How should you handle the fact that events received via webhooks can be received in random order ?
For instance, given the following ordered event:
A: invoiceitem.created (with quantity of 1)
B: invoiceitem.updated (with quantity going from 1 to 3)
C: invoiceitem.updated (with quantity going from 3 to 2)
How do you make sure receiving C-A-B does not result in corrupted data (ie with a quantity of 2 instead of 3)?
You could reject the webhook if the previous_attributes in Event#data do not correspond to the current state, but then you are stuck if your local model was updated already, as you will never find yourself in the state expected by the webhook.
Or you can just use treat any webhook as a hint to retrieve and update an object. You just disregard the data sent by the webhook and always retrieve it.
Even if you receive events ordered as update/delete/create it should work, as update would in fact create the object, delete would delete it, and create would fail to retrieve the object and do nothing.
But it feels like a waste of resources to retrieve data each time when the webhook offers it as event data.
This question was asked before but the answers don't cover the above solutions.
Thanks
If your application is sensitive to changes like this that can occur close in time, you really should just use the event as a signal to retrieve the object, as #koopajah noted in their comment. That's the only way to ensure you have the latest state.
Related
Let's say I have a pull request with three commits, commit-a, commit-b and commit-c.
For every commit, I would like to obtain the aggregate diff within the pull request. That is, when I'm using the GitHub API to request the diff for commit-b I would like to get the changes made in commit-b combined with the changes in commit-a.
One way to do that would be to compare each commit to the state of the pull request before the pull request was opened. That is, compare to the parent-commit of commit-a.
Using the compare commits API, for e.g. commit-b that would be: <base_url>/compare/parent-commit...commit-b.
However, in principle commit-a can have any number of parent commits, and I do not know how to distinguish between them and which was one to use as a reference point.
Hope my question is clear, otherwise I'm happy to provide more details. Thanks!
Let's consider I have 30 items in my db. And clientA will make an api call to get the first 10 records based on item updated time. And think of a use case where clientB updated the 11th record (item) by making some changes in it. But now when clientA makes an api call for next set of items based on the pagination page 2 (items from 11 to 20) It's because the clientB has updated the 11th item the pagination is going to break here (Bases on updated time 11th item will become 1 and 1 become 2, 2 become 3 ...10 becomes 11).There is a chance that clientA is will receive the duplicate data.
Is there any better approach for this kind of problem ??
Any help would be thankfull
I think you could retrieve all elements each time using no pagination at all, to prevent this kind of "false information" at your table.
If visualizing the actual values of each record is mandatory, you could always add a new function to your api working as a trigger. Each time a user modifies any record, this api's function will trigger a message for all active sessions to notify the user some data has been changed. As an example, think about something like the "twitter's live feed". In which when a new bunch of tweets are created, Twitter will notify all users to reload the page if they want to see realtime information.
I'm currently learning ABAP and trying to make an enhancement but have broken down in confusion on how to go about building on top of existing code. I have a program that runs periodically via a background job that disables user accounts X amount of days (in this case 90 days of inactive usage based on USR02~TRDAT).
I want to add an enhancement to notify the User via their email address (result of usr02~bname to match usr21~bname to pass the usr21~persnumber and usr21~addrnumber to adr6 which will point to the adr6~smtp_addr of the user, providing the usr02~bname -> adr6~smtp_addr relationship) based on their last logon date being 30, 15, 7, 5, 3, and 1 day away from the 90 day inactivity threshold with a link to the SAP system to help them reactivate the account with ease.
I'm beginning to think that an enhancement might not be a good idea but rather create a new program and schedule the background job daily. Any guidance or information would be greatly appreciated...
Extract
CLASS cl_inactive_users_reader DEFINITION.
PUBLIC SECTION.
TYPES:
BEGIN OF ts_inactive_user,
user_name TYPE syst_uname,
days_of_inactivity TYPE int1,
END OF ts_inactive_user.
TYPES tt_inactive_users TYPE STANDARD TABLE OF ts_inactive_user WITH EMPTY KEY.
CLASS-METHODS read_inactive_users
IMPORTING
min_days_of_inactivity TYPE int1
RETURNING
VALUE(result) TYPE tt_inactive_users.
ENDCLASS.
Then refactor
REPORT block_inactive_users.
DATA(inactive_users) = cl_inactive_users_readers=>read_inactive_users( 90 ).
LOOP AT inactive_users INTO DATA(inactive_user).
" block user
ENDLOOP.
And add
REPORT warn_inactive_users.
DATA(inactive_users) = cl_inactive_users_readers=>read_inactive_users( 60 ).
LOOP AT inactive_users INTO DATA(inactive_user).
CASE inactive_user-days_of_inactivity.
" choose urgency
ENDCASE.
" send e-mail
ENDLOOP.
and run both reports daily.
Don't create a big ball of mud by squeezing new features into existing code.
From SAP wiki:
The enhancement concept allows you to add your own functionality to SAP's standard business applications without having to modify the original applications. To modify the standard SAP behavior as per customer requirements, we can use enhancement framework.
As per your description, it doesn't sound like a use case for an enhancement. It isn't an intervention in an existing process. The original process and your new requirement are two different processes with some mutual logical part - selection of days of inactivity of users. The two shouldn't rely on each other.
