercial flavours.
Server frameworks: Scala, Play 2.2, ReactiveMongo, Heroku
I think I have quite interesting brain teaser for you:
In my trip-planning application I want to display weather forecast on a map(similar to this). I'm using a paid REST service to query weather data. To speed up user experience and reduce costs I plan to cache weather data for each location for one hour.
There are a few not-so obvious things to consider:
It might require to query up to 100 location for weather to display one weather map
Weather must be queried in parallel because it would take too long to query it in serial fashion considering network latency
However launching 100 threads for each user request is not an option as well (imagine just 5 users looking at a map at one time)
The solution is to have let's say 50 workers that query weather for user requests
Multiple users might be viewing the same portion of map
There is a possible racing condition where one location is queried multiple times.
However it should be queried only once and then cached.
The application is running in clustered environment meaning there will be several play instances.
Coming from a Java EE background I can come up with a pretty good solution using the Java EE stack.
However I wonder how to do this using something more natural to Scala/Play stack: Akka. There is an example (google "heroku scala akka") for similar problem but it doesn't solve one issue: Racing condition when multiple users query the same data at once.
How would you implement this?
EDIT: I have decided that the requirement to ensure that weather data is updated only once is not necessary. The situation would happen far too infrequently to be a real problem and all proposed solutions would bring too much overhead and complexity to the system to be viable.
Thanks everyone for your time and effort. I hope answers to this question will help someone in the future with similar problem.
In Akka you can choose from multiple routing strategies. ConsistentHashingRoutingLogic could serve you well in this situation. Since actors are single-threaded you can easily maintain a cache in each actor. This routing logic will assure that two equal messages will always hit the same actor.
Each actor can work in the following way:
1. check local cache (for example apache commons LRUMap)
- if found, return
2. check global cache (distributed memcache or any other key-value store)
- if found, store the result in the local cache and return
3. query the REST service
4. store the result in the global and local caches
You can have a look at this question, which I based my answer on.
I decided that I'll post my JMS solution as well.
Controller that processes the request for weather does following:
Query the DB for weather data. If there are NO locations with out-of-date data reply immediately. Otherwise continue:
Start listening on a topic (explained later).
For each location: Check whether the weather for the location isn't being updated.
If not send a weather update request message to queue.
Certain amount of workers (50?) listen to that queue.
Worker first marks the location weather as being updated
Worker retrieves updated weather and updates the DB.
Worker sends a message to a topic with weather data for that location.
When controller receives (via topic) weather updates for all out-of-date locations, combine it with up-to-date locations and reply.
Can people give me examples of why they would use coreData in an application?
I ask this because most apps are just clients to a central server where an API of some sort gives you the information you need.
In my case I'm writing a timesheet application for a web app which has an API and I'm debating if there is any value in replicating the data structure on my server in core data(Sqlite)
e.g
Project has many timesheets
employee has many timesheets
It seems to me that I can just connect to the API on every call for lists of projects or existing timesheets for example.
I realize for some kind of offline mode you could store locally in core data but this creates way more problems because you now have a big problem with syncing that data back to the web server when you get connection again.. e.g. the project selected for a timesheet no longer exists.
Can any experienced developer shed some light on there experiences on when core data is best practice approach?
EDIT
I realise of course there is value in storing local persistance but the key value of user defaults seems to cover most applications I can think of.
You shouldn't think of CoreData simply as an SQLite database. It's not JUST an SQLite database. Sure, SQLite is an option, but there are other options as well, such as in-memory and, as of iOS5, a whole slew of custom data stores. The biggest benefit with CoreData is persistence, obviously. But even if you are using an in-memory data store, you get the benefits of a very well structured object graph, and all of the heavy lifting with regards to pulling information out of or putting information into the data store is handled by CoreData for you, without you necessarily needing to concern yourself with what is backing that data store. Sure, today you don't care too much about persistence, so you could use an in-memory data store. What happens if tomorrow, or in a month, or a year, you decide to add a feature that would really benefit from persistence? With CoreData, you simply change or add a persistent data store, and all of your methods to get information out or in remain unchanged. The overhead for that sort of addition is minimal in comparison to if you were trying to access SQLite or some other data store directly. IMHO, that's the biggest benefit: abstraction. And, in essence, abstraction is one of the most powerful things behind OOP. Granted, building the Data Model just for in-memory storage could be overkill for your app, depending on how involved the app is. But, just as a side note, you may want to consider what is faster: Requesting information from your web service every time you want to perform some action, or requesting the information once, storing it in memory, and acting on that stored value for the remainder of the session. An in-memory data store wouldn't persistent beyond that particular session.
