Takeaways from PagerDuty Summit: Tracing + Timely Alerts
Last week, we spoke with a lot of folks at PagerDuty Summit, where we explored the power of error monitoring in a world that’s always on (or on-call).
What a neat tweet.
Throughout our conversations, one thing became clear: developers are seeking out increased efficiency — they want to minimize time to error detection and resolution. They want to know about issues, find the root cause, and fix them quickly so that they can move on to other things — like writing more code.
Here are two important contributors to efficiency that came up many times during PagerDuty Summit.
Tracing
Applications are frequently built on several communicating services. Without a way to monitor errors on each of those services, developers need to manually dig through each set of logs and hope that they locate the errors they’re looking for. Unfortunately, as we all know, digging through logs isn’t ideal for anyone, and they often don’t provide a clear enough picture of what needs to be reproduced. That digging also takes a lot of time and effort away from actual development, as you likely have to reproduce and obtain additional context before you can fully understand what is going on and move onto fixing it.
In an effort to get that time back and avoid navigating through various systems, developers are looking for a way to efficiently trace errors through services to their root. Here’s a look at how you can trace errors through your stack with Sentry using unique identifiers, allowing you to trace the error and pinpoint the service and code behaving unexpectedly.
Alerts
PagerDuty incidents provide immediate visibility into errors. If those alerts are from Sentry, they also uncover error impact and context as to why you’re being alerted. So, instead of reacting to an everything-is-on-fire-right-now series of issues, developer teams can make calculated, proactive choices about next steps.
Our integration with PagerDuty sends alerts for the incident response and intelligence workflows you define. You can associate a single PagerDuty integration key with as many Sentry Projects as you want. Then set up different alert rules within Sentry for each Project to determine the kinds of issues and errors that should trigger PagerDuty.
We’re curious as to which other ways developer teams are working to reduce time to error detection and resolution. If you’re at SpringOne Platform in Austin (or any of our other upcoming events), stop by our booth and let us know.