We are implementing support for tracking of Mailgun events in our application. We reviewed the proposed event polling algorithm but find ourselves not quite comfortable with it. First, we would prefer not to discard the data that we have already fetched and then retry from scratch after a pause. It is not very efficient and leaves a door open for a long loop of retries, as it is not clear when the loop is supposed to end. Second, the "threshold age" seems to be the key to determine "trustworthiness", but its value is not defined, only a very large "half an hour" is suggested.
It is our understanding that the events become "trustworthy" after some threshold delay, let us call it D_max, when the events are guaranteed to reside in the event storage. If so, we can implement this algorithm in a different way, so that we do not fetch the data that we know are not "trustworthy" and make use of all data which have been fetched.
We would be fetching data periodically, and on each iteration we would:
Make a request to the events API specifying an ascending time range from T_1 to T_2 = now() - D_max. For the first iteration, T_1 can be set to some time in the past, "e.g., half an hour ago". For the subsequent iterations, T_1 is set to the value of T_2 from the previous iteration.
Fetch all pages one by one while the next page URL is returned.
Use all fetched events, as they are all "trustworthy".
My questions are:
Q1: Are there any problems with this approach?
Q2: What is the minimum realistic value of D_max? Obviously, we can use "half an hour" for it, but we would like to be more agile in tracking events, so it would be great to know what is the minimum value we can set it to and still reliably fetch all events.
Thanks!
1: I see no problems with this solution (in fact I'm doing something very similar). I'm also storing ID's of the events to validate I'm not inserting duplicate entries.
2: I've been working through this similar process. Right now I am testing with D_max at 10 minutes.
Additionally, While going through a testing process I'm running an additional task nightly that goes back over the entire day to validate a few things:
Am I missing existing metrics?
Diagnose if there is a problem with the assumptions I've made about D_max.
I have a function which should have two modes of behaviour according to the place where it's called from.
The core functionality is to do an insert into a table in my database, but it has to be done in two different ways.
Normal mode: whenever it's called only one time (outside of a loop)
For example:
//...
myfunc(param1, record); // it should insert a single record into the database
//...
Batch mode: whenever it's called from inside of a loop
For example:
while(...){
myfunc(param1, record);
}
Inside the "while" loop, each time it's called, it only should store the record in a list and when it reaches the end of the loop, it should fetch all records from the list and prepare a "batch" query that inserts all in one go.
I am wondering how to make it to detect from where it's called in order to switch to the corresponding mode and also how to detect that it has reached the end of a loop and from now on, it should start getting records from the list, prepare the query and execute it.
Any tips or suggestions will be highly appreciated!
Thanks heaps!
It is not, in general, possible to tell whether you are being called in a loop, even with full source code access.
You might be able to do something with caching and delaying the actual database insert for a limited time in all cases. Go on caching until you go for x microseconds without a new call, and then insert the cached data.
However, that could give strange effects if you are not in control of all accesses to the database. In particular, you should do your cached inserts any time there is query that might be affected by them, even in a loop.
Sometimes it is useful to cache queries like this in order to minimize the number of database queries. You can have one function that builds a cache and a second function that sends the request and flushes the cache. If you are going to do that, I recommend using the same function for both single-entry and multiple-entries. The pseudocode will look something like this:
Single-entry usage:
myfunc(param1, record); # caches requests
sendRequests(); # sends all cached requests, flushes cache
Multiple-entry usage:
while(...){
myfunc(param1, record);
}
sendRequests();
sendRequest() will send as many queries as it finds: 1 or many. For efficiency, it can format the requests differently based on their size.
I'm using pqxx as an api to access postgres from c++ code. I'm trying to insert a large amount of data and am finding the performance isn't good enough. I've tried several things and nothing is giving the performance that I need.
