OCR 2 Engine service response does not work

Good afternoon.
Yesterday I was doing tests with the OCR 2 engine, since in my company they need an application that needs to read the photo information, for this I use this engine since the first one does not give me the results I expect. Today January 24, 2020 I started testing and the service does not return any response. Is this due to some maintenance or is the link disabled? Does this not happen if you buy the PRO plan ?, since my bosses are thinking about buying the plan.

Sorry about this issue, it should be fixed now.

Such a problem does not happen with PRO, since our PRO OCR API services run on completely different servers, and we have three independent data centers for PRO (US, EU, Asia).

1 Like

Thank you very much, in the same way my bosses will choose to hire the PRO service, so that these incidents do not happen.

I started hitting this issue with the free service a couple days ago. I just upgraded to pro this morning but it’s doing the same thing.

@Pullmyfinger it seems some endpoints are overloaded. We will be adding more servers ASAP. Meanwhile: The PRO3 endpoint looks best right now.

Have you never heard of auto scaling? It’s pretty easy to setup in cloud services like AWS.

OCR Engine 2 is new, and just out of beta. Next bigger update coming soon.

Over the last couple of days all 3 of the PRO end points seem to keep independently going down after 200s+ response times, will this be going on for long?

Sorry about this issue. It is fixed now and with the new server hardware the response times are faster than ever :slight_smile:

See https://status.ocr.space/ for OCREngine1 and (new!) the performance monitoring for OCREngine2 at OCR.Space API Status

Note that even in the “this should never happen” scenario from your screenshot, 3 of 5 endpoints where still up and working fine. If you need 100% uptime and not just 99.9% we always recommend that you add an automatic retry logic to your code. In other words, tell your software to e. g. try endpoint2 if endpoint1 fails. We use the same technique in our own tools, e. g. the OCR API calls in Copyfish (full source code in Github) and our RPA software with screen scraping (full source code in Github) and have had 100% uptime in the last years.

1 Like