Do you know the Datadog software?
Datadog is also the name of the company that produces it, and here we bring it in a comparison with Pandora FMS. From the outset we tell you that Datadog is actually Software as a Service (SaaS), while Pandora FMS is both an autonomous program and a service. Do you want to know more about the alternative to Datadog? Come and read below!
The alternative to Datadog
The alternative to Datadog is, as expected, Pandora FMS, a software that has been accumulating experience since 2004. Both software combine both free and proprietary solutions, although it is true that Datadog decided to settle solely in the cloud to centralize its operations and adapt to the American way of working. For this reason, they put support in the background and predict a learning curve of between one to two weeks.
As each monitoring software names its components differently, we will use the Pandora FMS glossary, which is more extensive due to the greater time it has been on the market; the Datadog glossary is at this link.
Free software and proprietary software
While Pandora FMS is open source and is required as a base for the Enterprise version, Datadog is proprietary at its core, running on Amazon Web Services® (AWS) servers. Pandora FMS has been on AWS since March 2016 with an Amazon Machine Image (AMI), so we can install our server in the cloud without any problem. Both develop for GNU/Linux, Windows (Datadog for Windows 7 onwards) and Mac OS X. Datadog is based on agile development or agile system and Pandora FMS on Continuous Software Integration and Continuous Software Release since version 7.0 NG.
The alternative to Datadog offers a high availability scheme; In the following graph, in broad strokes, we draw the high availability architecture of Pandora FMS:
Datadog high level architecture
Let us now explain component by component.
Agent Software (AS)
A sticky issue for Datadog was having an open source AS v5 based only on the Python language; later they went to v6, also open source, based on the Go language and some Python. Curiously, they have embedded a web server that only accepts local connections on port 5001 (on 32-bit Windows this feature is not included) and we watched tutorials to revert or go from v6 to v5.
On the other hand, the alternative to Datadog, Pandora FMS, offers perl versions for monitoring Unix / Linux devices without the need to install additional packages. In Windows environments, the Software Agent is written in C and compiled, so additional packages are not necessary. In addition, Pandora FMS has an agent auto-update system from the console.
The alternative to Datadog proposes Ansible and Puppet for AS installation, as well as the possibility of deploying agents using the Active Directory functionality of the Enterprise version, being much more comfortable for companies that use these services. This will serve to deploy many other applications. Datadog is designed to go from computer to device, one at a time (each software agent with its web console), although in the case of Docker (since 2015) the scenario is highly automated. On October 18, 2018, they presented the Datadog Cluster Agent that allows from 20 thousand pods, which entails a decrease in the workload of the API servers. All this compared to Pandora FMS falls into the distributed monitoring section with several servers (in this link we also present all the different possible combinations).
Both applications can connect their ASs through APIs but the ASs of the alternative to Datadog, Pandora FMS, connect using Tentacle as the primary option, accepting deliveries by FTP or SSH; in this case, the ASs do not use the API, but it is accessible to be consulted by whoever the administrator deems appropriate (plugins, scripts, integrations, etc.).
In the event of a communication failure, Datadog would lose the connection and, with it, the data. On the other hand, Pandora FMS, when sending the information in XML files, would not lose the data since it saves it until it can be sent. As soon as the connection is restored, they are sent, respecting the timestamps.
Like all SaaS, Datadog has its maximum limits on the number of API connections and additional charges in certain amounts.
Datadog highlights its support for third parties, and that it is capable of receiving data directly from a statsd agent that it sends via UDP (obviously, without data encryption and without confirmation or verification of delivery) without representing a greater load on the monitored device. However, Datadog incorporates it -this is the great thing about free software- into its own AS -it calls them DogStatsD– to allow tagging.
Datadog uses four reserved tags (system tags) called host, device, service, and source. We think that Datadog starts from a normalized or standard scenario (in two type of graphic screens called TimeBoards and ScreenBoards and that can be shared by public URLs and in JSON format) and that from there each user begins their customizations up to the specified limit.
There is a free version for up to five devices and data recording with a maximum of 24 hours without any alert; Pandora FMS in its OpenSource version is completely free, with no limit on the number of devices and no ties (except if we stay on AWS, as we explained previously).
Datadog Alternative Plugins
We counted 260 (Datadog calls them Integrations), both internal and external. We can mention SNMP (internal) that allows us to create our own MIBs with the help of Python (pysnmp) and to control our source code stored on GitHub with a web hook (in the same style as Jenkins). The alternative to Datadog, Pandora FMS, has 530 plugins that are totally free and 158 in the Enterprise version for a total of 688 at the end of November 2018; not counting those that each user has specifically designed for their environment. The simplicity of plugin development in Pandora FMS is one of its strong points.
Watchdog looks at patterns and trends in application metrics such as request rate, error rate and latency, and unexpected behavior. Watchdog evaluates all services and resources without the need to configure a monitor for each service. Of course, you need a certain amount of time to collect information. It is the equivalent to the Prediction Server in Pandora FMS available since 2008.
In both Pandora FMS and Datadog we can define them via the console, but in Datadog there are third-party solutions such as Barkdog (Ruby Gem), Dogpush (YAML) or datadog_monitor.
Do you want to know more about the alternative to Datadog? Do you need to monitor a large number of devices? Enter here to learn more about Pandora FMS Enterprise: https://pandorafms.com/es
Request a free demo now for more than 100 devices and start experiencing full monitoring flexibility: https://pandorafms.com/en/free-demo