April 8 – Upwork Upgrading it Systems
Upwork has grown rapidly over the past few years and our infrastructure has grown with it. However, parts of these legacy systems have caused site outages and slowdowns. It’s time to do something to fix that.
On Saturday, Apr. 8, we are going to do a massive data migration that will improve site speed and stability. However, we can’t do that without taking Upwork’s site and services temporarily offline.
We know how disruptive extended downtime can be and are doing everything we can to make this process as brief and seamless as possible. Over the next week, we’ll notify all users via email about the update.
Here’s a look at what we have planned and how you might be affected.
Upgrades to support your business
Upwork has grown rapidly over the past few years; three million jobs posted every year, worth $1 billion USD. At the same time, we’ve rolled out new services to support your business, such as messaging and video conference/screensharing capabilities.
It takes a lot of resources to power the world’s largest freelancing website, and we’re quickly reaching capacity. That’s an issue that disrupts your ability to get work done.
Right now, our production environment is split between an aging data center and the Amazon Web Services (AWS) region in Northern California. These two environments are connected by a dedicated network link. That link was the source of a lot of instability until we upgraded it last year and built in redundancy—but it isn’t a good long-term solution.
The infrastructure in our legacy data center is aging—and it isn’t aging well. It’s been in production for several years, and as demand has gone up we’ve leaned hard on those systems. We replace failed hard drives in our database servers almost weekly. Many of the site outages and slowdowns we’ve run into are due to hardware failures in our database servers.
In contrast, AWS is a cloud service platform that’s built to be scalable, flexible, and reliable. By migrating our entire production environment to the AWS Northern California region, we’ll be able to serve more customers, expand our services, be more efficient with our resources, and cut back on downtime.
Unfortunately, we can’t make the move and keep Upwork’s database servers live at the same time. But this upgrade will bring us much closer to our goal: To be able to rebuild Upwork’s production environment from scratch, using automation, in any AWS region as needed. That resiliency can help us provide a faster and more stable site.
Working through the outage
On Apr. 8, we’ll start at 15:00 UTC, taking Upwork offline and moving through the migration process. We estimate this will take six to eight hours. We’ll share regular updates through our social media channels and Upwork’s status page.
During that time, you will not be able to access:
- Upwork.com, including your account, payments, time tracking, and job search
- The Upwork desktop app
- Messages—whether through upwork.com, the mobile app, or the desktop app
- The Upwork blog
- The Hiring Headquarters
- The Community forums
- Help pages
- Information via Upwork’s mobile apps
You will be able to log your time via the Upwork desktop app during this time, but the Messages portion of the app will not work. If you plan to use the app, download the latest version (18.104.22.168) and ensure you are logged in prior to 15:00 UTC.
We apologize for any inconvenience this upgrade might cause. Rest assured that our team will be working carefully to get the Upwork site back online as quickly as possible. On Apr. 8, you’ll be able to get the latest updates through our social media channels or Upwork’s status page.
Source – upwork.com (official announcement)