Datadog Enterprise

My main use case for Datadog  is monitoring our servers.

A specific example of how I'm using Datadog  to monitor my server is that we are maintaining request and latency and looking for errors.

I really enjoy the user interface of Datadog, and it makes it easy to find what I need. In my opinion, the best features Datadog offers are the customizable dashboards and the Watchdog.

The customizable dashboards and Watchdog help me in my daily work because they're easy to find and easy to look at to get the information I need. Datadog has positively impacted my organization by making finding and resolving issues a lot easier and efficient.

I think Datadog can be improved by continually finding errors and making things easy to see and customize.

I have been using Datadog for one month.

Datadog is stable.

Datadog's scalability has been easy to put on each server that we want to monitor.

I have not had to contact customer support yet, but I've heard they are great.

Neutral

We previously used our own custom solution, but Datadog is a lot easier.

I'm not sure if I've seen a return on investment.

My experience with pricing, setup cost, and licensing is that it was easy to find and easy to purchase and easy to estimate.

I did not make the decision to evaluate other options before choosing Datadog.

I would rate Datadog a nine out of ten.

I give it this rating because I think just catching some of the data delays and latency live could be a little bit better, but overall, I think it's been great.

I would recommend Datadog and say that it's easy to customize and find what you're looking for.

My main use case for Datadog  involves working on projects related to our sales reps in terms of registering new clients, and I've been using Datadog  to pull up instances of them while they're beta testing our product that we're rolling out just to see where their errors are occurring and what their behavior was leading up to that.

I can't think of all of the specific details, but there was a sales rep who was running into a particular error message through their sales registration process, and they weren't giving us a lot of specific screenshots or other error information to help us troubleshoot. I went into Datadog and looked at the timestamp and was able to look at the actual steps they took in our platform during their registration and was able to determine what the cause of that error was. I believe if I remember correctly, it was user error; they were clicking something incorrectly.

One thing I've seen in my main use case for Datadog is an option that our team can add on, and it's the ability to track behavior based on the user ID. I'm not sure at this time if our team has turned that on, but I do think that's a really valuable feature to have, especially with the real-time user management where you can watch the replay. Because we have so many users that are using our platform, the ability to filter those replay videos based on the user ID would be so much more helpful. Especially in terms where we're testing a specific product that we're rolling out, we start with smaller beta tests, so being able to filter those users by the user IDs of those using the beta test would be much more helpful than just looking at every interaction in Datadog as a whole.

The best features Datadog offers are the replay videos, which I really find super helpful as someone who works in QA. So much of testing is looking at the UI, and being able to look back at the actual visual steps that a user is taking is really valuable.

Datadog has impacted our organization positively in a major way because not even just as a QA engineer having access to the real-time replay, but just as a team, all of us being able to access this data and see what parts of our system are causing the most errors or resulting in the most frustration with users. I can't speak for everybody else because I don't know how each other segment of the business is using it, but I can imagine just in terms of how it's been beneficial to me; I can imagine that it's being beneficial to everybody else and they're able to see those areas of the system that are causing more frustration versus less.

I think Datadog can be improved, but it's a question that I'm not totally sure what the answer is. Being that my use case for it is pretty specific, I'm not sure that I have used or even really explored all of the different features that Datadog offers. So I'm not sure that I know where there are gaps in terms of features that should be there or aren't there.

I will go back to just the ability to filter based on user ID as an option that has to be set up by an organization, but I would maybe recommend that being something part of an organization's onboarding to present that as a first step. I think as an organization gets bigger or even if the organization starts using Datadog and is large, it's going to be potentially more difficult to troubleshoot specific scenarios if you're sorting through such a large amount of data.

I have been working in this role for a little over a year now.

As far as I can tell, Datadog has been stable.

I believe we have about 500 or so employees in our organization using our platform, and Datadog seems to be able to handle that load sufficiently, as far as I can tell. So I think scalability is good.

