Datadog vs. InfluxDB
By
Al Sargent /
Product, Use Cases, Developer
May 05, 2020
Navigate to:
If you’re responsible for monitoring, chances are you’ve heard of Datadog. Like InfluxDB, Datadog is a monitoring platform for cloud applications, bringing together data from containers, servers, databases and third-party services.
InfluxData and Datadog approach monitoring from different starting points. InfluxDB is an open-source time series data platform that can be used for a range of use cases, one of which is monitoring. Datadog started out as an application purpose-built for infrastructure monitoring, and from there created purpose-built offerings for monitoring application performance, logs, security, and more.
Both companies focus on making things easier for engineering teams. For instance, today on review site G2, “easy to use” is in the headline of both companies’ lead reviews.
Given this ease of use, a number of our customers use both InfluxDB for infrastructure monitoring, and Datadog for other types of monitoring.
But that led us to ask: why are these customers using InfluxDB in addition to Datadog? After all, Datadog aims to be a place where you can “see it all in one place.”
So, we decided to investigate. Here’s what we found.
How are InfluxDB and Datadog used?
InfluxDB and Datadog are two very similar solutions in that they bring together data from containers, servers, databases and third-party services. It’s key to understand, however, that they approach the idea of monitoring from two totally different starting points.
InfluxDB is an open-source time series data platform that can itself be used for a wide range of different purposes, only one of which is monitoring. Datadog, on the other hand, began life as an application that was specifically built for infrastructure monitoring. Only later were offerings added regarding the monitoring of application performance, logs, security and more.
If they share anything in common, it’s that both offerings are focused on making things as easy as possible for engineering and DevOps teams everywhere. While there is some degree of overlap, both bring their own innovations to the table to the point where many people use both solutions rather than picking one or the other. A lot of professionals use InfluxDB for infrastructure monitoring, for example, while using Datadog for other needs that they have.
With regards to those customers that use both, many prefer InfluxDB for infrastructure monitoring and for storing performance metrics for assets like servers, containers, virtual machines and databases, and for similar tasks of that nature. They tend to use Datadog for application performance management metrics, otherwise known as APMs.
Yes, it’s absolutely true that seeing all of your logs, metrics and traces in one place would make it easier to find and fix problems as quickly as possible posing the question of why someone would want to use two similar solutions to begin with. This is mostly because people tend to run into problems with Datadog in two key areas: Datadog pricing and the flexibility of some Datadog features. First, let’s provide a quick overview of what Datadog is and how it works.
What is Datadog?
At its core, Datadog is a monitoring platform designed for cloud applications that brings together critical information from servers, containers, databases and even third-party services to make a stack totally observable from the top down. These unique features and capabilities are intended to help DevOps teams all over avoid downtime while also solving any performance issues that they’re dealing with, all in service of the most important goal of all: guaranteeing that customers are getting the best experience out of those cloud applications, no matter what.
Datadog also offers a variety of turnkey integrations designed to make the lives of DevOps teams easier by seamlessly aggregating metrics and events across the entirety of their DevOps stack. This includes working with not only data from SaaS and cloud providers, but also automation tools, other monitoring and instrumentation platforms, source control and bug tracking features, databases and common server components and more.
What does Datadog do?
As stated, Datadog is a viable way to gain full visibility into modern cloud-based applications. This allows teams working on them to monitor, troubleshoot and optimize application performance relative to their goals at the moment.
Datadog is used for, among other things:
- Tracing requests from end to end across all distributed systems that you might be working with.
- Tracking application performance with automatically generated service overviews, thus guaranteeing that actionable insight and intelligence is always available for those who need it.
- Graphing and alerting on error rates or latency percentiles that could negatively impact the end user experience.
- Instrumenting code using open source tracing libraries to make the entire process easier across the board.
How much does DataDog cost?
While Datadog’s per-server, per-month pricing for infrastructure monitoring is easy to understand, customers tell us that it becomes very expensive at scale.
For example, suppose you’re an enterprise with 2,000 hosts to monitor. That’s actually a small number by today’s standards, given microservice architectures where many specialized backend components work together to power an application. (Note that, for billing purposes, Datadog counts virtual machines running on AWS, Google Cloud, Azure, or VMware vSphere as individual hosts.)
At $23 per monitored host per month, Datadog Enterprise will have a list price of over $500,000. Even with a discount, that’s not a number to sneeze at.
On top of that, Datadog charges extra for:
- Serverless tasks $1 each
- Serverless function $5 each per month
- Custom metrics (list pricing not available on website)
- Individual containers beyond specified limits
- Usage spikes (Datadog excludes the top 1% of usage much different from billing on average use.)
<figcaption> A fast-growing volume of serverless functions and containers can add up to a significant chunk of change when using Datadog. Photo source: @sharonmccutcheon via unsplash</figcaption>
How much does InfluxDB cost?
