We just released a new website! Pretty happy with the result.
Our intention here was to focus more on the purpose of PDFShift, which is a technical tool. We also wanted to show to our visitors that we are a reliable service that is here for the long haul.
Hope you'll like it!
PDFShift retweeted
Ever needed to save a webpage as a PDF? Our latest guide has you covered! From browser tricks to advanced API use with @pdfshift , we've got all the tips and tricks. Perfect for offline access or archiving. muzedevelopment.com/blog/gui…
Today at 2:49 UTC, our two APIs encountered an issue caused by a deployment. We got notified on the issue and quickly worked on resolving the problem. The outage lasted for 40 minutes.
We apologize for the troubles during this incident.
Our LinkedIn page is up (that's about time!) - you can follow us at linkedin.com/company/pdfshif… 😇
We had around one hour of downtime today following our work on making our system more redundant. We continued to run some tests after we resolved the issue, which is why our status page shows for longer issues, but the API was back to normal sooner.
That's why you can see a longer downtime per server (for example at SBG) while the API was working properly; we voluntarily killed sbg to ensure the redundancy system was 100% working, which it now is.
We had some issues caused by a script stopping one of our servers. With the redundancy we have, the IPs should have handled that, but some users still got problem accessing our API.
So we updated how we accept incoming and dispatch requests across our network :) All is good now!
Half the processing servers were down Saturday because of an issue at OVH. This did not cause any issues at all on PDFShift because of the way it was built! (redundancy and performances).
We've started migrating to OVH and updated the status page to reflect our servers with direct health checks : status.pdfshift.io/ (the grey bars are because the checks are new)
Now, what we'll do is first reach out to a few key users about this new infrastructure, and see how it operates.
Then, in a few weeks, we'll start balancing the requests between AWS and OVH, and once we are sure the new code is working well, we'll stop AWS.
PDFShift will run on two distinct regions and each region can handle the current load of what we serve at AWS ... TWICE! This should give us plenty of room for the next steps.
Great news! The new servers are up and running (but we haven't changed the DNS yet). You can try with api.sbg.pdfshift.io and api.gra.pdfshift.io.
Yes, we are on two distinct regions now, so one can catch fire ;)