I haven't had an instance where I've reached out to customer support for Datadog, so I do not know.

I do not believe we used a different solution previously for this.

I cannot answer if I have seen a return on investment; I'm not part of the leadership in terms of making that decision. Regarding time saved, in my specific use case as a QA engineer, I would say that Datadog probably didn't save me a ton of time because there are so many replay videos that I had to sort through in order to find the particular sales reps that I'm looking for for our beta test group. That's why I think the ability to filter videos by the user ID would be so much more helpful. I believe features that would provide a lot of time savings, just enabling you to really narrow down and filter the type of frustration or user interaction that you're looking for. But in regards to your specific question, I don't think that's an answer that I'm totally qualified to answer.

I was not part of the decision-making process before choosing Datadog, so I cannot speak to whether we evaluated other options.

Right now our users are in the middle of the beta test. At the beginning of rolling the test out, I probably used the replay videos more just as the users were getting more familiar with the tool. They were probably running into more errors than they would be at this point now that they're more used to the tool. So it kind of ebbs and flows; at the beginning of a test, I'm probably using it pretty frequently and then as it goes on, probably less often.

It does help resolve issues faster, especially because our sales reps are used to working really quickly in terms of the sales registration, as they're racing through it. They're more likely to accidentally click something or click something incorrectly and not fully pay attention to what they're doing because they're just used to their flow. Being able to go back and watch the replay and see that a person clicked this button when they intended to click another button, or identifying the action that caused an error versus going off of their memory.

I have not noticed any measurable outcomes in terms of reduction in support tickets or faster resolution times since I started using Datadog. For myself, looking at the users in our beta test group, none of those came as a result of any sort of support ticket. It came from messages in Microsoft Teams  with all the people in the beta group. We have resulted in fewer messages in relation to the beta test because they are more familiar with the tool. Now that they know there might be differences in terms of what their usual flow is versus how their flow is during the beta test group, they are resulting in fewer messages because they are probably being more careful or they've figured out those inflection points that would result in an error.

My biggest piece of advice for others looking into using Datadog would be to use the filters based on user ID; it will save so much time in terms of troubleshooting specific error interactions or occurrences. I would also suggest having a UI that's more simple for people that are less technical. For example, logging into Datadog, the dashboard is pretty overwhelming in terms of all of the bar charts and options; I think having a more simplified toggle for people that are not looking for all of the options in terms of data, and then having a more technical toggle for people that are looking for more granular data, would be helpful.

I rate Datadog 10 out of 10.

We use Datadog  for all of our observability needs and application performance monitoring. We recently transitioned our logs to Datadog . We also use it for incident management and on-call paging. We use Datadog for almost everything monitoring and observability related.

We use Datadog for figuring out the root cause of incidents. One of the more recent use cases was when we encountered a failure where one of our main microservices kept dying and couldn't give a response. Every request to it was getting a 500. We dug into some of the traces and logs, used the Kubernetes  Explorer in Datadog, and found out that the application couldn't reach some metric due to its scaling. We were able to figure out the root cause because of the Kubernetes  Event Explorer in Datadog. We pushed out a hotfix which restored the application to working condition.

Our incident response team leverages Datadog to page relevant on-calls for whatever service is down that's owned by that team, so they can get the appropriate SMEs and bring the service back up. That's the most common use case for our incident response. All of our teams appreciate using Datadog on-call for incident response because there are numerous notification settings to configure. The on-call schedules are very flexible with overrides and different paging rules, depending on urgency of the matter at stake.

As an administrator of Datadog, I really appreciate Fleet  Automation. I also value the overall APM  page for each service, including the default dashboards on the service page because they provide exactly what you need to see in terms of request errors and duration latency. These two are probably my favorite features because the service page gives a perfect look at everything you'd want to see for a service immediately, and then you can scroll down and see more infrastructure specific metrics. If it's a Java app, you can see JVM metrics. Fleet  Automation really helps me as an administrator because I can see exactly what's going on with each of my agents.