I've been trying to use pqxx::pipeline to get asyn inserts, but it seams with this, I can either wait until the end of all inserts to commit the transaction in which case I run the risk of losing a very large amount of data if the process crashes before the commit. Or I can commit occasionally (like every 5 minutes) in which case I have a blocking call every 5 minutes which takes quite a large amount of time.
Is there a way to do this without having a transaction, or to have asynchronous commits for my transaction?
How do I run a second batch on a HIT but ensure I have a new set of workers? I want to make a small change to the HIT and start a new batch, but I don't want any of the workers who participated in the first batch to participate in the second.
Your best bet is to add assignments to your existing HITs (you can do this easily through the RUI), which will exclude workers who have already done those HITs.
But, before you do that, you'll need to change the HITs, which is more difficult (but relatively easy through the API using a ChangeHITTypeOfHIT operation for title/description/duration/qualification changes). If you need to change the Question parameter of a HIT (the actual displayed content of the HIT) or the amount it pays (reward), then you need to create new HITs and send them out as a new batch.
To prevent workers from redoing the HITs you can either put a qualification on the HIT and assign all of your current workers to a score below that level.
Or, you can put a note on the HIT saying that duplicate work will be unpaid. If you do this, you should include a link on the HIT that takes workers to a list of past workerids so that they can check whether they've already done the task.
Update 2018:
You don't need to modify the content of the HIT anymore.
Instead, you assign a qualification to the previous participants in a csv sheet.
Then, in the new HIT, you set as requirement that this qualification "has not been granted".
Detailed explanation at the bottom here and step-by-step procedure describe in this pdf.
Long story short, I'm rewriting a piece of a system and am looking for a way to store some hit counters in AWS SimpleDB.
For those of you not familiar with SimpleDB, the (main) problem with storing counters is that the cloud propagation delay is often over a second. Our application currently gets ~1,500 hits per second. Not all those hits will map to the same key, but a ballpark figure might be around 5-10 updates to a key every second. This means that if we were to use a traditional update mechanism (read, increment, store), we would end up inadvertently dropping a significant number of hits.
One potential solution is to keep the counters in memcache, and using a cron task to push the data. The big problem with this is that it isn't the "right" way to do it. Memcache shouldn't really be used for persistent storage... after all, it's a caching layer. In addition, then we'll end up with issues when we do the push, making sure we delete the correct elements, and hoping that there is no contention for them as we're deleting them (which is very likely).
Another potential solution is to keep a local SQL database and write the counters there, updating our SimpleDB out-of-band every so many requests or running a cron task to push the data. This solves the syncing problem, as we can include timestamps to easily set boundaries for the SimpleDB pushes. Of course, there are still other issues, and though this might work with a decent amount of hacking, it doesn't seem like the most elegant solution.
Has anyone encountered a similar issue in their experience, or have any novel approaches? Any advice or ideas would be appreciated, even if they're not completely flushed out. I've been thinking about this one for a while, and could use some new perspectives.
The existing SimpleDB API does not lend itself naturally to being a distributed counter. But it certainly can be done.
Working strictly within SimpleDB there are 2 ways to make it work. An easy method that requires something like a cron job to clean up. Or a much more complex technique that cleans as it goes.
The Easy Way
The easy way is to make a different item for each "hit". With a single attribute which is the key. Pump the domain(s) with counts quickly and easily. When you need to fetch the count (presumable much less often) you have to issue a query
SELECT count(*) FROM domain WHERE key='myKey'
Of course this will cause your domain(s) to grow unbounded and the queries will take longer and longer to execute over time. The solution is a summary record where you roll up all the counts collected so far for each key. It's just an item with attributes for the key {summary='myKey'} and a "Last-Updated" timestamp with granularity down to the millisecond. This also requires that you add the "timestamp" attribute to your "hit" items. The summary records don't need to be in the same domain. In fact, depending on your setup, they might best be kept in a separate domain. Either way you can use the key as the itemName and use GetAttributes instead of doing a SELECT.