InfluxData doesn’t price by monitored host or metric, but by the size of the server or cluster that stores your monitoring data. This is a critical difference that means InfluxDB is only 30% the cost of Datadog for most real-world infrastructure monitoring use cases.
For our pricing analysis, we looked at pricing for the fully managed InfluxDB Cloud product, since this is closest to Datadog’s offering. Like Datadog, InfluxDB Cloud is fully managed 24/7 by our in-house site reliability engineering (SRE) team. This is in contrast to InfluxDB Enterprise, which customers manage on their own infrastructure.
After looking at prices paid by InfluxDB Cloud customers, we found that the median price is $6.75 per monitored host per month.
Slash your infrastructure monitoring costs by 70%
When you compare InfluxDB’s $6.75 price point to Datadog’s $23, you see savings of 70%.
Your mileage might vary, but put us to the test! Talk to one of our specialists to see how much we can reduce your infrastructure monitoring bill.
To be clear, we’re talking about pricing for infrastructure monitoring. Datadog charges extra for its APM, log, security, and other monitoring products, which wasn’t included in the above analysis.
Why does InfluxDB cost less than Datadog?
You might ask, why can’t Datadog simply match InfluxData’s price?
We don’t believe it’s that simple.
One reason InfluxDB is typically less expensive is our lower cost structure. InfluxDB Cloud is based on an open-source product InfluxDB, along with some closed source functionality like clustering and role-based access control (RBAC). A global community of engineers freely contributes code to InfluxDB, Telegraf, and our Flux language, as shown below. This lets us deliver features at a lower cost and we pass those savings on to our customers.
<figcaption> Some of the contributors to the InfluxDB, Telegraf, and Flux open source repositories (April 2020)</figcaption>
In contrast, Datadog’s product is largely proprietary software it builds itself, leading to a higher cost structure that it passes on to its customers. You can see for yourself on Datadog’s GitHub repository; while some of its agents and libraries are open source, its core product is not.
Now you might be thinking: InfluxDB is lower-cost but is it as good? Read on
Datadog vs. InfluxDB flexibility
The second difference our joint customers brought up was flexibility. Specifically, they mentioned the ability for developers to tailor infrastructure monitoring to their exact needs, and make it manageable using GitOps (storing operations configurations in repositories like GitHub). Let’s dive in.
Broader observability
On its website, Datadog states that they monitor 400 different technologies. While that’s not bad, it’s less than half of the 735 FluentD plugins plus 181 Telegraf input plugins that can send data to InfluxDB.
Maybe Datadog’s 400 plugins will cover your needs today, and maybe they won’t. But if you want to ensure that your future needs are covered, it makes sense to go with a product that taps into both the Telegraf and FluentD communities. By doing so, you can ensure the broadest possible observability to detect problems sooner and fix them faster now and going forward.
<figcaption> Broader observability with InfluxDB. Photo source: @awerin via unsplash</figcaption>
One customer also told us that Datadog agents can get overwhelmed, crash, and send duplicates of data none of which happens with Telegraf. Your mileage may vary, but at the very least, agent stress testing is something you might want to incorporate into your Datadog and Telegraf evaluations.
Unleash your monitoring data
Another challenge our customers told us about is getting data out of Datadog. Datadog’s documentation states that, by default, it rate-limits its output API to 100 requests per hour.
<figcaption> With InfluxDB, you can unleash your data for effective use. Photo source: @theodorrr via unsplash</figcaption>
InfluxData’s stance is the complete opposite; we don’t artificially rate-limit our output. We realize that no monitoring platform is an island. It needs to integrate with a broad set of vendors: visualization, alerting, ML/AI, and more. So we can’t be stingy around data export.
Because of this, InfluxData provides first-class outbound integrations across your DevOps and monitoring toolchain:
- Grafana for visualization
- PagerDuty, Slack, and many other alerting tools
- AWS Kinesis, Google Cloud Pub/Sub, Azure Monitor to send to those clouds' AI/ML engines
- Jupyter and Zeppelin notebooks for machine learning and AI-powered analysis
- 30+ Telegraf output plugins
- 8 client libraries in popular languages
Our philosophy is: this is your data. You paid to collect it. You should be free to use it to drive further efficiencies in your monitoring and better experiences for your customers.
Efficiency through analytics
Datadog query filters and query language let you do min, max, and average, and counts. But that’s a small fraction of the dozens of functions in InfluxDB’s Flux data scripting language. These InfluxDB functions streamline common tasks for teams running production systems, such as:
- Calculate percentiles to track SLA compliance, which are often measured at the 90th, 95th, or 98th percentile.
- Window and aggregate data to pick out trends from noisy data sets.
- Enrich monitoring data with data in SQL databases, like account data, to facilitate customer outreach during outages.
- Forecast with Holt-Winters to predict outages and capacity issues.
- Geographically track your monitoring metrics to better determine which regions are experiencing problems.
