Hey everyone, we've been using Datadog for our web APM needs, and as our data volume has skyrocketed in the last few months, the costs are becoming too high to handle. We're particularly interested in gaining insights into our Python and Node.js services without major code changes or resource-intensive integrations. Since we already use Elastic for one of our internal services and like it, I'm curious if anyone has any recommendations on alternatives. I'm open to any suggestions, whether it's Honeycomb, Sentry, Splunk, New Relic, or any others that might do the job. The key for us is having support for traces, as cloud metrics aren't what we need. Also, if there's a way to connect web APM traces with database traces, that would be awesome. Thanks for any help!
8 Answers
Check out KloudMate! It's built on OTel and is much cheaper than Datadog. It also offers great features like full correlation between metrics, logs, and traces, plus some extras like Incident Management. Definitely worth a look if you're concerned about costs.
Have you considered using Elastic APM? Since you're already comfortable with Elastic for logging, it could make sense to integrate APM into your existing setup.
You might want to check out Parseable! It's an open-source observability platform that focuses on affordability. It runs on object storage like S3, so you won’t hit unexpected costs. No major code changes are needed, and it integrates well into your setup. Plus, they’re working on enhancing DB visibility to help tie together web traces and database calls! Might be worth your time.
Our team made the switch from Datadog to New Relic, and honestly, it has worked out well. The APM agents provide solid code insights and handle distributed tracing well between microservices. Only tried the .NET agent myself, but others on the team have had a good experience with various stacks too.
Take a look at Last9.io. They promise to cut your observability costs by a significant margin and also offer tools for importing DataDog data. Some good names are already on their customer list, so they might be a worthy contender.
If you're using Kubernetes, using Grafana along with Loki can simplify centralized logging for you. It's a solid option for real-time log aggregation.
How about using Loki, Grafana, and Tempo? Tempo is particularly good for handling traces, and you can also replace whatever you're using with AWS managed Prometheus if that fits into your workflow better!
Datadog really does provide a lot of value for the cost. If you want something similar without all the bells and whistles, you might consider a combo like Grafana Cloud and Sentry. This setup can get you up and running fairly quickly with decent initial features, but if you want to customize it like you would with Datadog, be ready to put in some development effort. Just a heads up about Dynatrace; it's pretty good too but can get pricier than Datadog.
Yeah, I heard Dynatrace has a lot of features but doesn't come cheap! Seems like most robust options require some financial commitment.