Additionally, with CoreData you get a lot of other great features like saving, fetching, and undo-redo.
There are basically two kinds of apps. Those that provide you with local functionality (games, professional applications, navigation systems...) and those that grant access to a remote service.
Your app seems to be in the second category. If you access remote services, your users will want to access new or real-time data (you don't want to read 2 week old Facebook posts) but in some cases, local caching makes sense (e.g. reading your mails when you're on the train with unstable network).
I assume that the value of accessing cached entries when not connected to a network is pretty low for your customers (internal or external) compared to the importance of accessing real-time-data. So local storage might be not necessary at all.
If you don't have hundreds of entries in your timetable, "normal" serialization (NSCoding-protocol) might be enough. If you only access some "dashboard-data", you will be able to get along with simple request/response-caching (NSURLCache can do a lot of things...).
Core Data does make more sense if you have complex data structures which should be synchronized with a server. This adds a lot of synchronization logic to your project as well as complexity from Core Data integration (concurrency, thread-safety, in-app-conflicts...).
If you want to create a "client"-app with a server driven user experience, local storage is not necessary at all so my suggestion is: Keep it as simple as possible unless there is a real need for offline storage.
It's ideal for if you want to store data locally on the phone.
Seriously though, if you can't see a need for it for your timesheet app, then don't worry about it and don't use it.
Solving the sync problems that you would have with an "offline" mode would be detailed in your design of your app. For example - don't allow projects to be deleted. Why would you? Wouldn't you want to go back in time and look at previous data for particular projects? Instead just have a marker on the project to show it as inactive and a date/time that it was made inactive. If the data that is being synced from the device is for that project and is before the date/time that it was marked as inactive, then it's fine to sync. Otherwise display a message and the user will have to sort it.
It depends purely on your application's design whether you need to store some data locally or not, if it is a real problem or a thin GUI client around your web service. Apart from "offline" mode the other reason to cache server data on client side might be to take traffic load from your server. Just think what does it mean for your server to send every time the whole timesheet data to the client, or just the changes. Yes, it means more implementation on both side, but in some cases it has serious advantages.
EDIT: example added
You have 1000 records per user in your timesheet application and one record is cca 1 kbyte. In this case every time a user starts your application, it has to fetch ~1Mbyte data from your server. If you cache the data locally, the server can tell you that let's say two records were updated since your last update, so you'll have to download only 2 kbyte. Now you should scale up this for several tens of thousands of user and you will immediately notice the difference of the server bandwidth and CPU usage.
I'm looking to re-code an application to better handle spikes in tweets. I'm moving to Heroku and MongoDB (either MongoLab or MongoHQ) for the database solution.
During certain news events, tweet volume might spike to 15,000 / second. Typically with each tweet, I parse the tweet and store various pieces of data such as user data, etc. My idea is to store the raw tweets in a separate collection, and have a separate process grab raw tweets and parse them. The goal here is when there is a massive spike in tweets, my application isn't trying to parse all of these, but is essentially backlogging the raw tweets in another collection. As the volume slows, the process can take care of the backlog over time.
My question is three fold:
Can MongoDB handle this type of volume with regards to inserts into a collection at a rate of 15,000 tweets per second?
Any idea on the better setup: MongoHQ or MongoLab?
Any feedback on the overall setup?
Thanks!
The write volume that it will handle depends on lots of factors - hardware, indexes, size of each document, etc. Your best bet is to test it in the environment you're planning to use. If the demands of the write load exceed the capacity of a single mongo server, you can always use just multiple shards.
They are very similar, but there are some differences in pricing and the actual site design has a bunch of differences. There's a thread of discussion about it here: https://webmasters.stackexchange.com/questions/20782/mongodb-hosting-mongolab-vs-mongohq-vs-mongomachine
Overall it seems to make sense. Sounds like you will probably want to flesh out some details about how you will be processing the backlog. Will you be polling it by querying periodically, deleting tweets from the backlog as it processes them, etc.
Completely agree on the need to test this. In general, mongo can handle that many writes, but in practice it depends on the size of your set up, other operations, indexes, etc.
I had to do a similar approach for collecting tons of metrics data. I used a lightweight event-machine process to accept incoming requests in parallel, and store them in a simple format, then another process would take those requests and send them up to a central server. The main goal was to make sure no data was lost if the central server was down, but it also allowed me to put in some throttling logic so that the spikes in data wouldn't overwhelm the system.
I'd be interested to see how this works out for you price-wise, vs. a vps like linode. (I'm a huge Heroku fan, but with certain architectures it can get pricey quickly)