My SRE team is responsible for upgrading and maintaining the agents, and with Fleet Automation, we've been able to leverage remote agent upgrades, which is fantastic because we no longer need to deploy to our servers individually, saving us considerable time. We can see all the integration errors on Fleet Automation, which is super helpful for our product teams to figure out why certain metrics aren't showing up when enabling certain integrations. On Fleet Automation, we can see each variant of the Datadog configuration we have on each host, which is very useful as we can try to synchronize all of them to the same version and configuration.

The Kubernetes Explorer in Datadog is particularly valuable. It gives us a look at each live pod YAML and we can see specific metrics related to each pod. I appreciate the ability to add custom Kubernetes objects to the Orchestration Explorer. It gives our team an easier time to see pods without having to kubectl because sometimes you have permission errors related to that. Sometimes it's just quicker than using kubectl.

Our teams use Datadog more than they used their old observability tool. They're more production-aware, conscious of how their changes are impacting customers, how the changes they make to their application speed up or slow down their app, and the overall request flow. It's a much more developer-friendly tool than other observability tools.

Datadog needs to introduce more hard limits to cost. If we see a huge log spike, administrators should have more control over what happens to save costs. If a service starts logging extensively, I want the ability to automatically direct that log into the cheapest log bucket. This should be the case with many offerings. If we're seeing too much APM , we need to be aware of it and able to stop it rather than having administrators reach out to specific teams.

Datadog has become significantly slower over the last year. They could improve performance at the risk of slowing down feature work. More resources need to go into Fleet Automation because we face many problems with things such as the Ansible  role to install Datadog in non-containerized hosts.

We mainly want to see performance improvements, less time spent looking at costs, the ability to trust that costs will stay reasonable, and an easier way to manage our agents. It is such a powerful tool with much potential on the horizon, but cost control, performance, and agent management need improvement. The main issues are with the administrative side rather than the actual application.

I have been using Datadog for about a year and nine months.

We face a high amount of issues with niche-specific outages that appear to be quite common. AWS  metrics being delayed is something that Datadog posts on their status page. We face a relatively high amount of Datadog issues, but they tend to be small and limited in scope.

We have not experienced any scalability issues.

I have interacted with support. Support quality varies significantly. Some support agents are fantastic, but some tickets take months to resolve.

Neutral

We used Dynatrace  previously, and I believe the switch was due to cost, but that decision was outside my scope as I'm not a decision-maker in that situation.

The initial setup in Kubernetes is not particularly difficult.

I cannot definitively say MTTR has improved as I don't have access to those numbers and don't want to make misleading statements. Developers use it significantly more than our old observability tool. We've seen some cost savings, but we have to be significantly more cost-aware with Datadog than with our previous observability tool because there's more fluctuation and variation in the cost.

One pain point is that it has caused us to spend too much time thinking about the bill. Understand  that while it is an administrative hassle, it is very rewarding to developers.

On a scale of 1-10, I rate Datadog an 8 out of 10.

Our main use case for Datadog  includes monitoring and logs, custom metrics, as well as utilizing the APM  feature and synthetic tests in our day-to-day operations.

A quick specific example of how Datadog  helps with our monitoring and logs comes from all our applications sending logs into Datadog for troubleshooting purposes, with alerts built on top of the logs, and for custom metrics, we send our metrics from the applications via Prometheus to Datadog, building alerts on top of those as well, sometimes sending critical alerts directly to PagerDuty.

We generally have monitors and alerts set up for our applications and specifically rely on them to check our critical business units, such as databases; in GCP, we use Cloud SQL, in AWS , we use RDS , and we also monitor Scylla databases and EC2  instances running Kafka services, which we heavily depend upon. Recently, we migrated from US one to US five, which was a significant shift, requiring us to migrate all alerts and monitors to US five and validate their functionality in the new site.