Now getting the count is a two step process. You have to pull the summary record and also query for 'Timestamp' strictly greater than whatever the 'Last-Updated' time is in your summary record and add the two counts together.
SELECT count(*) FROM domain WHERE key='myKey' AND timestamp > '...'
You will also need a way to update your summary record periodically. You can do this on a schedule (every hour) or dynamically based on some other criteria (for example do it during regular processing whenever the query returns more than one page). Just make sure that when you update your summary record you base it on a time that is far enough in the past that you are past the eventual consistency window. 1 minute is more than safe.
This solution works in the face of concurrent updates because even if many summary records are written at the same time, they are all correct and whichever one wins will still be correct because the count and the 'Last-Updated' attribute will be consistent with each other.
This also works well across multiple domains even if you keep your summary records with the hit records, you can pull the summary records from all your domains simultaneously and then issue your queries to all domains in parallel. The reason to do this is if you need higher throughput for a key than what you can get from one domain.
This works well with caching. If your cache fails you have an authoritative backup.
The time will come where someone wants to go back and edit / remove / add a record that has an old 'Timestamp' value. You will have to update your summary record (for that domain) at that time or your counts will be off until you recompute that summary.
This will give you a count that is in sync with the data currently viewable within the consistency window. This won't give you a count that is accurate up to the millisecond.
The Hard Way
The other way way is to do the normal read - increment - store mechanism but also write a composite value that includes a version number along with your value. Where the version number you use is 1 greater than the version number of the value you are updating.
get(key) returns the attribute value="Ver015 Count089"
Here you retrieve a count of 89 that was stored as version 15. When you do an update you write a value like this:
put(key, value="Ver016 Count090")
The previous value is not removed and you end up with an audit trail of updates that are reminiscent of lamport clocks.
This requires you to do a few extra things.
the ability to identify and resolve conflicts whenever you do a GET
a simple version number isn't going to work you'll want to include a timestamp with resolution down to at least the millisecond and maybe a process ID as well.
in practice you'll want your value to include the current version number and the version number of the value your update is based on to more easily resolve conflicts.
you can't keep an infinite audit trail in one item so you'll need to issue delete's for older values as you go.
What you get with this technique is like a tree of divergent updates. you'll have one value and then all of a sudden multiple updates will occur and you will have a bunch of updates based off the same old value none of which know about each other.
When I say resolve conflicts at GET time I mean that if you read an item and the value looks like this:
11 --- 12
/
10 --- 11
\
11
You have to to be able to figure that the real value is 14. Which you can do if you include for each new value the version of the value(s) you are updating.
It shouldn't be rocket science
If all you want is a simple counter: this is way over-kill. It shouldn't be rocket science to make a simple counter. Which is why SimpleDB may not be the best choice for making simple counters.
That isn't the only way but most of those things will need to be done if you implement an SimpleDB solution in lieu of actually having a lock.
Don't get me wrong, I actually like this method precisely because there is no lock and the bound on the number of processes that can use this counter simultaneously is around 100. (because of the limit on the number of attributes in an item) And you can get beyond 100 with some changes.
Note
But if all these implementation details were hidden from you and you just had to call increment(key), it wouldn't be complex at all. With SimpleDB the client library is the key to making the complex things simple. But currently there are no publicly available libraries that implement this functionality (to my knowledge).
To anyone revisiting this issue, Amazon just added support for Conditional Puts, which makes implementing a counter much easier.
Now, to implement a counter - simply call GetAttributes, increment the count, and then call PutAttributes, with the Expected Value set correctly. If Amazon responds with an error ConditionalCheckFailed, then retry the whole operation.
Note that you can only have one expected value per PutAttributes call. So, if you want to have multiple counters in a single row, then use a version attribute.
pseudo-code:
begin
attributes = SimpleDB.GetAttributes
initial_version = attributes[:version]
attributes[:counter1] += 3
attributes[:counter2] += 7
attributes[:version] += 1
SimpleDB.PutAttributes(attributes, :expected => {:version => initial_version})
rescue ConditionalCheckFailed
retry
end
I see you've accepted an answer already, but this might count as a novel approach.
