# observability

Ashish Hanwadikar

02/27/2023, 2:39 AM
Hi all, Would love to know what tools/products people use for observability? SaaS products like Datadog, New Relic, or Open source tools such as, Prometheus, OpenSearch, etc or home grown tools?

Adnan Rahić

02/27/2023, 11:28 AM
I wrote this a few days ago. Hope it can help. 🙌

Cruise Hall

02/27/2023, 12:39 PM
I’ve used Datadog and NewRelic extensively, and AWS CloudWatch intermittently. In my completely subjective opinion: • Datadog is a 2023 Land Rover (luxury for a premium) • NewRelic is a 1995 Toyota Corolla (usually gets you from point A to point B, at a discount; some point Bs not supported) • AWS CloudWatcch is a 2015 Prius (practical and lightweight if you’re already in AWS, few bells+whistles)

Assaf Shutan

02/28/2023, 7:05 PM
There are many observability tools but they all index and send your data to their hot storage for analysis, leading to • High cost of ownership • Gaps in visibility/coverage into production workloads • Inability to ingest all data • High latency, poor scaling & performance, reactive\manual alerting Coralogix is the only open-telemetry observability platform (Logs, Metrics, Traces, Security) that analyzes your data in-stream without relying on storage or indexing for analysis. This next-gen approach to observability is allowing engineering leaders to • Gain full-stack coverage • Eliminate the inability to ingest all data • Store their data in their own S3 bucket VS vendor hot storage (=unlimited retention + lower costs on the market) • Leverage the pro-active approach of decoupling real-time analytics from storage • Reduce observability spend by 70% annually In addition, everything in Coralogix is open-source friendly with no vendor lock-in. For example: Query language is Lucene\PromQL\SQL, Visualisation are in OpenSearch\Grafana and so on (: 📥 Feel free to hit me up if you’re interested in learning more

Kyle Campbell

04/20/2023, 3:00 PM
We went from NewRelic to using Grafana/Prometheus. We were finding the cost to observe everything was too exorbitant in our original NR, resulting us in leaving areas uncovered, and DataDog was going to be even worse, The open source approach has it’s drawbacks definitely but it’s definitely workable