At Onespacemedia we make use of a variety of third-party services to build, deploy and monitor our websites. We choose our services based on a variety of factors including feature set, per-user pricing and integration with our other services.
We’ve been using Sentry for a while now to monitor any errors which happen during the day-to-day usage of our websites and it’s incredibly useful for us as a team. We get fast notifications of problems and we get good insights into which lines of code caused the errors and the environmental conditions surrounding them. Recently we discovered the well-advertised Opbeat, which seems to be a Django focussed Sentry-like tool which seemed to build upon the feature set we were used to in Sentry. One of the main selling points for us was their pricing structure, free for the first 3 users and $14 per month for each user thereafter. We were paying monthly for Sentry at a fixed price of $29 per month, so while Opbeat wouldn’t be cheaper in the medium- to long-term future, it would be better value for our current setup of 4 developers and potentially provide a better resource for us going forward.
With this pricing structure and feature-set of Opbeat in mind, we moved forwards with integrating it into our development process, including our fully automated project configuration tools all the way through to our deployment and update processes. Everything was working well and we had a number of projects running Sentry and Opbeat side-by-side so we would make real world comparisons - and Opbeat was doing well in them.
Then this afternoon I received an email from Opbeat announcing two things; their new “New Relic”-like metrics service which would provide insights into slow code or long-running queries, but also that their pricing model was changing. Instead of keeping with the pricing model of 3 free plus $14 per month per user, they decided to split their offering into two tiers: free and pro. The free tier had all of the features but only retained data for 24 hours whereas the Pro tier would retain error data forever, and metric data for 72 hours.
With this new pricing, it no longer made sense for us to even contemplate using Opbeat. With one change they’ve gone from costing $14 per month under our current structure to costing $56 per month - no longer better value compared to Sentry. Now it seems that we will revert our tooling back to Sentry - which will allow us to scale to 10 users and still pay only $29pm (vs $140pm with Opbeat) - and then use Opbeat’s free plan for the code metrics, but not use the error tracking feature at all. To me this seems crazy, we were fully onboard with Opbeat and were looking forward to migrating all of our sites over and use it full-time, but now it doesn’t make any sense to.
I replied to the newsletter I received from Opbeat and let them know my thoughts, but I felt it was also worth going into a little more detail in this blog post. I’m really hoping that Opbeat look into making the pricing a little more appealing to companies - both small and large. I really don’t feel the pricing scales well at all, and I don’t like the way that they went from the value proposition of “the first 3 seats are always free” to this new model where the first 3 seats are absolutely not free. The cynical side of me feels like they went all-out on their social media and general marketing to hook a lot of developers and small teams in on the “first 3 seats” message and then they switched it up as soon as they reached a good enough number. Whether that’s true or not, I have no idea.