PagerDuty Blog

New Product Launch: PagerDuty Vacations

Today, PagerDuty is pleased to announce our long-awaited solution for making on-call life better: PagerDuty Vacations. PagerDuty Vacations is a revolutionary new approach to managing team health and incentivizing engineers to join on-call rotations.

For people on an on-call rotation, life can be incredibly stressful. Being woken up in the middle of the night, interruptions during family dinners, and canceled weekend plans are just a few of the common ways that being on call can lead to burnout.

We know that one of the best ways to combat burnout is to take a relaxing vacation that melts away all the stress. But knowing who can take a vacation and when they need it can be a daunting task for even the most proactive teams. Why not let PagerDuty do that for you automatically?

PagerDuty Vacations predicts who needs a vacation the most, where they should go, and the best times for them to go. Being on call now becomes your fast track to enjoying tropical beaches, sunny skies, and clear blue waters. With PagerDuty Vacations you’ll have teams lining up to be on call!

How PagerDuty Vacations Works

PagerDuty has historical data gathered across your various teams that captures timestamps for triggered alerts, when incidents were resolved, and who participated in resolution. But PagerDuty Vacations allows you to do so much more.

PagerDuty Vacations uses the latest advancements in machine learning to automatically predict which on-call responders are most in need of a vacation. Using this new feature starts with some simple set up on your part. You provide details about the optional activities that your team enjoys such as eating, sleeping, off-hours entertainment, celebrating birthdays or anniversaries, going on dates, participating in their child’s sporting events, and so forth.

PagerDuty Vacations also collects data via the PagerDuty app, which has been installed on countless mobile devices. With your consent, we can now leverage devices you already use to compile data on how on-call impacts your real-world teams.

We use geolocation data when an incident is acknowledged to determine if the disruption occurs at the movies, in a restaurant, at home, or wherever a responder may be. Thanks to connected wearable fitness devices, we can also find disruptions in sleep patterns, elevated heart rates, and can even tell if an incident made a responder stress eat afterward (weigh in daily for best results). That may sound creepy, but wait until you see the benefits!

We churn all that data into efficient algorithms that track who Has Acknowledged Lots Presently (or a HALP score) to predict who’s most at risk of burnout based on which incidents they’ve recently participated in. Here’s how it works.

  • Pat and Morgan are both engineers who are on call.
  • Morgan has been taking on many high-profile incidents in the last few weeks and has a HALP score of 95.
  • Pat recently came back from PTO. Therefore, Pat has relatively high spirits and a low HALP score of 12.
  • Pat is woken up at 3 a.m. because an entire data center is down until 5 a.m. Pat’s wearable device tells PagerDuty Vacations that Pat didn’t actually get back to sleep until 6 a.m., just 30 minutes before waking up to go to work. PagerDuty Vacations calculates that Pat needs +32 HALP points.
  • Morgan is interrupted at the movies because a deployment to production had a bug. As a secondary responder, Morgan and only spends about 15 minutes on the call. Using geolocation data, we determine the exact theater within the cineplex Morgan was in. With the new Fandango and Metacritic movie data PagerDuty integrations, we can determine what movie was being watched, if any crucial plot points were missed, or if Morgan was able to finish watching an enjoyable movie. PagerDuty Vacations calculates that Morgan needs only +10 HALP points (Metacritic rates the movie 52%, so it wasn’t a big loss).
  • During the team’s weekly operations review, the team sees that Pat’s HALP score is now 44—more stressed than expected for someone fresh off PTO, but within tolerable parameters. Even though Morgan was in a smaller incident, Morgan’s HALP score is now 105—well above the threshold for needing immediate halp! Morgan’s manager decides it’s time for Morgan to go on vacation ASAP.

“I used to think that being on call was terrible. Now, I can’t wait to go to Fiji. I hope I get woken up in the middle of the night at least four times next week!”
– Jonathan Curry

When and Where to Take a Vacation

But everyone knows that going on vacation just isn’t that simple. Morgan’s team is currently understaffed. Since everyone is overworked while they try to fill open positions, Morgan’s team hasn’t been as diligent about knowledge sharing as they should be. If Morgan goes on vacation right now, it’s possible the team may have to reach out to interrupt a vacation for an incident they don’t know how to deal with otherwise (automatic minimum of +50 HALP points).

Not to worry, PagerDuty Vacations can help there too!

Whenever PagerDuty Vacations selects someone to go on vacation, we use machine learning algorithms to determine when and where they should go. We analyze several factors like a person’s response to incidents across all services, how involved they were during the incident, the overall availability of those services, and other proprietary parameters to predict the best times for an interrupt-free vacation. PagerDuty Vacations suggests the top three time frames least likely to have service interruptions that might require a response from this person.

Unfortunately, the nature of incidents is unpredictable. Although teams try their best not to interrupt vacations, they’re not always successful. With PagerDuty Vacations, your teams can also minimize impact during these extreme cases by intelligently routing vacationers to the best destinations that help ensure the least amount of disruption.

PagerDuty Vacations analyzes when incidents likely to impact a vacationer occur most frequently and then recommends popular destinations in time zones that lead to significantly better outcomes. That 4 a.m. interruption is a lot less terrible if it’s 10 a.m. where you are. PagerDuty Vacations proactively highlights these potential danger zones so that vacation itineraries can be structured around them. Ideally, if and when an alert comes in, you would have just finished breakfast and be taking a break anyway right before your afternoon snorkel. Amazing!

Please note: During the launch beta, PagerDuty Vacations only makes basic recommendations based on time zone. We expect future releases to integrate with TripAdvisor and Expedia to build more complete vacation itineraries. The tentative release date for these integrations is next April Fool’s Day.