The best feature Datadog offers is its user-intuitive interface, making it very easy to track logs and custom metrics. We also appreciate the APM  feature, which has helped reduce our log volumes and custom metric volumes, allowing us to turn off some custom metrics.

We recently learned how tags contribute to custom metrics volume, which led us to exclude certain tags to further reduce that volume, and we implement log indexing and exclusion filters, leaving us with much to explore and optimize in our use of Datadog as our major observability platform.

Regarding metrics showing our improvements, the MTTR has been reduced by about 40% after integrating Datadog with PagerDuty, and we've seen our costs significantly drop in the most recent renewal after three years' contract.

Operationally, we spend about 30-40% less time correlating logs and metrics across services, while potential areas for improvement in Datadog include its integration depth and providing more flexible pricing models for large metric and log volumes.

I would suggest having an external Slack channel for urgent requests, which would enable quicker access to support or a dedicated support team for our needs.

I choose eight because, while we have used Datadog for three years and experienced growth in our business and services, the cost has also increased with the growth in metrics and log volumes, and proactive cost management feedback has not been provided to help manage or budget those rising costs. Thus, I'd like to see more proactive cost management in the future, as the pricing model seems to escalate quickly with increasing metrics ingestion and monitoring across clouds. Datadog is a powerful and reliable observability platform, but there is still room for improvement in cost efficiency and usability at scale.

Regarding pricing, setup costs, and licensing, I find Datadog's pricing model transparent but scaling quickly; the base licensing for host integration is straightforward, but costs can rapidly climb as we add custom metrics and log ingestion, especially in dynamic Kubernetes  or multi-cloud environments, with the pricing being moderate to high, and while cost visibility is straightforward, it could become challenging with growing workloads. The upfront setup cost is minimal, mainly involving fine-tuning dashboards, tags, and alerts, making licensing very flexible to enable features as needed.

I have been working in my current field for roughly around 10 years, starting my AWS  journey about 10 years ago, mainly focused on infrastructure and observability.

I believe Datadog is stable.

Datadog's scalability is impressive, as it has the necessary integrations, supports agent-based and cloud-native solutions, and accommodates multi-cloud, multi-region features, making overall performance very good.

Customer support has improved recently with online support available through a portal, allowing for quicker access to help.

Neutral

Previously, we used Splunk SignalFx for a couple of years, switching to Datadog because of Datadog's user-intuitive interface, which was lacking in SignalFx at the time.

Datadog has had a significant positive impact on our organization overall, particularly in visibility, reliability, and cost efficiency, allowing us to centralize metrics, logs, and traces across our cloud, moving from reactive to proactive monitoring, with improvements including faster incident detection and resolution, enhanced service reliability, better cost and resource optimization, and shared dashboards providing the engineering and product teams a single source of truth for system health and performance, thus enhancing our overall observability and operational efficiency.

I believe Datadog has delivered more than its value through reduced downtime, faster recovery, and infrastructure optimization; although we sometimes miss critical alerts, overall, it has improved our team's efficiency by maybe 30% less time spent troubleshooting logs and custom metrics while providing measurable ROI through enhanced system reliability, reduced incident costs, and infrastructure spending optimization.

We only evaluated SignalFx before choosing Datadog, as Datadog offered simpler scaling, better management, broader integrations, and dashboards, allowing for easier monitoring of our multi-cloud setup.

After reducing log and custom metric volumes, we notice a significant reduction in costs without any performance issues on our end, actually seeing a lot of cost reductions.

I strongly recommend using Datadog, but suggest being proactive about resource usage and tracking anomalies monthly.

I find the interview process okay, although it runs longer than I expected, exceeding the anticipated 10 minutes.

My rating for Datadog is 8 out of 10.

My main use case for Datadog  is monitoring performance of Grainger.com and all the components that are involved within it.