If you're building a web app then you can use Google's Analytics product to track page impressions (if the page to domain-item mapping fits) and then to use the Analytics API to periodically push that data up into the items themselves.
I haven't thought this through in detail so there may be holes. I'd actually be quite interested in your feedback on this approach given your experience in the area.
Thanks
Scott
For anyone interested in how I ended up dealing with this... (slightly Java-specific)
I ended up using an EhCache on each servlet instance. I used the UUID as a key, and a Java AtomicInteger as the value. Periodically a thread iterates through the cache and pushes rows to a simpledb temp stats domain, as well as writing a row with the key to an invalidation domain (which fails silently if the key already exists). The thread also decrements the counter with the previous value, ensuring that we don't miss any hits while it was updating. A separate thread pings the simpledb invalidation domain, and rolls up the stats in the temporary domains (there are multiple rows to each key, since we're using ec2 instances), pushing it to the actual stats domain.
I've done a little load testing, and it seems to scale well. Locally I was able to handle about 500 hits/second before the load tester broke (not the servlets - hah), so if anything I think running on ec2 should only improve performance.
Answer to feynmansbastard:
If you want to store huge amount of events i suggest you to use distributed commit log systems such as kafka or aws kinesis. They allow to consume stream of events cheap and simple (kinesis's pricing is 25$ per month for 1K events per seconds) – you just need to implement consumer (using any language), which bulk reads all events from previous checkpoint, aggregates counters in memory then flushes data into permanent storage (dynamodb or mysql) and commit checkpoint.
Events can be logged simply using nginx log and transfered to kafka/kinesis using fluentd. This is very cheap, performant and simple solution.
Also had similiar needs/challenges.
I looked at using google analytics and count.ly. the latter seemed too expensive to be worth it (plus they have a somewhat confusion definition of sessions). GA i would have loved to use, but I spent two days using their libraries and some 3rd party ones (gadotnet and one other from maybe codeproject). unfortunately I could only ever see counters post in GA realtime section, never in the normal dashboards even when the api reported success. we were probably doing something wrong but we exceeded our time budget for ga.
We already had an existing simpledb counter that updated using conditional updates as mentioned by previous commentor. This works well, but suffers when there is contention and conccurency where counts are missed (for example, our most updated counter lost several million counts over a period of 3 months, versus a backup system).
We implemented a newer solution which is somewhat similiar to the answer for this question, except much simpler.
We just sharded/partitioned the counters. When you create a counter you specify the # of shards which is a function of how many simulatenous updates you expect. this creates a number of sub counters, each which has the shard count started with it as an attribute :
COUNTER (w/5shards) creates :
shard0 { numshards = 5 } (informational only)
shard1 { count = 0, numshards = 5, timestamp = 0 }
shard2 { count = 0, numshards = 5, timestamp = 0 }
shard3 { count = 0, numshards = 5, timestamp = 0 }
shard4 { count = 0, numshards = 5, timestamp = 0 }
shard5 { count = 0, numshards = 5, timestamp = 0 }
Sharded Writes
Knowing the shard count, just randomly pick a shard and try to write to it conditionally. If it fails because of contention, choose another shard and retry.
If you don't know the shard count, get it from the root shard which is present regardless of how many shards exist. Because it supports multiple writes per counter, it lessens the contention issue to whatever your needs are.
Sharded Reads
if you know the shard count, read every shard and sum them.
If you don't know the shard count, get it from the root shard and then read all and sum.
Because of slow update propogation, you can still miss counts in reading but they should get picked up later. This is sufficient for our needs, although if you wanted more control over this you could ensure that- when reading- the last timestamp was as you expect and retry.