<figcaption> Analyze metrics by geography to determine where outages occur. Photo source: @martinsanchez via unsplash</figcaption>
We built all of this flexibility into InfluxDB to make it a complete time series data platform. This lets you give your developers the monitoring flexibility they crave, so they can quickly find and fix problems fast to keep critical systems running 24/7.
Deployment flexibility
Datadog is deployed only on AWS. In contrast, you can deploy InfluxDB on dozens of AWS, Microsoft Azure, and Google Cloud regions, as well as on your own servers. This is especially significant given customer privacy regulations such as GDPR and CCPA that might have business implications for where you host your data.
Datadog vs. InfluxDB features
Take a look at this detailed comparison of Datadog vs. InfluxDB. You’ll see that InfluxDB matches Datadog in many areas and surpasses it in others.
Datadog | InfluxDB | |
---|---|---|
# of technologies monitored | 400, per its website | 900 700 FluentD plugins, 200 Telegraf plugins |
Cost/host/month | $15 to $23 | $7 on average |
Extra $ for custom metrics | Yes | No |
Extra $ for containers | Yes | No |
Extra $ for functions | Yes | No |
Max metrics output per hour | 100 | Unlimited |
Out of the box dashboards | Yes | Yes |
Role-based access control (RBAC) | Yes | Yes |
Auth0 single sign-on | Yes | Yes |
Azure AD single sign-on | Yes | Yes |
GitHub single sign-on | No | Yes |
GitLab single sign-on | No | Yes |
Google single sign-on | Yes | Yes |
Heroku single sign-on | No | Yes |
Okta single sign-on | Yes | Yes |
Oauth single sign-on | No | Yes |
LDAP single sign-on | No | Yes |
# of user accounts | Unlimited | Unlimited |
Full resolution data retention | 15 months | Unlimited |
# of alerts | Unlimited | Unlimited |
Alerting channels |
5 - OpsGenie PagerDuty Webhooks Slack Dingtalk |
20 - complete list here |
Max containers per host | 20 | Unlimited |
Max custom metrics per host | 200 | Unlimited |
Outlier Detection | Yes | Yes |
Holt-Winters Forecasting | No | Yes |
TensorFlow support | No | Yes |
Jupyter Notebooks support | Yes | Yes |
AWS AI Services support | No | Yes |
Google Cloud AI support | No | Yes |
Microsoft Azure AI support | No | Yes |
# of cloud providers supported | 1 - AWS | 3 - AWS, Google, Azure |
On-prem deployment | No | Yes |
How to migrate off Datadog?
If you’d like to migrate off Datadog, we’ve made it easy and low-risk by allowing you to migrate gradually, and on your own timeline. Here’s an overview of what that looks like.
<figcaption> Migrating from Datadog to InfluxDB</figcaption>
Let’s walk through this diagram.
Telegraf, our data integration agent, can ingest metrics from StatsD format used by the Datadog DogStatsD plugin. Just set the datadog_extensions
flag to true in your telegraf.conf
file, as shown below, and Telegraf will be able to ingest Datadog metrics:
## Parses extensions to statsd in the datadog statsd format
## currently supports metrics and datadog tags.
## http://docs.datadoghq.com/guides/dogstatsd/
datadog_extensions = true
And, Telegraf lets you dual-write monitoring metrics to both Datadog and InfluxDB. For Datadog, you simply put the following into your telegraf.conf
file:
[[outputs.datadog]]
apikey = "<datadog api key>"
This lets you verify that your monitoring metrics are coming into InfluxDB; explore our query capabilities and outbound integrations; and eventually, migrate with no monitoring outages.
Dual-write also means that, if you can’t get monitoring data out of Datadog due to their 100 output API requests/hour limit, you can use both InfluxDB and Datadog for a period of time until you no longer need your Datadog monitoring data. At that point, you can fully switch over to InfluxDB.
Summary
While any Datadog review shows that it does bring a lot of benefits to DevOps teams and engineers everywhere, it’s clear that there are also certain limitations baked into the DNA of the solution that are difficult for a lot of people to overcome.
In addition to the fact that Datadog charges extra for a lot of the functionality that has become critical to the modern application development environment, it’s just not as flexible as various use cases require. Datadog can ONLY be deployed on Amazon Web Services, whereas InfluxDB gives you the choice of AWS, Microsoft Azure, Google Cloud and more. You can even deploy on your own servers if you need to, which is particularly important for customers dealing with compliance and privacy regulations like the GDPR and the CCPA.
So, if you’re looking for a Datadog alternative that lets you:
- Slash your infrastructure monitoring budget by up to 70%
- Increase observability for your developers, engineers, and SREs
- Unlock monitoring data so it can be used across your DevOps toolchain
- Increase DevOps and SRE team efficiency using a broad range of analytics
- Store data wherever business needs dictate
then speak with one of our monitoring specialists for a free consultation to learn how we can help you transition off of Datadog.