If you are planning on using a cloud service provider, there are a couple of topics on your mind from day one; is my data going to be secure and is the service going to be available at all times? Launching the Compliance Initiatives section helps us answer those questions, especially when discussing security and privacy of data, but it does not add as much to the availability discussion. That’s why we are happy to announce the relaunch of the OneLogin Availability page.
Whether you call it “availability”, “uptime”, or “status” page, there are two basic reasons why someone navigates to this page; check to see if the service is currently up, and gauge what the service’s track record is in this regard. Our old page met these basic requirements, but we felt we needed to improve upon it and thereby increase our transparency. Here’s what we did:
New availability categories Previously, we only had one category that we reported issues on. This binary approach was doing us a great disservice as we had no clear way of communicating issues that were not actual outages on our uptime page and provided some misleading historical stats. To remediate this, we defined the following categories to clearly differentiate the main types of availability issues you might experience:
- Service Disruption: a majority of end users are unable to connect to or interact with the service
- Partial Service Disruption: some end users are unable to connect to or interact with the service or sub components of the service
- Degraded Performance: some or all end users experience slow response times or timeouts when connecting to or interacting with the core service or sub components of the service
- Maintenance: planned maintenance may impact the availability of certain backend services, but never impacts end users
We had to do some data cleanup to realign incidents based on the new categories, and with the exception of Partial Service Disruptions, they pretty much mapped to how we were already tracking them behind the scenes. Going forward, we will start using the Partial Service Disruptions bucket as well. Needless to say, we made sure that they matched our published service notifications and did not downgrade any issues that were clearly Service Disruptions and impacted uptime.
When you say “uptime”, do you mean “uptime”? Speaking of uptime, we took the opportunity to clearly define on the page how we calculate it, which is a lot simpler to explain now that we have clear buckets to group availability incidents into:
- Uptime is calculated as the percentage of time that the service is available, i.e., no Service Disruptions over a given period.
Our new page reports monthly and historic uptime, in addition to the last 6 and 12 months. It also uses a new calendar format, so you can see uptime as of the previous day. This is a huge improvement over our previous page that only updated once a month.
One stop shopping: current status + historical Previously, we had two distinct sites to report the current status (status.us.onelogin.com and status.eu.onelogin.com) and one for historical data (www.onelogin.com/trust). We wanted to take all guesswork out of it and have a single source for all availability data. The new page accomplishes this and while I have your attention, it’s a good time to remind you that you can subscribe for updates, so you can receive status updates via SMS, RSS feed, Webhooks, or Email.
The Eyes have it Last, but not least, a big part of the page redesign, was just that; changing the visual design to more clearly communicate the data. We wanted to have a clean design that wouldn’t distract the end user from getting the data they needed quickly and with the fewest number of clicks. I think our Design team did a great job meeting this challenge and they were able to incorporate some great ideas already in use by some of the great partners we work with like Slack and Heroku.
This is by no means the end of the redesign, but rather our first iteration of it. As we get customer feedback and consider how we can improve the page, we will continue to tinker and tweak it, just like we do the rest of the site and our service offering.