A specific example of how I use Datadog  to monitor performance is finding out an issue with an internal bot that we use. We had some issues with some of the commands and we looked into the logs which showed the events from that Slack bot. This was quite useful.

I use Datadog day-to-day to monitor the performance of key services, endpoints, and resources. Currently, we have a migration project for which I created a dashboard to help visualize the performance of key services and endpoints being migrated. At a high level, it helps to capture the performance and health of the services and endpoints.

Datadog has impacted my organization positively as this is our main observability tool when it comes to monitoring services, traces, and all resources within key services. This is our go-to tool and it has helped us to pinpoint issues. One aspect that needs improvement about Datadog is the Watchdog. If there are any escalated conditions or errors happening, it does not indicate which service is causing the issue or which line of code is responsible unless we recreate Watchdog monitors and add the dependency of the GitHub  repo to that service.

When pinpointing issues, it helps us focus on where the problem is. Sometimes it's finding a needle in a haystack, especially when it comes to network issues. This has been our key concern lately. During network outages, we don't know exactly which device has the issue, but network observability is an area we're working towards improving. For regular issues within services, we can see the errors, but we must configure the GitHub  repo associated with that service to see the key issue. Overall, it helps us to pinpoint issues. While I'm not certain about the exact timing of resolution, it does help overall.

In my opinion, the best features Datadog offers are their APM  traces and ability to create dashboards with many customizable metrics, from CPU to thread count to host errors by host and errors by service. Having customized dashboards is really useful, and exploring traces is one of my favorite parts.

We have a list of dashboards primarily showing the key services and APIs related to orders, generating orders, customer direct, and main customer services. Within that list, we have RUM dashboard as well, which shows us the customer impact and the performance of key services which can directly impact customers. During code red or major escalations, I refer to these dashboards for quick analysis of any issues for the services or endpoints.

To make Datadog better, it should be able to pick up error codes automatically. Currently, you have to programmatically configure every single step. In our previous tool, Dynatrace , it could pick up error codes without developers having to explicitly code that into the configuration. Sometimes the APMs are missing the exact error code and error message which is frustrating.

Some minor improvements could include adjusting unit display on dashboards. When request counts go from 900,000 to 1.5 million or 2.2 million for endpoints, the graph keeps all units in thousands rather than converting to millions, which would be more useful and visually appealing.

Datadog Watchdog hasn't been as effective as Dynatrace  Davis, which pinpoints key errors or latency within a specific service and drills down to the specific endpoint. This is an area where Datadog could improve.

We fully migrated to Datadog last year.

In my experience, Datadog is stable, though there's typically at least one or two incidents per week. This amounts to approximately four incidents per month that cause disruption. These incidents are related to log service, indexes, and metric capturing issues, which occur in the Datadog platform more frequently compared to other tools we have.

Datadog's scalability for my organization is pretty straightforward. When it comes to installation, we just have to install it on the respective service hosts and configure it. There's a new way of installing these agents, though I haven't worked on it in a while, but the process is straightforward for installing.

The customer support rates eight out of ten. They require all information upfront and there's still back and forth communication happening. Overall, they provide good service.

Positive

We switched from Dynatrace to Datadog after conducting a survey amongst team members from various service teams. We found that developers preferred using Datadog over Dynatrace. The user interface was more intuitive, modern, and more cloud-focused. Since everybody was moving to cloud, we determined that Datadog would be a suitable tool for us.

When comparing the setup between Dynatrace and Datadog, Datadog required more time and effort. Dynatrace was more straightforward - you simply install the agent and it picks up all the traffic with minimal configuration needed for capturing specific things. Overall, the setup for Datadog was more challenging compared to Dynatrace setup.

I would rate Datadog overall as eight out of ten.

My advice for others looking into using Datadog is to be ready to spend a lot of time setting it up and make sure you have a good plan in terms of analyzing the finances because it can easily cost a lot of money to install agents on your service hosts.