Our Magento store uses the Website Payments Pro payment gateway through PayPal. For the most part this works pretty well. Occasionally (about once a day), we get failed transaction emails and the errors are very generic (i.e. "Internal Error" or "There was an error processing your order. Please contact us or try again later."). I've been working with our PayPal rep, but he has not yet been able to figure out why. This doesn't happen with every transaction, and typically if the customer re-submits the order it will work the second time. This has been very frustrating for me as the developer, my employers, and the customers who see the same generic error message in a popup dialog box.
My questions are as follows:
Is this normal for a site that gets about 300+ orders/day?
Are we at the mercy of bugs in the PayPal system?
Do any of you have experience with similar problems? Or, on the contrary, do you have experience with a site that gets at least as much traffic and never has any errors with PayPal?
I'm at a loss for what to do. I want to believe that it is not an error in what we are sending to PayPal. But if every client using PP as their gateway experienced this much pain I would think someone with a loud enough voice would be pushing for improvements. I don't have any experience with other payment gateways, so perhaps PP is the lesser of all evils...
you certainly need to be prepared for issues like this and error messages for clients need to be clear and informative what to do if error happens.
yes you are and this can't be only paypal issue this can be n of issues with network latency user interaction etc.
there are n amount of sites experiencing this as the system is engineered with lot of break points and you can't assure that issues do not happen on 3rd party system or sites and you can have several precautions to avoid that by doubling the payment methods with several gateways etc.
Related
I'm using Python and Plaid's development environment to download bank balances and transactions. To get the initial access tokens, I'm launching Link from quickstart, and can do that in standard and update mode.
The problem I'm running into is how frequently my API call returns the ITEM_LOGIN_REQUIRED error and I have to re-authenticate. For a Regions account I've been testing, this happens a few times throughout the day. For a Pinnacle Financial Partners bank, this happens almost immediately after updating the access token. As in, I can log in through link, successfully fire an auth/get request, and by the time I can send another request (e.g., balance/get), I'm already getting ITEM_LOGIN_REQUIRED again.
As I'm evaluating Plaid for production use, is this frequent authentication atypical? Is it a known limitation with development, or with specific banks even on production? I've seen some banks (Bank of America) only work in production, but I'm hoping what I'm experiencing is just the nature of working in development. Thanks.
Development vs. Production environments are virtually identical and shouldn't impact how often you hit ITEM_LOGIN_REQUIRED.
What you're seeing is atypical, though. Unless you have multi-factor auth turned on and configured not to trust known devices, this shouldn't happen.
Assuming you don't have that configured, would you mind submitting a support ticket so Plaid Support can look into this and help figure out why it's happening?
I'm having a problem with payment procedure in heroku account(billing section).
I got verified first time successfully and in one month all charges of my card are giving an error: "Unable to verify your card. Please try later or contact your financial institution for help'".
I tried with 10 credit cards, all of which are from Mexico (and I triple checked all the info and I’m sure it’s correct). I also talked to my bank and there is no attempt of charging cards.
Heroku support can't help me out. This is their answer to my ticket "I truly wish I could be of more assistance, but unfortunately I do not have access to anything that would be able to make a difference here. I apologize for the poor experience."
Maybe someone had the same problem?
Thank you.
The answer is ... that noone can help me in Heroku.
Support of Heroku answered: "Unfortunately, the many layers of security make it impossible for us to further investigate. I find it frustrating myself, since my job is to try and help customers and in cases like this I'm told there's nothing I can do. I'm truly sorry for the poor experience here. The only choices are pretty much to use only Free Dynos and no add-ons except the free version of Heroku Postgres (and no custom domains), or to move to a different platform. I apologize that these are really the only choices".
So I'm moving to other platform. Lol
If that many cards are not working, one of the possibilities is that your user footprint (things that can compare your online activity to the other user profiles, like location, e-mail, credit card) could be triggering fraud prevention tools, which can happen if patterns are identified that are similar to ones used by spam or fraudulent accounts. It doesn't necessarily mean that your account is identified as fraudulent or spam, but it does mean that the virtual footprint is considered too risky to approve.
Yes, Braintree and Taxamo have a couple of widgets facilitating their integration, but the way they work they are tied to the process of authorization of a credit card while what the tax services are interested in are SETTLED transactions (i.e. what is actually billed from the client at the end of the day),
settled transactions aren't the same thing as authorized transactions, voids, IO problems in API communication, client-side software bugs, even credit card company policies (chargebacks, presentment rejections etc) result with what is authorized not being equal to what is settled / billed,
special world of pain continues when you consider the complex formulas for mid-cycle subscription updates (proration), discount and addon calculations and the fact that none of my subscription API calls done using the sandbox don't have a transaction object (so I don't know the amounts being billed from the client),
meaning I have to reimplement a bunch of advanced logic done by Braintree to deduce what the client will see on his bank statement and hope to have gotten everything right, or have IRS on our back -_- (in Croatia companies get blocked over 1-3 euro imbalances in the tax reports)
So the question, am I missing something here?
is there a simple way to submit to Taxamo only settled transactions?
does maybe the sandbox and mock webhooks behave differently than production, is it possible that Subscription API calls and webhooks in production return a transaction object and that I just don't see it?
does the current Braintree and Taxamo integration at least "kind of work in practice 100% of the time" and I'm worrying too much?