Structurally I think it is best to have a separate program for computing which e-mails need to be sent and when, and a separate program for actually sending them.
I would copy your original program to a new one, and modify it a little bit so that instead of disabling a user, it records into some table for each user: 1) an e-mail 2) a date when to send 3) how many days left (30, 15, 7, etc) 4) status if the e-mail was sent or not. Initially you can even have multiple such jobs for each period (30, 15, 7 etc) and pass it as a parameter (which you use inside instead of 90).
This program you run daily as a job and it populates that table with e-mail "tasks" of what needs to be sent today. It just adds new lines, so lines from yesterday should stay in there.
The 2nd program should just read that table and send actual e-mails and update the statuses. You run that program daily as well.
This way you have:
overview: just check the table to see what's going on
control: if the e-mailer dies or hangs, you can restart it and it will continue where it left off; with statuses you avoid sending duplicate mails
you can make sure that you don't send outdated e-mails if in your mailer script you ignore all tasks older than say 2 days
I want to clarify your confusion about the use of enhancements:
You would want to use enhancements in terms of 'something' happens or is going to happen in the system and you would want to change this standard way.
That something, let's call it event or process could be for example an order is placed, a certain user is logging onto the system or a material has been or is going to be changed.
The change could be notifying another system of an order or checking the logged on user with additional checks for example his GUI version and warn him/her if not up-to-date.
Ask yourself, what process on the system does the execution of your program or code depend on. Does anything need to happen before the program is executed? No, only elapsing time.
Even if you had found an enhancement, you would want to use. If this process using the enhancement would not be run in 90 days, your mails would not be sent, because the enhancement would never been called.
edit: That being said, supposing you mean by enhancement 'building on your existing program' instead of 'creating a new one' would be absolutely not the right terminology for enhancement in the sap universe.
I would extend the functionality of your existing program, since you already compute how many days are left and you would have only one job to maintain.
I am trying to make a notification system with Redis rather than using MySQL which is what I use for the rest of the system. The reason for this is that I don't really need to save that much data so it can be saved in memory and I want it to be lightweight and fast.
The notifications will be kept temporarily. What I mean by that is that I do not want to save all notifications, but more like 50 latest unseen notifications for each user. So first thing I thought about was to use a linked list with a capped length of 50.
I would need to save this information for the notification:
postId
commentId
type
time
userId
username
image
So perhaps a JSON serialized string like this:
{"postId":1,"commentId":10,"type":1,"time":1462960058,"userId":2,"username":"Alexander","image":"ntfpRrgx.png"}
The notifications would be output like this on the client side:
Alexander commented on your post.
Alexander replied to your comment.
Where the type determines what kind of notification it is. I can handle "type" checks client side and output notification format accordingly. But here is the part I am having difficult with.
1) I need to be able to save the notifications in an ordered way so that I know which notification is newest.
2) I need to be able to know when a notification has been seen, so that it is not registered as not seen anymore.
3) I need to have a count of unseen notifications that I can show to the user. And If the user clicks on a notification, I need to mark that as a seen notification and decrement the count of unseen notifications.
4) I need to be able to mark all notifications as marked seen if the user wishes to do that.
5) I need to be able to get a subset of the notifications, whether seen or unseen, like an offset and limit on MySQL. For example, the user sees the newest 5 notifications, but he could click a next button and see the next 5, and the next 5 and so on.
I have no idea how to do all of this on Redis.
The key for the list or set could be user:1:notification. I know a list is sorted, and we can add and remove from the head and tail. But how do I achieve all these points?
1: You can use redis sorted sets (zset) operations and use timestamp as a score, and event id (or the entire event json) as a member.
ZADD my-set-key timestamp event-id
Then to get a page newest items you use zrevrange command. If you choose to use event id as a member, then you need additional structure to store event fields. I would recommend HSET eventid, field, value.
2: You can remove an item by member (event-id)
ZREM my-set-key event-id
3: Assuming your zset only keeps unseen, then you can use ZCARD to get size of the set
ZCARD my-set-key
4: You can remove an entire set in one shot using
DELETE my-set-key
5: You can paginate using zrange/zrevrange:
ZREVRANGE my-set-key start-position to-position
If you need to keep both seen and unseen items, then you need an extra zset where you only add, but don't remove once an item is seen
Thank to the libgit api, I can get the Object corresponding to my entry in the tree.
- Get the last commit (git_reference_peel)
- Get the tree (git_commit_tree)
- Get the tree entry (git_tree_entry_bypath)
- Get the object (git_tree_entry_to_object)
I wonder if there is a way to get the author/signature of the last user who has make a change on this object.
The signature of the commit doesn't my requirement
Best Regards
Vincent
The libgit2 API allows you to retrieve the identity of the person that performed the last changes on an object. However, that requires to orchestrate the calls according to the following algorithm.
Revwalk the history and perform diffs between the current commit and its parents until the file you're tracking is detected as changed. Once the commit is isolated, retrieve the identity of the author of this commit.
In order to provide you with a quick start, I'd suggest you to get some inspiration from the code of pull request #1965 that demonstrates how to do this.