Item: Datadog
Rating: 3
Author: Verified User

Overall Satisfaction with Datadog

Use Cases and Deployment Scope

We assessed Datadog as a provider for telemetry and user monitoring to aid in our optimization and troubleshooting efforts of our web application. We were looking to measure overall performance of HTTP endpoints and background jobs, surface errors and inefficient areas, and drill down into problem areas and dependencies (DB, AWS services, external web services, inter-service communication, etc)

Pros and Cons

Pros

Datadog agent was really good at analyzing performance
UI allowed you to cross reference traces, logs, errors, metrics, etc
Customizable

Cons

In my experience, .NET Tracing Agent caused severe and untraceable performance issues
In my opinion, usage and billing structures were opaque and surprising
In my experience, documentation was incomplete, contradicting or sometimes completely wrong, even for common infrastructure (AWS Fargate)
I feel support was unhelpful at times, and bounced us back and forth to other teams
In my opinion, multiple methods of sample rate control were ineffective, adding to excessive usage and cost

Return on Investment

Made us consider our tracing and logging practices in more depth

Usability

The Datadog features and platform/UI is rich and powerful. This would be a 9/10 if all the features required worked -- but the places where it didn't work really fell apart.
Primarily:
- In my experience, usage reports/dashboard were inaccurate, contradicting, and opaque
- In my opinion, rate sampling controls were also ineffective and inconsistent, via multiple avenues (remote config, local config, code config). Trace sampling reports were inaccurate and inconsistent as well
- In my opinion, documentation was challenging to dig through for various platforms, and support was not as helpful as we'd hoped

Alternatives Considered

New Relic, Amazon CloudWatch, OpenTelemetry, Jaeger and Zipkin

Datadog seems to be the most feature-rich of all the alternatives we've considered, however due to problems outlined earlier, some of the others have benefits. OpenTel can give us a way to make our platforms compatible with a variety of vendors, and can be done without injecting proprietary tracers into our infrastructure, giving us more confidence that our performance will remain stable. AWS Cloudwatch (with appsignals/xray), New Relic are vendors that offer some of the same features we were looking for with Datadog, but have other benefits. AWS is integrated into our infrastructure, and New Relic has a very approachable and transparent pricing model. Open Source tools like Jaeger and Zipkin are also on our radar as ways to achieve our telemetry goals simply and cost effectively.

Key Insights

Do you think Datadog delivers good value for the price?

Are you happy with Datadog's feature set?

Yes

Did Datadog live up to sales and marketing promises?

Did implementation of Datadog go as expected?

Would you buy Datadog again?

Other Software Used

New Relic

Likelihood to Recommend

Datadog may be better suited for teams that have a more out-of-the-box infrastructure, on the primary platforms Datadog supports. You may also have better results if you have a bigger team dedicated to devops and/or a bigger budget. We found that trying to adapt it to our use case (small team, .NET on AWS Fargate) wasn't feasible. We continually ran into roadblocks that required us to dig through documentation (and at times, having to figure out some documentation was wrong), go back and forth with support, and in my opinion, waste money on excessive and unintended usages due to opaque pricing models and inaccurate usage reports, as well as broken/non-functional rate sampling controls.

Optimization
Observability
User Monitoring

Likelihood to Renew

Definitely will not revisit after our issues and, in my opinion, poor support.

Evaluating Datadog and Competitors

Products Replaced

Key Differentiators

Cloud Solutions
Scalability
Integration with Other Systems
Ease of Use

Evaluation Lessons Learned

Would time evaluation when we could do a more thorough testing and rollout.

Datadog Implementation

Implementation Rating

Documentation was difficult to work through, rollout was catastrophic (completely outage)

Implementation Details / Implementation Partner

Implemented in-house
Don't know

Tracer caused severe performance issues causing system outage

Datadog Support

Support Rating

In my opinion, support was often not helpful, repeating wrong answers, or transferring us to another person multiple times.

Datadog Customer Support Pros and Cons

Pros	Cons
Kept well informed	Slow Resolution Poor followup Less knowledgeable Problems left unsolved Escalation required Difficult to get immediate help Need to explain problems multiple times Support doesn't seem to care