Full disclosure: I work at Braintree. If you have any further questions, feel free to contact support.
Braintree and Taxamo don't have an official integration package (the widgets available are not supported by Braintree, and Braintree had no hand in developing them). However, the two can be, and often are, used alongside each other. Because there's no official integration, the two integrations don't interact with one another, and the logic concerning when any action in Taxamo should take place in relation to an action in Braintree is up to a user to implement.
With that in mind, there's no 'out-of-the-box' solution to submit only settled transactions to Taxamo. Braintree does offer settlement batch summary reports, which contain the transactions that settle every night. Using the information there, you could submit your transactions to Taxamo every night alongside the settlement batches. However, be aware that transactions don't always settle immediately, so there will be a gap between the time that you submit a transaction for settlement and the time you submit the transaction to Taxamo.
Alternatively, you could store the transaction in Taxamo immediately and keep track of the transaction key that Taxamo generates to reference the tranasction. Then, if anything happens that prevents the transaction from settling, you can delete that transaction in Taxamo.
These are just two options to consider for an open-ended problem. As always, you should evaluate your business needs carefully, and do whatever you believe will work best.
As BladeBarringer mentioned in the comments to your question, the subscription response object does contain an array of transactions that you can reference. The most recent transaction is always at index [0] of that array.
Does anyone have any idea of the average Adaptive Payments application approval time at the moment? It appears there may be a backlog as we've heard nothing from the moderators and it has been over 15 days (their usual timescale apparently). Our case is a crowdfunding website built on CodeIgniter with the Adaptive Payment gateway installed and ready to go live when we are given application approval. Our site has been active for 6months using other methods via PayPal, but it's time to move up a notch.
It depends on the permissions your application is requesting; less restricted permissions usually goes through quicker.
Even so, check with PayPal Developer Technical Services at https://www.paypal.com/dts/ if you're looking for an update.
Does any US wireless carrier offer individuals or companies with a direct connection to the SMSC?
The number is 747-772-3101 (repalce 7's with 6's)
This number is registered to t-mobile, also verified by t-mobile to be a valid subscriber sending 160,000+ text messages monthly and that all they have is an unlimited text messaging plan on top of the cheapest voice plan. This company of the number verified to me that they don't use gsm modems as they are too slow.
So I know it's possible but who would I contact, Sales or anyone else reachable through a 1-800 is ignorant to these services and developer.t-mobile is worthless and doesn't reply to emails.
Any info??
Most likely they are connected to an Aggregator (Sybase 365, Mblox, Netsize, Verisign, etc. Smaller guys like multimode and Clickatell are more open to this) that is connected to T-Mobile. As they have chosen to not use a shortcode they simply buy a regular T-Mobile SIM/MSISDN and use the full longcode as the origination address of the messages.
A lot of companies use aggregators to enable Oracle Applications Server to send SMS messages.
does anybody have more info on this????
i did a little investigation and here is what i have determined. the company that the long code is tied to is broadtexter.com. they offer a free service to people who want to follow bands/comedians/acts/ etc. you basically join their fan club.
when i text help to the phone number it immediately autoresponds from that same number. that means they are either using an agg with a dedicated vmn (totally possible) of they are using a mobile modem with a sim (totally possible as well, and probably cheaper), but basically all they are doing from this phone number is pushing traffic to their website.
once you go to their website and sign up to a fan club; ALL FUTURE COMMUNICATION is mo/mt thru the SMTP gateway. dead giveaway is that they ask for your carrier when you sign up. second dead givaway is the caller id is xxxxxxx#broadtexter.com every time.
so the simple answer is that they are only using the vmn (long code) to drive people to their website to sign up...then all future communication is over SMTP. so the 160K+ messages are occuring thru the SMTP gateway. since they appear to be non-commercial (no ads, no spamming, etc.) and they are somewhat of a peer to peer setup, they probably fly under the radar (or are accepted by) the carriers.
if anyone can offer more insight to this, i would love to read it!
I've got more info, and its kinda blowing my mind. I'm interested because I've got a social networking website and I'd like to set up an interactive SMS service without a shortcode. So I went to one of the profile pages on broadtexter.com and used their flash widget to join that bands club. I entered my mobile number and for provider I choose AT&T - option 1 (there were 3 options listed). Almost immediately I got a SMS asking me to reply with a Y to confirm. Here's the crazy thing: The number it was from was 1 (410) 000-001. At first I thought it was a regular cell number, then I realized it's missing a digit. Unless the area code is actually 141 and my iPhone is just formatting it weird. Except 141 isn't an area code?
Next, I replied with a Y and I got another text that asked me to reply with a photo for my profile (something that would be prefect for my site). This time the text from from 1 (410) 000-002.
So I sent back an MMS with a pic from my iPhone gallery, and I didn't get a reply yet. It's been about 15 minutes. It's kinda weird because to sign into their site, I need a username, which they never gave me. I haven't given them my email yet either. Anyway, I don't really care about that. I just wanna know what's going on with those numbers? How are they doing this?!