Atatus

Rationale

Atatus is our Application Performance Management (APM) tool where the engineering team gets insights into our products' behavior and information on how our users are experiencing them. It is a powerful and valuable tool that lights the way for technical decisions and analysis.

The main reasons why we chose it over other alternatives are:

1. It provides a lot of data and information in one place, making it possible to analyze some behaviors without changing to other apps or tools.
2. It integrates Real User Monitoring (RUM) for our web applications.
3. It provides descriptive information on the errors within the products.
4. It works with an agent that is inside the code and constantly collects precise performance information.
5. It gathers non-sensible data about user interactions with our products, giving insights to improve our code or catch bottlenecks.
6. It provides explicit and precise information about Web Vitals for our web applications.
7. It breaks down the performance information for our products, allowing us to analyze it for each part of the products and not only in a general way.
8. It provides dashboards with explicit graphics for numerous types of metrics.
9. It allows us to compare the performance between an extensive list of previous deployments.

Alternatives

The following alternatives were considered but not chosen for the following reasons:

1. New Relic: At the moment it was tried, there was a lack of support for asynchronous Python and other tools.
2. Sentry: It does not provide a considerable improvement in the observability of our products.
3. Datadog: It was the more robust and complete alternative, but its price model related to our autoscaling strategy was incompatible.
4. Elastic APM: It did not bring added value to improve the observability aspects of our components.
5. OpenTelemetry with AWS X-Ray, AWS Managed Prometheus and AWS Managed Grafana: This alternative provides powerful tools but requires much engineering effort.
6. Cloudflare Core Web Vitals: It does not provide better metrics than its alternatives, bringing insufficient graphics and less specific data.

Note: Currently, Coralogix is being tested. A review is pending.

Usage

We use Atatusfor:

1. Products' performance monitoring.
2. Real User Monitoring.
3. Synthetic monitoring.
4. Error tracking.
5. Database monitoring.
6. Transaction monitoring.
7. External request monitoring.
8. Session traces.
9. Ajax monitoring.
10. Alerting on events.
11. Dashboard analysis.

We do not use Atatusfor:

1. Infrastructure monitoring.
2. Logs monitoring.
3. Kubernetes monitoring.

Guidelines

You can access the Atatus hub by entering the Atatus application through Okta. There you can manage the different projects through the Projects tab.