Airbrake vs. Scout APM

Airbrake is very good at what it does, I don’t really have any criticism at all on that front. It’s less well-suited when bugfixing goes beyond the immediate error and means looking at a lot of context (particularly asynchronous context) like logs.

Incentivized

To me, it's the best APM out there for a Rails application. It is really easy to set up, configure and start debugging any kind of bad-performing endpoint. We have solved issues going from High Response Time to Memory Leaks/Bloats. All of this can be achieved even if you are hosting your app in a PaaS like Heroku. (Where sometimes is difficult to install an agent like in other vendors).

Incentivized

• Integration with most of languages and frameworks as well support via existing logging configs.
• Aggregations on occurrences which helps to understand and analyse issues better.
• Web UI displays all error related parameters which I never could have figured out in manual debugging.

• Rails instrumentation
• Segmentation by controllers and sometimes methods
• Detecting low performing endpoints

Incentivized

• We use Airbrake in conjunction with OpsGenie, but I feel like there could be more room for integration between the two.
• I think it would also be nice if there was a GitHub integration that would comment on recently merged error-prone PRs, currently, we need to dig into the error to find the commit.
• Generally, more integrations would be nice as people often forget about Airbrake when they are stressed out about an issue.

Incentivized

• Errors (is a new functionality but I didn't get well how it can replace rollbar for example).
• Usage measuring: In some cases is easy to get "disabled" due to a peak of bad requests.
• Editing app's names

Incentivized

I love to use Airbrake and New Relic in conjunction. New Relic has better metrics and data that you can really dig into (especially for optimizations), but the error part has always been kinda meh. I fee like Airbrake has done an awesome job at this

We have preferred Scout APM over Datadog due to: - Easy to configure and install on Heroku Platform (Datadog's agent can give some trouble when installing it on Heroku's Dynos) - Easy to monitor normal behavior and be actionable on endpoints who are out of normal - It has an easier pricing which facilitates to make predictions on budget

Incentivized

• Hugely reduced frontend errors in devices we don’t often test: means more conversions and sales
• Lets us know immediately when there is an error: shorter error durations also means higher sales
• Ultimately once we’ve solved all the most common bugs: more time for developers to focus on other things

Incentivized

• Speed to solve incidents
• Easy to monitor new endpoints or deployments
• Easy to monitor background jobs (like sidekiq)

Incentivized