March 09, 2019
Score 8 out of 10
Overall Satisfaction with Datadog
We use Datadog on almost all of our production systems to give us granularity into how systems are running, and additionally, use Datadog's logging feature to aggregate all of our logs across systems in real time. Datadog's simple dashboards enable visibility from different providers to be displayed all in one place, even if systems are not from the same vendor.
- Does a good job at log management with full search, live-access, and automatic archiving to S3 is also simple.
- APM is fantastic and gives great insights into production machines, but is not cheap.
- Because APM is billed by instance, it can be very expensive -- perhaps even more than the cost of the underlying instances depending on the kinds of systems you are running.
- While it's not difficult to deploy, it certainly has quirks owing to the limits of cloud platforms -- we wish it was easier to set up for some services.
- Visibility into systems and performance gives us a go-to overview of all of our services.
- APM provides additional context for which parts of our workloads are taking the most time from our servers, and gives us the ability to try to reduce these pain points.
- Log management makes viewing logs from remote systems much easier, ultimately saving time and a lot of developer frustration.
Ultimately, Datadog had the most already-built bridges into our existing infrastructure -- third parties that we're using for certain services are far more likely to work with Datadog than other systems. This means that, while expensive, Datadog has done a tremendous amount of work for us and leaves us with a greater ability to see what our systems are doing and why.
For organizations that understand and require the value-add of Datadog, it's a great choice for log management, APM, and system visibility. Because of its costs, it's not well suited to smaller organizations, or organizations running lots of small workloads on inexpensive VMs, where you are stuck paying the same price that an organization would pay for a server 100x the size.